ПРЕОБРАЗОВАНИЕ ИЗ ЗОННОГО ФОРМАТА В ДЕСЯТИЧНЫЙ ФОРМАТ С ПЛАВАЮЩЕЙ ТОЧКОЙ

Группа изобретений относится к вычислительной технике и может быть использована для преобразования данных. Техническим результатом является повышение производительности. Способ содержит этапы получения процессором машинной команды для выполнения, причем машинная команда определена для выполнения компьютером согласно архитектуре компьютера и содержит по меньшей мере одно поле кода операции, предоставляющее код операции, причем код операции идентифицирует функцию преобразования из зонного в десятичный с плавающей точкой; поле первого регистра, определяющее ячейку первого операнда; поле второго регистра и поле смещения, причем содержимое второго регистра, определенного полем второго регистра, комбинируется с содержимым поля смещения с образованием адреса второго операнда; и директиву знака, использующуюся для указания того, имеет ли второй операнд поле знака; и выполнение машинной команды, включающее преобразование второго операнда в зонном формате в десятичный формат с плавающей точкой; и ...

УСТРОЙСТВО АНАЛОГО-ЦИФРОВОГО ПРЕОБРАЗОВАНИЯ С ПРОМЕЖУТОЧНЫМ ПРЕОБРАЗОВАНИЕМ АНАЛОГОВОГО СИГНАЛА В ЦИФРОВОЙ КОД С ПЛАВАЮЩЕЙ ЗАПЯТОЙ

Изобретение относится к технике электрической связи и может быть использовано в цифровых системах передачи. Сущность заявляемого технического решения является повышение точности аналого-цифрового преобразования в цифровых системах передачи путем уменьшения количества весовых резисторов в аналого-цифровом преобразователе за счет введения в устройство дискретного делителя напряжения, определяющего порядок числа разрядностью II и логического преобразователя, осуществляющего преобразование разрядного кода с плавающей запятой в стандартный 8-разрядный код ИКМ.

Последовательный многоканальный анализатор импульсов

ПРЕОБРАЗОВАТЕЛЬ ЧИСЕЛ

WANDLER FUER NUMERISCHE AUSDRUECKE UND IHN BENUETZENDER VEKTORPROZESSOR

Verknüpfte Quantisierung und Normalisierung stützend auf eine Mantisse - Exponent Darstellung einer Quantisierungsparameterfunktion

VERFAHREN ZUR CODIERUNG ODER DECODIERUNG VON SPRACHSIGNALABTASTWERTEN SOWIE CODER BZW. DECODER

Method and apparatus for floating point operations and format conversion operations

Method and apparatus for floating point operations and format conversion operations

A data processing apparatus and method for converting a number between fixed-point and floating-point representations

DIGITALE SIGNALVERARBEITUNGSEINRICHTUNG

CONVERSION OF A FLOATING DECIMAL POINT BINARY NUMBER TO A COMPLEMENT ON TWO BINARY NUMBER

PROCEDURE FOR THE CODING OR DECODING OF LANGUAGE SIGNAL SCANNING VALUES AS WELL AS CODER AND/OR DECODER

PROCEDURE AND DEVICE FOR THE SIGNALKOMPRESSION IN THE FIXED POINT FORMAT WITHOUT BIAS

ELECTRONIC COMPUTERS

FLOATING POINT ENCODING SYSTEMS AND METHODS

FLOATING POINT ENCODING SYSTEMS AND METHODS

... ²Systems and methods for encoding floating point numbers. A system can include ²encoding logic which encodes invalid floating point representations as valid ²data. ²Decoding logic can be used to recognize the invalid floating point ²representations and map ²can provide the invalid floating point representations to valid data values. ²The decoding ²logic then can provide the valid data values so that operations on the valid ²data values can ²be performed in accordance with instructions received from an associated ²program.² ...

AN IMAGE DECODING APPARATUS FOR OBTAINING A DECODED IMAGE BY DECODING AN ENCODED DATA

An image decoding apparatus for obtaining a decoded image by decoding an encoded data is provided. The apparatus comprises a dequantization module for obtaining a transform coefficient K[i][j] by dequantizing the quantization value L[i][j] for each block obtained by dividing an image; an inverse transformation module for inverse integer transforming the transform coefficient derived by the dequantization module; and a normalization module for normalizing the inverse integer transformed transform coefficient by the inverse transformation module, the inverse integer transform is executed by an inverse transform matrix in which a different basis function has a different norm, the dequantization module uses a mantissa portion matrix element Bm(QP)[i][j] which is a function of a quantization parameter QP and an exponential portion Be(QP) which is a function of the quantization parameter QP to derive the transform coefficient K[i][j] as follows: K[i][j] =[L[i][j]×Bm(QP)[i][j]] < Подробнее

METHOD FOR REDUCED BIT-DEPTH QUANTIZATION

A method is provided for the quantization of a coefficient. The method comprises: supplying a coefficient K; supplying a quantization parameter (QP); forming a quantization value (I) from the coefficient K using a mantissa portion (Am (QP)) and an exponential portion (x Ae(QP)). Typically, the value of x is 2. In some aspects of the method, forming a quantization value (I) from the coefficient K includes L = K*A(QP) = K*Am(QP)* (2Ae(QP)). In other aspects, the method further comprises: normalizing the quantization value by 2N as follows Ln = L/2N = K*Am(QP)/2(N- Ae(QP)) In some aspects, forming a quantization value includes forming a set of recursive quantization factors with a period P, where A(QP+P) = A(QP)/x. Forming recursive quantization factors includes forming recursive mantissa factors, where Am(QP) = Am(QP mod P), and forming recursive quantization factors includes forming recursive exponential factors, where Ae(QP) = Ae(QP mod P) - QP/P.

CALCOLATORE ELETTRONICO PROGRAMMABILE.

Model calculation unit and control unit for calculating a data-based function model having data in various number formats

A model calculation unit (3) for calculating a data-based function model, in particular a Gaussian process model, in a control unit (1), including: a processor core (31) to carry out a strictly hardware-based calculation of an algorithm for a data-based function model, the data-based function model being calculated using provided calculation data, in particular hyperparameters and node data; and a strictly hardware-based conversion unit (35) to provide the processor core with at least a portion of the calculation data, in particular the provided node data, in a predefined number format.

ANALOGUE SIGNAL SAMPLE AMPLIFIER

A method for improving the conditions of use of a digital computer and calculating machine using this process

Encoder/Decoder for improving a voice quality in G.711 codec

압축 및 압축해제를 이용한 향상된 멀티 프로세서 파형 데이터 교환

... 멀티코어 프로세싱 환경에서 파형 데이터의 설정가능한 압축 및 압축해제는 코어들 간의 데이터 전송 효율을 향상시키고 데이터 저장매체의 자원을 절약한다. 파형 데이터 처리 시스템들에서, 입력, 중간, 및 출력 파형 데이터는 종종 코어들 간에 그리고 오프칩 메모리와의 사이에서 교환된다. 각각의 코어에서, 하나의 단일 설정가능 압축기 및 하나의 단일 설정가능 압축해제기는, 정수 또는 부동소수점 파형 데이터를 압축 및 압축해제하도록 설정될 수 있다. 메모리 제어기에서, 상기 설정가능 압축기는 오프칩 메모리로의 전송을 위하여 정수 또는 부동소수점 파형 데이터를 압축하고, 상기 설정가능 압축해제기는 오프칩 메모리로부터 수신된 압축된 패킷들을 압축해제한다. 압축은, 반도체 또는 자성체 메모리 안에 파형 데이터를 저장하기 위하여 필요로 하는, 메모리 또는 저장매체를 감소시킨다. 압축은 파형 데이터를 교환하기 위하여 필요한 지연시간 및 대역폭 모두를 감소시킨다. 본 요약서는 청구범위의 청구항들에 의하여 정의되는 본 발명의 범위를 제한하지 않는다.

JOINT QUANTIZATION AND INTEGER TRANSFORM NORMALIZATION USING A MANTISSA-EXPONENT REPRESENTATION OF A QUANTISATION PARAMETER

SYSTEM AND METHOD FOR THE TRANSFER OF FLOATING POINT NUMBERS IN XML FORMAT WITHOUT LOSS OF INFORMATION

Die Erfindung betrifft ein System und Verfahren zur verlust- freien Übertragung von Fließkommazahlen in XML, das insbesondere im Bereich der Automatisierungstechnik zum Einsatz kommt, um beispielsweise Prozessdaten eines Prozessautomatisierungssystems in XML insbesondere über Internet an ein Bedien- und Beobachtungssystem und/oder an ein weiteres Automatisierungssystem etc. zu übermitteln. Zur Vermeidung von Rundungsfehlern, die vor allem bei der Umwandlung von Zahlenwerten aus einer binären in eine dezimale Darstellung vorkommen, wird vorgeschlagen, bei einer textuellen Darstellung einer Fließkommazahl keine dezimale Basis vorzusehen, sondern eine Hexadezimale.

MULTIFUNCTION FLOATING POINT ADDITION/SUBTRACTION PIPELINE AND BIPARTITE LOOK-UP TABLE

An add/subtract pipeline has far and close data paths. The far data path handles effective addition operations, and effective subtraction operations for operands having an absolute exponent difference greater than one. The close data path handles all other effective subtraction operations. Selection of the output value in the close data path effectuates the round-to-nearest operation. Floating point-to-integer conversion may be executed in the far data path integer-to-floating point instructions in the close data path. The execution unit may include a pluraliy of add/subtract pipelines, allowing vectored add, subtract, and integer/floating point conversion instructions to be performed. Additional arithmetic instructions (such as reverse subtract and accumulate functions minimum/maximum and comparison) may also be implemented. A method for generating entries for a bipartite look-up table having base and difference table portions is also disclosed. So is a multi-function look-up table.

Method and apparatus for compressing/decompressing data using floating point

Disclosed are a method and apparatus for compressing/decompressing data using floating points. As technology for compressing/decompressing data using floating points for efficient memory management, there are provided a method and apparatus for compressing/decompressing data using floating points, in which a log table is used to compress/decompress data, whereby not only data loss caused by compression/decompression can be minimized, but also degradation of performance can be prevented through a floating point representation even though the same number of bits are used for compression.

Fixed-point data/floating-point data converting apparatus

A fixed-point data/floating-point data converting apparatus has: a priority encoder with input terminals respectively receiving output signals from logic gates (first logic gates) for use in detecting whether or not all of the digits fixed-point data are "0", to obtain a number of continuous "zero" digits from the most significant digit of the fixed-point data, in accordance with the contents of the signals received thereby; a shifter for shifting the fixed-point data to the left by the number of digits of continuous "zero" digits, thereby obtaining a mantissa part; an adder for subtracting the number of continuous "zero" digits from a reference value, thereby obtaining an exponent part; and a data selector for generating the floating-point data in accordance with the subtraction results from the adder, the shift results from the shifter, all-zero data, and an output signal from the logic gate (first logic gate) for detecting whether or not all digits of the fixed-point data are "0", in ...

Binary data identification circuit

A binary data identification circuit including first and second potential terminals set to first and second logical potential levels, a series circuit including first to (n-1)th transfer gates whose conduction states are controlled responsive to 1st to (n-1)th bit signals of an input operand containing first to nth bit signals, one end of the series circuit being connected to the first potential terminal, nth to (2n-2)th transfer gates which are controlled, responsive to the first to (n-1)th bit signals, so as to have opposite conduction states with respect to those of the first to (n-1)th transfer gates, the nth to (2n-2)th transfer gates being connected at their source to the second potential terminal and at their drain to the drains of the first to (n-1)th transfer gates; and first to nth logic gates whose first input terminals receive the first to nth bit signals and whose second terminals are connected to the first potential terminal and to the drains of the first to (n-1)th transfer ...

Systems and methods for reduced bit-depth processing in video-related data with frequency weighting matrices

Embodiments of the present invention comprise systems and methods for processing of data related to video wherein reduced bit depth intermediate calculations are enabled.

MACHINE LEARNING TRAINING ARCHITECTURE FOR PROGRAMMABLE DEVICES

A programmable device may be configured to support machine learning training operations using matrix multiplication circuitry implemented on a systolic array. The systolic array includes an array of processing elements, each of which includes hybrid floating-point dot-product circuitry. The hybrid dot-product circuitry has a hard data path that uses digital signal processing (DSP) blocks operating in floating-point mode and a hard/soft data path that uses DSP blocks operating in fixed-point mode operated in conjunction with general purpose soft logic. The hard/soft data path includes 2-element dot-product circuits that feed an adder tree. Results from the hard data path are combined with the adder tree using format conversion and normalization circuitry. Inputs to the hybrid dot-product circuitry may be in the BFLOAT16 format. The hard data path may be in the single precision format. The hard/soft data path uses a custom format that is similar to but different than BFLOAT16. 1. An integrated circuit , comprising:first digital signal processing (DSP) blocks configured to operate in a floating-point mode;second digital signal processing (DSP) blocks configured to operate in a fixed-point mode that is different than the floating-point mode; andan adder configured to receive a first signal from the first DSP blocks operating in the floating-point mode and a second signal from the second DSP blocks operating in the fixed-point mode.2. The integrated circuit of claim 1 , wherein the first DSP blocks are part of a hard data path claim 1 , and wherein the second DSP blocks are part of a hard and soft data path.3. The integrated circuit of claim 1 , wherein the first and second DSP blocks are configured to receive input signals of a first floating-point format claim 1 , and wherein the first DSP blocks are configured to output signals in a second floating-point format that is different than the first floating-point format.4. The integrated circuit of claim 3 , wherein the ...

NEURAL NETWORK DEVICE FOR NEURAL NETWORK OPERATION, METHOD OF OPERATING NEURAL NETWORK DEVICE, AND APPLICATION PROCESSOR INCLUDING THE NEURAL NETWORK DEVICE

Provided are a neural network device for performing a neural network operation, a method of operating the neural network device, and an application processor including the neural network device. The neural network device includes a direct memory access (DMA) controller configured to receive floating-point data from a memory; a data converter configured to convert the floating-point data received through the DMA controller to integer-type data; and a processor configured to perform a neural network operation based on an integer operation by using the integer-type data provided from the data converter.

Model calculation unit and control unit for calculating a data-based function model having data in various number formats

A model calculation unit for calculating a data-based function model, in particular a Gaussian process model, in a control unit, including: a processor core to carry out a strictly hardware-based calculation of an algorithm for a data-based function model, the data-based function model being calculated using provided calculation data, in particular hyperparameters and node data; and a strictly hardware-based conversion unit to provide the processor core with at least a portion of the calculation data, in particular the provided node data, in a predefined number format. 1. A model calculation unit for calculating a data-based function model in a control unit , comprising:a processor core configured to carry out a strictly hardware-based calculation of an algorithm for a data-based function model, the data-based function model being calculated using provided calculation data; anda strictly hardware-based conversion unit configured to provide the processor core with at least a portion of the calculation data in a predefined number format.2. The model calculation unit of claim 1 , wherein the conversion unit is configured to carry out claim 1 , as a function of a selection signal claim 1 , a conversion of at least a portion of the provided calculation data of one number format which differs from the predefined number format claim 1 , into the predefined number format.3. The model calculation unit of claim 1 , wherein the conversion unit includes at least one conversion block for converting a first number format of data into the predefined number format claim 1 , as well as a multiplexer claim 1 , in order claim 1 , as a function of the selection signal claim 1 , to forward to the processor core either the provided calculation data or the calculation data converted by one of the at least one conversion blocks into the predefined number format.4. The model calculation unit of claim 3 , wherein the predefined number format corresponds to a 32-bit floating point number ...

Floating point to fixed point conversion

A binary logic circuit converts a number in floating point format having an exponent E of ew bits, an exponent bias B given by B=2ew-1−1, and a significand comprising a mantissa M of mw bits into a fixed point format with an integer width of iw bits and a fractional width of fw bits. The circuit includes a shifter operable to receive a significand input comprising a contiguous set of the most significant bits of the significand and configured to left-shift the significand input by a number of bits equal to the value represented by k least significant bits of the exponent to generate a shifter output, wherein min{(ew−1),bitwidth(iw−2−sy)}≤k≤(ew−1) where sy=1 for a signed floating point number and sy=0 for an unsigned floating point number, and a multiplexer coupled to the shifter and configured to: receive an input comprising a contiguous set of bits of the shifter output; and output the input if the most significant bit of the exponent is equal to one.

LOSSLESS CODING METHOD FOR DIGITAL SIGNAL IN FLOATING-POINT FORMAT, DECODING METHOD, APPARATUS THEREFOR AND PROGRAMS THEREFOR

Umsetzung einer Gleitkommabinärzahl in eine Zweierkomplementbinärzahl

Multifunktionelles Gleitkomma-Additions/Subtraktions-Fliessband

Arithmetic pipeline

Conversion between floating point data and scalar data in different architectural registers and memory

Floating point arithmetic and dot product pipelineoperations

A method and apparatus for parallel conversion of color values from a single precision floating point format to an integer format

A data compression method

A method for transmitting the value of a parameter in a compressed form, the method comprising the steps of: accepting successive numbers representing the value of a parameter; manipulating each number, the manipulation comprising placing the number in a form comprising a mantissa and an exponent, and defining a transmission mantissa to be transmitted; transmitting to a receiver, in turn, the transmission mantissas only of the successive numbers; receiving the transmission mantissas of the successive numbers at the receiver; maintaining a record, at the receiver, of a receiver variable, the receiver variable initially corresponding to the exponent of an initial number; formulating at the receiver, for each received transmission mantissa, a reconstructed number comprising at least the transmission mantissa and an exponent corresponding to the receiver variable; and altering the receiver variable in a first manner if the transmission mantissa of the current number fulfils a first criterion ...

Floating point to fixed point conversion

A floating point format has an exponent of ew bits, an exponent bias B = 2 ew-1 1 and a mantissa of mw bits. A fixed point format has an integer width of iw and a fractional width of fw. A circuit to convert from the floating point format to the fixed point format has a left shifter 401 and a right shifter 403. Both shifters take the mantissa of the floating point number 409 as an input. The shift count for the left shift is the first k bits of the exponent, where min((ew 1), bitwidth(iw 2 sy)) ≤ k ≤ (ew 1), where sy = 1 for a signed floating point number and sy = 0 for an unsigned one. The shift count for the right shifter is the inverse of the p least significant bits of the inverted exponent where min((ew 1), bitwidth(fw)) ≤ p ≤ (ew 1). A multiplexer 405 outputs the output of the left shifter if the most significant bit of the floating point number is one and the output of the right shifter otherwise.

FLOATING POINT TO FIXED POINT CONVERSION

TRANSFORMATION CODER, - DECODER AND CODER/CDECODER WITH LOW BIT RATE F�R AUDIO'S APPLICATIONS OF HIGH ONES QUALIT�T

ADAPTIVE BIT ALLOCATION FOR AUDIO KODER AND DEKODER

TRANSFORMATION CODER, - DECODERS AND CODER/DECODER WITH LOW BIT RATE FOR AUDIO'S APPLICATIONS OF HIGH QUALITY.

LOW BIT RATE TRANSFORM CODER, DECODER AND ENCODER/DECODER FOR HIGH-QUALITY AUDIO

Method and apparatus for floating point operations and format conversion operations

LOW BIT RATE TRANSFORM CODER, DECODER AND ENCODER/DECODER FOR HIGH-QUALITY AUDIO

LOW BIT RATE TRANSFORM CODER, DECODER AND ENCODER/DECODER FOR HIGH QUALITY AUDIO

SYSTEM FOR CONVERSION BETWEEN CODED BYTE AND FLOATING POINT FORMATS

DIGITAL SIGNAL PROCESSING DEVICE

The invention relates to a digital signal processing device comprising: input storage means (3; 5); a computational device (4) that is connected to said means, defines a data path (9) and contains at least one arithmetic unit (6) in addition to a control input (2a) for specifying calculation operations; and output storage means (8). The data path (9) between the arithmetic unit (6; 7) and the output storage means (8) is equipped with a number-format conversion unit (10) comprising a shift unit (17). A number-format specification unit (11) and a control unit (17'), which is connected to the latter and calculates required shift operations on the basis of the number-format specification, are assigned to the number-format conversion unit (10). Formatting operations are calculated automatically using input and output format information and corresponding commands are applied to the shift unit (17).

METHOD FOR REDUCED BIT-DEPTH QUANTIZATION

A method is provided for the quantization of a coefficient. The method comprises: supplying a coefficient K; supplying a quantization parameter (QP); forming a quantization value (L) from the coefficient K using a mantissa portion (Am(QP)) and an exponential portion (xAe(QP)). Typically, the value of x is 2. In some aspects of the method, forming a quantization value (L) from the coefficient K includes L = K*A(QP) = K*Am(QP)* (2Ae(QP)). In other aspects, the method further comprises: normalizing the quantization value by 2N as follows: Ln = L/2N = K*Am(QP)/2(N-Ae(QP)). In some aspects, forming a quantization value includes forming a set of recursive quantization factors with a period P, where A(QP+P) = A(QP)/x. Forming recursive quantization factors includes forming recursive mantissa factors, where Am(QP) = Am(QP mod P), and forming recursive quantization factors includes forming recursive exponential factors, where Ae(QP) = Ae(QP mod P) - QP/P.

METHOD FOR REDUCED BIT-DEPTH QUANTIZATION

A method is provided for the quantization of a coefficient. The method comprises: supplying a coefficient K; supplying a quantization parameter (QP); forming a quantization value (I) from the coefficient K using a mantissa portion (Am (QP)) and an exponential portion (x Ae(QP)). Typically, the value of x is 2. In some aspects of the method, forming a quantization value (I) from the coefficient K includes L = K*A(QP) = K*Am(QP)* (2Ae(QP)). In other aspects, the method further comprises: normalizing the quantization value by 2N as follows Ln = L/2N = K*Am(QP)/2(N- Ae(QP)) In some aspects, forming a quantization value includes forming a set of recursive quantization factors with a period P, where A(QP+P) = A(QP)/x. Forming recursive quantization factors includes forming recursive mantissa factors, where Am(QP) = Am(QP mod P), and forming recursive quantization factors includes forming recursive exponential factors, where Ae(QP) = Ae(QP mod P) - QP/P.

METHOD FOR REDUCED BIT-DEPTH QUANTIZATION

A method is provided for the quantization of a coefficient. The method comprises: supplying a coefficient K; supplying a quantization parameter (QP); forming a quantization value (I) from the coefficient K using a mantissa portion (Am (QP)) and an exponential portion (x Ae(QP)). Typically, the value of x is 2. In some aspects of the method, forming a quantization value (I) from the coefficient K includes L = K*A(QP) = K*Am(QP)*(2Ae(QP)). In other aspects, the method further comprises: normalizing the quantization value by 2N as follows Ln = L/2N = K*Am(QP)/2(N- Ae(QP)). In some aspects, forming a quantization value includes forming a set of recursive quantization factors with a period P, where A(QP+P) = A(QP)/x. Forming recursive quantization factors includes forming recursive mantissa factors, where Am(QP) = Am(QP mod P), and forming recursive quantization factors includes forming recursive exponential factors, where Ae(QP) = Ae(QP mod P) - QP/P.

METHOD FOR REDUCED BIT-DEPTH QUANTIZATION

A method is provided for the quantization of a coefficient. The method comprises: supplying a coefficient K; supplying a quantization parameter (QP); forming a quantization value (I) from the coefficient K using a mantissa portion (Am (QP)) and an exponential portion (x Ae(QP)). Typically, the value of x is 2. In some aspects of the method, forming a quantization value (I) from the coefficient K includes L = K*A(QP) = K*Am(QP)* (2Ae(QP)). In other aspects, the method further comprises: normalizing the quantization value by 2N as follows Ln = L/2N = K*Am(QP)/2(N-Ae(QP)) In some aspects, forming a quantization value includes forming a set of recursive quantization factors with a period P, where A(QP+P) = A(QP)/x. Forming recursive quantization factors includes forming recursive mantissa factors, where Am(QP) = Am(QP mod P), and forming recursive quantization factors includes forming recursive exponential factors, where Ae(QP) = Ae(QP mod P) - QP/P.

METHOD FOR REDUCED BIT-DEPTH QUANTIZATION

A method is provided for the quantization of a coefficient. The method comprises: supplying a coefficient K; supplying a quantization parameter (QP); forming a quantization value (I) from the coefficient K using a mantissa portion (Am (QP)) and an exponential portion (x Ae(QP)). Typically, the value of x is 2. In some aspects of the method, forming a quantization value (I) from the coefficient K includes L = K*A(QP) = K*Am(QP)* (2Ae(QP)). In other aspects, the method further comprises: normalizing the quantization value by 2N as follows Ln = L/2N = K*Am(QP)/2(N- Ae(QP)) In some aspects, forming a quantization value includes forming a set of recursive quantization factors with a period P, where A(QP+P) = A(QP)/x. Forming recursive quantization factors includes forming recursive mantissa factors, where Am(QP) = Am(QP mod P), and forming recursive quantization factors includes forming recursive exponential factors, where Ae(QP) = Ae(QP mod P) - QP/P.

A data compression method

A method for transmitting the value of a parameter in a compressed form, the method comprising the steps of: accepting successive numbers representing the value of a parameter; manipulating each number, the manipulation comprising placing the number in a form comprising a mantissa and an exponent, and defining a transmission mantissa to be transmitted; transmitting to a receiver, in turn, the transmission mantissas only of the successive numbers; and receiving the transmission mantissas of the successive numbers at the receiver, characterised by the steps of maintaining a record, at the receiver, of a receiver variable, the receiver variable initially corresponding to the exponent of an initial number; formulating at the receiver, for each received transmission mantissa, a reconstructed number comprising at least the transmission mantissa and an exponent corresponding to the receiver variable; and altering the receiver variable in a first manner if the transmission mantissa of the current ...

Apparatus and method for selectively storing depth information of a 3-D image

In a computer system for creating images of three-dimensional objects, an apparatus and method for processing depth values representing the relative depths of the objects. The depth values are transformed according to projections in order to give the appearance of depth when the objects are displayed on a two-dimensional computer screen. These transformed depth values are then interpolated using N bits of precision. Next, the depth values are encoded into a format whereby the depth values have less than N bits. These encoded depth values are stored into memory. In displaying the objects, the encoded depth values are read from memory and compared to determine the relative depths of the objects.

TRAINING NEURAL NETWORKS USING MIXED PRECISION COMPUTATIONS

A system for training a neural network receives training data and performing lower precision format training calculations using lower precision format data at one or more training phases. One or more results from the lower precision format training calculations are converted to higher precision format data, and higher precision format training calculations are performed using the higher precision format data at one or more additional training phases. The neural network is modified using the results from the one or more additional training phases. The mixed precision format training calculations train the neural network more efficiently, while maintaining an overall accuracy.

Apparatus for processing floating-point data having exponents of variable length

Herein disclosed is a floating-point data processing apparatus for generating exponent data (105) of fixed length from floating-point data (100) composed of: a sign bit indicating a mantissa sign; a first exponent part which has a bit length determined in dependence upon a significant bit length necessary for binarily expressing an exponent and which has all its bits determined at 1 or 0 in dependence upon said mantissa sign and the sign of said exponent: a second exponent part which has its bit length determined in dependence upon the bit length of said first exponent part, which has a predetermined relationship determined in dependence upon the sign of said exponent and said mantissa sign with the significant bit part when said exponent is binarily expressed, and the leading bit of which has a value different from the value of one bit of said first exponent part; and a mantissa part which has a plurality of bits having bit length determined in dependence upon the value of said exponent ...

浮動小数点データを符号化する方法、浮動小数点データを復号化する方法並びに関連する符号器及び復号器

СПОСОБ И УСТРОЙСТВО ДЛЯ СЖАТИЯ СИГНАЛОВ В ФОРМАТЕ С ФИКСИРОВАННОЙ ЗАПЯТОЙ БЕЗ ВВЕДЕНИЯ СМЕЩЕНИЯ

Изобретение относится к сжатию сигналов с фиксированной запятой без введения смещения. Техническим результатом является обеспечение сжатия одноразрядного сигнала с минимальным количеством аппаратных средств. Сигналы сжимают в соответствии с методом псевдослучайного округления, в котором значения сигнала, как в большую, так и в меньшую стороны с приблизительно равной вероятностью. Числовые свойства входного сигнала используются для того, чтобы определить, необходимо ли значение сигнала округлять в большую или меньшую стороны. Поэтому, сжатие сигнала позволяет вводить многочисленные запятые внутри системы без накопления смещения сигнала и ухудшения характеристик работы. 3 н. и 2 з.п. ф-лы, 8 ил.

СПОСОБ И УСТРОЙСТВО ПЕРЕКОДИРОВКИ N-БИТНЫХ СЛОВ ВМ-БИТНЫЕ СЛОВА, ГДЕ М МЕНЬШЕ N

... 1. Способ перекодировки N-битного слова в M-битное слово, N > M, отличающийся тем, что он содержит следующие этапы, на которых разбивают N-битное слово на порядок числа и мантиссу, каждая из которых имеет размер, который варьируется согласно значению упомянутого N-битного слова, причем размер мантиссы увеличивается вместе со значением упомянутого N-битного слова, и кодируют порядок числа N-битного слова на переменное число бит A и удаляют при необходимости самые младшие значащие биты мантиссы, чтобы получить мантиссу с переменным числом бит B, где A+B=M. 2. Устройство перекодировки N-битного слова в M-битное слово, N > M, отличающееся тем, что оно содержит средство для разбивки N-битного слова на порядок числа и мантиссу, каждая из которых имеет размер, который варьируется согласно значению упомянутого N-битного слова, причем размер мантиссы увеличивается вместе со значением упомянутого N-битного слова, и средство для кодирования порядка числа N-битного слова на переменное число бит A и ...

Устройство для преобразования двоичных кодов

УСТРОЙСТВО ДЛЯ ПРЕОБРАЗОВАНИЯ ПАРАЛЛЕЛЬНОГО ДВОИЧНОГО КОДА ЧИСЛА

VORRICHTUNG UND VERFAHREN MIT EINER AUFLÖSUNGSVERBESSERTEN CHARAKTERISTIK ZUR ERHÖHTEN GENAUIGKEIT BEI DER UMWANDLUNG DES BENÖTIGTEN CHEMISCHEN MECHANISCHEN POLIERDRUCKS IN KRAFT, WELCHE BEI EINEM POLIERKOPF AUF EINEM WAFER ANGEWENDET WIRD

Graphics data parallel conversion method for multimedia system

Several color components in floating point format are stored in 128-bit register set (906) in packed format. The floating point format of the color components is converted into integer format and is stored in 64-bit register set (904) in packed format. The color components for the pixels are assembled using integer number stored in 64-bit register set. An independent claim is also included for graphics data parallel conversion system.

FIXED POINT TO FLOATING POINT CONVERSION

Generating fixed point data with reduced circuit scale

A decoding rate is improved while reducing a circuit scale, in a fixed point data generating circuit. When a plurality of floating point data are inputted, for example, the maximum floating point data is detected as a reference data among the plurality of floating point data, in a MAX value detecting circuit 10. Then, in an exponent part subtractor 20, differences are obtained between the values of exponent parts of the plurality of inputted floating point data and the value of an exponent part of the maximum floating point data. Thereafter, in the shift register 30, mantissa parts of the inputted floating point data are shifted by the differences obtained in the exponent part subtractor 20, and, in a bit extracting portion 40, a predetermined number of bits of the shifted mantissa parts are extracted as fixed point data to be inputted to a Viterbi decoder.

Converting floating point numbers to reduce the precision

Floating point to fixed point conversion

Arithmetic pipeline

A method and apparatus for parallel conversion of color values from a single precision floating point format to an integer format

Electronic calculators

... 1,145,661. Electric digital calculators; fixed or floating point representation. CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE. 2 March, 1966 [2 March, 1965], No. 9208/66. Heading G4A. An electronic arithmetic calculator may perform calculations on numbers expressed in fixed and/or floating point representation and produces results in either representation according to certain predetermined conditions. An initial check is made to see if both numbers to be operated on are expressed in the same notation, if not, the fixed point number is converted to floating point representation and the calculation is performed in floating point. If both are in floating point, the calculation is performed in floating point whereas, if both are in fixed point, the result is expressed in fixed point unless overflow occurs or greater accuracy can be achieved in floating point. Floating point numbers may be converted to fixed point numbers when the value of the exponent is such that this can be done. 35-bit words ...

FLOATING POINT

MULTIFUNCTION ARITHMETIC LOGIC UNIT CIRCUITS

TRANSFORMATION CODER, - DECODERS AND CODER/CDECODERS WITH LOW BIT RATE FOR AUDIO'S APPLICATIONS OF HIGH QUALITY

LINKED QUANTIZATION AND NORMALIZATION SUPPORTING BY A MANTISSA - EXPONENT REPRESENTATION TO A QUANTIZATION PARAMETER FUNCTION

Apparatus and methods with resolution enhancement feature for improving accuracy of conversion of required chemical mechanical polishing pressure to force to be applied by polishing head to wafer

Method for encoding floating-point data, method for decoding floating-point data, and corresponding encoder and decoder

An algorithm for efficiently compressing floating-point data in 3D meshes is disclosed. 3D meshes are represented by topology data, geometry data and property data. Geometry data specify vertex locations and are usually represented by floating-point coordinates. While geometry data are usually compressed by quantization, prediction and entropy coding, the present invention uses no prediction. A floating-point number consists of mantissa and exponent, and normally the exponent, sign and mantissa are compressed separately. A method for encoding floating-point formatted data comprises determining if a current floating-point value was previously stored in a memory, storing the current value in the memory if it was not previously stored in the memory, and encoding it. Otherwise, if the current floating-point value was previously stored in a memory, the storage position of the value within the memory is determined and a reference pointing to the storage position is encoded.

Conversion between packed floating point data and packed 32-bit integer data in different architectural registers

AMPLIFICATEUR D'ECHANTILLONS DE SIGNAUX ANALOGIQUES AVEC REGLAGE DE GAIN AUTOMATIQUE AVEC COMPENSATION D'ERREUR

AN IMAGE DECODING APPARATUS FOR OBTAINING A DECODED IMAGE BY DECODING AN ENCODED DATA

A computer-implemented dequantization method utilized in video decoding using a processor and memory is provided. The method determines a two- dimensional transform coefficient matrix K by dequantizing a quantization value matrix L. The method comprising steps of inputting an element L [i][j] of the quantization value matrix; inputting a quantization parameter OP; and determining the transform coefficient matrix K. The element K [i][j] of the transform coefficient matrix is determined by using a mantissa portion matrix element B(QP) [i][j] being a function of QP according to the following equation: K [i][j] = L [i][j] x B(QP mod P) [i][j] x 2 B0+QP/P, where B0 and P are constants.

METHOD FOR REDUCED BIT-DEPTH QUANTIZATION

DEVICE OF CONVERSION OF NUMBERS.

DATA CODING

Embodiments of methods, systems and/or devices for data coding are disclosed. © KIPO & WIPO 2009 ...

ANCHORED DATA ELEMENT CONVERSION

An apparatus is provided, that includes an instruction decoder responsive to an anchored-data processing instruction, to generate one or more control signals. Conversion circuitry is responsive to the one or more control signals to perform a conversion from a data value to an anchored-data select value. The conversion is based on anchor metadata indicative of a given range of significance for the anchored-data select value. Output circuitry is responsive to the one or more control signals, to write the anchored-data select value to a register.

Floating point to logarithm converter

A method converts an n-bit floating point number to a logarithmic representation, the number having a first set of bits assigned to a mantissa, and a second set of bits assigned to an exponent. The logarithmic representation has a characteristic and a fraction. The method: (a) sets the characteristic of the logarithm equal to the second set of bits; (b) selects from a logarithm table a first logarithm of a first subset of the first set; (c) selects from the logarithm table a slope of a logarithmic function at the first logarithm; (d) multiplies the slope times a second subset of the first set to obtain an interpolated value; (e) adds the interpolated value to the first logarithm to obtain the fraction; and (f) adds the characteristic and the fraction to obtain the logarithmic representation.

FLOATING POINT COMPUTATION FOR HYBRID FORMATS

Various embodiments are provided for performing hybrid precision floating point format computation via a simplified superset floating point unit in a computing system. One or more inputs, represented as a plurality of floating point number formats, may be converted into a superset floating point format prior to computation by one or more simplified superset floating point units (ssFPUs). A compute operation may be performed on the one or more inputs represented as the superset floating point format using the one or more ssFPUs. 1. A method for performing hybrid precision floating point format computation via a simplified superset floating point unit by one or more processors comprising:converting one or more inputs, represented as a plurality of floating point number formats, into a superset floating point format prior to computation by one or more simplified superset floating point units (ssFPUs); andperforming a compute operation on the one or more inputs represented as the superset floating point format using the one or more ssFPUs.2. The method of claim 1 , further including:identifying the plurality of floating point number formats as a very low precision (“VLP”) format comprising a sign bit, exponent bits (e), and mantissa bits (m), wherein the VLP is an 8-bit floating point format (“FP8”); andidentifying the superset floating point format as a single floating point format, wherein the superset floating point format is an 9-bit floating point format (“FP9”) comprising a sign bit, exponent bits (e), and mantissa bits (m) and the one or more ssFPUs is 9-bit floating point unit.3. The method of claim 1 , further including converting the one or more inputs claim 1 , represented as a plurality of 8-bit floating point formats (“FP8”) into the superset floating point format prior to computation by the one or more ssFPUs claim 1 , wherein the superset floating point format is an 9-bit floating point format (“FP9”) and the one or more ssFPUs is 9-bit floating point ...

Data processing apparatus and method for converting a number between fixed-point and floating-point presentations

The present invention provides a data processing apparatus and method for converting a number between fixed-point and floating-point representations. More particularly, the data processing apparatus comprises a data processing unit operable to execute instructions, with the data processing unit being responsive to a format conversion instruction to apply a format conversion operation to a number to perform a conversion between the fixed-point representation of the number and the floating-point representation of the number. Furthermore, a control field is provided which is arranged to provide a programmable value specifying a decimal point location within the fixed-point representation of the number, and the data processing unit is operable to reference the control field and to control the formal conversion operation in accordance with the programmable value. This enables a single instruction to be used to directly convert between fixed-point formats (where the position of the decimal point ...

Decomposition of Decimal Floating Point Data

A decimal floating point finite number in a decimal floating point format is composed from the number in a different format. A decimal floating point format includes fields to hold information relating to the sign, exponent and significand of the decimal floating point finite number. Other decimal floating point data, including infinities and NaNs (not a number), are also composed. Decimal floating point data are also decomposed from the decimal floating point format to a different format. 1. A method of decomposing a decimal floating point datum , said method comprising:obtaining, by a processor of an emulated processing environment, a datum in a decimal floating point format;converting a significand of the datum in the decimal floating point format to the significand in a first format; andconverting an exponent of the datum in the decimal floating point format to the exponent in a second format, wherein in response to converting the significand and converting the exponent the datum is provided in a format other than the decimal floating point format.2. The method of claim 1 , wherein the first format comprises a packed decimal format and the second format comprises a binary integer format.3. The method of claim 1 , further comprising providing a sign for the datum.4. The method of claim 1 , wherein the converting the significand comprises:converting a number of least significant digits of the significand into a packed decimal format and storing the result in a first register pair;shifting the significand of a first floating point register pair a specified number of places;converting remaining digits of the significand into a packed decimal format and storing the result in a second register pair;combining a plurality of digits of the second register pair and contents of the first register pair to provide a decomposed value.5. The method of claim 4 , wherein the combining comprises concatenating the plurality of digits and the contents of the first register pair in ...

Systems and methods for efficient video-related data processing

Embodiments of the present invention comprise systems and methods for quantization or dequantization of data related to video wherein reduced bit depth intermediate calculations are enabled.

Method for Compressing Information

Provided is a method of compressing information. The method includes converting compression target information into a binary number, converting the binary number into a decimal number a, performing operation of a discriminant S = 1 + 1 + 8  a 2 and obtaining a result S of the discriminant in order to operate “b” and “k” of a one-to-one correspondence function of the decimal number, obtaining “b” and “k” as a result of the discriminant, determining whether or not compression target information can be compressed based on the obtained “b” and “k” and the decimal number, and obtaining “b”, “k”, and least significant digit of the compression target information converted into binary number by repeating the first to third steps until it is determined that the compression target information can be compressed, and outputting a compressed information by incorporating the obtained “b”, “k”, and least significant digit of the binary number, wherein the least significant digit of the compression target information converted into binary number is removed and input as the compression target information during the repeating.

System, methods, and media for providing in-memory non-relational databases

Providing a first control process that executes in a hardware processor; providing a first server process that executes in a hardware processor, that responds to write requests by storing objects in in-memory, non-relational data store, and that responds to read requests by providing objects from in-memory, non-relational data store, wherein the objects each have an object size; forming a plurality of persistent connections between the first control process and the first server process; using the first control process, pipelining, using a pipeline having a pipeline size, requests that include the read requests and the write requests over at least one of the plurality of persistent connections; using the first control process, adjusting the number of plurality of persistent connections and the pipeline size based on an average of the object sizes; and using the first control process, prioritizing requests by request type based on anticipated load from the requests.

Binary-shift operations such as for header compression in packet-based communications

In one embodiment of a header-compression method, a timestamp value is divided by a stride value using a plurality of binary-shift operations corresponding to a Taylor expansion series of the reciprocal stride value in a base of ½. When the division-logic circuitry of an arithmetic logic unit in the corresponding communication device is not designed to handle operands that can accommodate the length of the timestamp and/or stride values, the header-compression method can advantageously be used to improve the speed and efficiency of timestamp compression in communication devices.

Bar code generation method using color code, data compression method, and internet service method thereof

A method and a system for exchanging electronic information data, using a color code recognition device are provided. An electronic information data from the first computer is recognized in either in the ASCII code or in the binary digit format. The electronic information data is encoded by substituting each character of the ASCII codes or the binary digit recognized with a color value. The encoded data is transmitted to the second computer, and the transmitted data is decoded into the original electronic information data. The system for exchanging an electronic information data includes a storage device for storing the user identification database and the legend database, an encoding device for encoding the electronic information data, a decoding device for decoding the encoded data to restore the encoded data into its original format, and a network device for transmitting the data between a plurality of computers over the network.

Adaptive compression

Technology for adaptive compression is described (“the technology”). The technology may identify two or more partitions of a data stream; optionally pre-process data in each partition; create one or more evaluation functions to evaluate a suitability for compression of the data in each partition using a set of potential compression methods; process the created one or more evaluation functions; choose a subset of the set of potential compression methods for each segment at least partly by analyzing the evaluation functions; select a compression method for each segment based on a compression ratio of compressing the sequence of used compression methods and a compression rate of the data; compress the data in each partition using the selected compression method for the partition; compress a subsequence that indicates which compression method is used for each segment.

Systems and methods for identifying long matches of data in a compression history

Systems and methods of storing previously transmitted data and using it to reduce bandwidth usage and accelerate future communications are described. By using algorithms to identify long compression history matches, a network device may improve compression efficiently and speed. A network device may also use application specific parsing to improve the length and number of compression history matches. Further, by sharing compression histories, compression history indexes and caches across multiple devices, devices can utilize data previously transmitted to other devices to compress network traffic. Any combination of the systems and methods may be used to efficiently find long matches to stored data, synchronize the storage of previously sent data, and share previously sent data among one or more other devices.

Method and Apparatus to Perform Function Compression

A functional compression scheme involves determining a minimal entropy coloring of a characteristic graph of a random variable. Various scenarios are disclosed where the determination of the minimal entropy coloring of a characteristic graph is easy and tractable.

Interference resistant compressive sampling

Embodiments provide for dramatically improved interference resistance in advanced communications applications, where the frequency range can exceed 1 GHz. Such embodiments may be implemented using wideband technology to provide a wideband compressive sampling architecture that is capable of superior interference rejection through RF front end cancellation.

Method and device for iterative blind wideband sampling

Devices and methods are for iteratively sampling a wideband signal in order to recover one or more narrowband signals are disclosed. In one aspect, a wideband signal is received and the signal is sampled using a sampling device, which includes an amplifier with an initial gain level, to produce a plurality of sampled signals. A first set of narrowband signals may be recovered from the plurality of sampled signals. Then, the wideband signal is re-sampled to produce a second plurality of sampled signals. The re-sampling includes increasing the gain of the amplifier to a second level and suppressing a component of the wideband signal. A second set of narrowband signals may then be recovered from the second set of sampled signals.

Compression Analyzer

Techniques are described herein for automatically selecting the compression techniques to be used on tabular data. A compression analyzer gives users high-level control over the selection process without requiring the user to know details about the specific compression techniques that are available to the compression analyzer. Users are able to specify, for a given set of data, a “balance point” along the spectrum between “maximum performance” and “maximum compression”. The point thus selected is used by the compression analyzer in a variety of ways. For example, in one embodiment, the compression analyzer uses the user-specified balance point to determine which of the available compression techniques qualify as “candidate techniques” for the given set of data. The compression analyzer selects the compression technique to use on a set of data by actually testing the candidate compression techniques against samples from the set of data. After testing the candidate compression techniques against the samples, the resulting compression ratios are compared. The compression technique to use on the set of data is then selected based, in part, on the compression ratios achieved during the compression tests performed on the sample data.

Method and apparatus for compression and de-compression of spectral data

A method and apparatus for data compression, particularly applicable to spectral signals such as Fast Fourier Transforms of vibration data. The data is merged to remove redundant frequencies when recorded at multiple sample rates, thresholded with respect to a noise floor to remove even more redundant data, and then the positions of non-zero signal values, with respect to the noise floor, are recorded in a first dataword and the non-zero signal values themselves are all recorded concatenated to form a second dataword. The compressed data set consists of the first and second datawords, together with the value of the noise floor, maximum original amplitude and the broadband power. In the event of successive data sets having the same or similar locations for non-zero signal values a re-use flag may be set and the locations dataword discarded. Preferably the signal values are non-linearly quantized to further reduce the amount of data.

Inverted Order Encoding in Lossless Compresssion

A method of compressing an electronic file is provided. The method comprises reading a first electronic file in reverse order sequence from bottom to top, while reading the first file, identifying patterns in a content of the first file and while reading the first file, building a dictionary comprising a plurality of entries, each entry defining an association of a code to one of the patterns identified in the content of the first file. The method further comprises, while reading the first file, building a second electronic file that is a compressed version of the first file, wherein the second electronic file comprises a compressed content portion and a dictionary portion, wherein the compressed content portion comprises codes from the dictionary and wherein the dictionary portion comprises the dictionary.

Data encoding apparatus, data decoding apparatus and methods thereof

a In a data encoding apparatus, an acquisition unit acquires an input data including a plurality of arranged data. A conversion unit searches a data group including a first consecutive data and a second data adjacent to the first consecutive data, each data of the first consecutive data having a first value and the second data having a second value, and converts the data group into an intermediate data. A selection unit selects a selected encoding method among a plurality of encoding methods, at a time of encoding the intermediate data based on where a beginning of the data group corresponding to the intermediate data exists in the input data. An encoding unit can encode the intermediate data by using the selected encoding method.

Compression format for high bandwidth dictionary compression

Method, apparatus, and systems employing dictionary-based high-bandwidth lossless compression. A pair of dictionaries having entries that are synchronized and encoded to support compression and decompression operations are implemented via logic at a compressor and decompressor. The compressor/decompressor logic operatives in a cooperative manner, including implementing the same dictionary update schemes, resulting in the data in the respective dictionaries being synchronized. The dictionaries are also configured with replaceable entries, and replacement policies are implemented based on matching bytes of data within sets of data being transferred over the link. Various schemes are disclosed for entry replacement, as well as a delayed dictionary update technique. The techniques support line-speed compression and decompression using parallel operations resulting in substantially no latency overhead.

CO-COMPRESSION AND CO-DECOMPRESSION OF DATA VALUES

A first and a second data value are co-compressed by generating a sequence of symbols having a most significant symbol that is the most significant symbol of a compressed representation of the first data value and a least significant symbol that is the most significant symbol of a compressed representation of the second data value. The compressed representation of the first data value corresponds to at least a portion of the symbols of the sequence of symbols starting from the most significant symbol and extending towards the least significant symbol in a first reading direction. The compressed representation of the second data value also corresponds to at least a portion of the symbols of the sequence of symbols, however, starting from the least significant symbol and extending in an opposite reading direction towards the most significant symbol. 1. A data compressing method comprising:co-compressing at least a first data value and a second data value by generating a sequence of symbols having a most significant symbol that is the most significant symbol of a compressed representation of said first data value and a least significant symbol that is the most significant symbol of a compressed representation of said second data value, wherein said compressed representation of said first data value comprises at least a portion of said sequence of symbols starting from said most significant symbol of said sequence of symbols and extending in a first reading direction towards said least significant symbol and said compressed representation of said second data value comprises at least a portion of said sequence of symbols starting from said least significant symbol and extending in a second, opposite reading direction towards said most significant symbol of said sequence of symbols; andstoring said sequence of symbols as a compressed representation of at least said first data value and said second data value.2. The method according to claim 1 , wherein co-compressing ...

Methods and systems for compressing and comparing genomic data

Systems and methods are disclosed for compressing and comparing data such as genomic data. The disclosed systems and methods may include selecting a segment, creating a delta representation of the segment, the delta representation comprising a script, and storing the script. Furthermore, the disclosed systems and methods may include receiving a first script comprising a compressed version of a first segment and receiving a second script comprising a compressed version of a second segment. The disclosed systems and methods may further include comparing the first script to the second script and determining if the first segment matches the second segment based upon the comparison of the first script to the second script.

MODIFIED GABOR TRANSFORM WITH GAUSSIAN COMPRESSION AND BI-ORTHOGONAL DIRICHLET GAUSSIAN DECOMPRESSION

A signal processor for compressing signal data, including a function shapes generator for receiving as input time and frequency scale parameters, and for generating as output a plurality of shape parameters for a corresponding plurality of localized functions, wherein the shape parameters govern the centers and spreads of the localized functions, a matrix generator for receiving as input the plurality of shape parameters and a sequence of sampling times, and for generating as output a matrix whose elements are the values of the localized functions at the sampling times, a signal transformer for receiving as input an original signal and the matrix generated by the matrix generator, and for generating as output a transformed signal by applying the matrix to the original signal, and a signal compressor for receiving as input the transformed signal, and for generating as output a compressed representation of the transformed signal. 1. A signal processor for compressing signal data , comprising:a function shapes generator for receiving as input time and frequency scale parameters, and for generating as output a plurality of shape parameters for a corresponding plurality of localized functions, wherein the shape parameters govern the centers and spreads of the localized functions;a matrix generator, coupled with said function shapes generator, for receiving as input the plurality of shape parameters generated by said function shapes generator, and a sequence of sampling times, and for generating as output a matrix whose elements are the values of the localized functions at the sampling times, wherein each column of the matrix corresponds to one of the localized functions, and wherein each row of the matrix corresponds to one of the sampling times;a signal transformer, coupled with said matrix generator, for receiving as input an original signal, and the matrix generated by said matrix generator, and for generating as output a transformed signal by applying the matrix to ...

DIGITAL MEDIA USAGE IN RESPONSE TO IMPACT DATA

Aspects provide digital media content as a function of embedded environmental impact data. When a digital multimedia item is received, an embedded environmental impact value is decoded and the item is utilized as a function of the decoded embedded environmental impact value. Some aspects account for amounts and types of energy used in creating the item to calculate an environmental impact value for the item, and the calculated environmental impact value is embedded within digital data of the item. 1. A method for providing digital media content with an embedded environmental impact data value , the method comprising:a programmable device determining at least one of an amount of energy used in creating a first digital multimedia item comprising a digital multimedia file on a tangible machine-readable article, and a type of energy used in creating the first digital multimedia item; an amount of energy used in providing the first item to a user;', 'an amount of energy used to physically deliver the tangible machine-readable article to the user; and', 'an amount of energy used to recycle or dispose of the tangible machine-readable article;, 'the programmable device calculating a first environmental impact value for the first digital multimedia item as a function of the determined at least one of the amount of energy used in creating the first digital multimedia item and the type of energy used in creating the first digital multimedia item, by determining at least one ofthe programmable device encoding the determined first environmental impact value; andthe programmable device embedding the encoded first environmental impact value within digital data of the first digital multimedia item for decoding by a receiver of the first digital multimedia item.2. The method of claim 1 , further comprising:receiving a second digital multimedia item comprising a second embedded environmental impact value;decoding the second embedded environmental impact value;determining an amount of ...

LOW BIT RATE SIGNAL CODER AND DECODER

An improved mixed oscillator-and-external excitation model and methods for estimating the model parameters, for evaluating model quality, and for combining it with known in the art methods are disclosed. The improvement over existing oscillators allows the model to receive, as an input, all except the most recent point in the acquired data. Model stability is achieved through a process which includes restoring unavailable to the decoder data from the optimal model parameters and by using metrics to select a stable restored model output. The present invention is effective for very low bit-rate coding/compression and decoding/decompression of digital signals, including digitized speech, audio, and image data, and for analysis, detection, and classification of signals. Operations can be performed in real time, and parameterization can be achieved at a user-specified level of compression. 1. A non-transitory computer-readable storage medium with an executable program stored thereon , in which the program for analyzing and modeling a data signal instructs a processor to perform a method comprising the following steps of:pre-processing the data signal to generate a data frame, said pre-processing being capable of: sampling the data signal, transforming the data signal, smoothing the data signal, and decomposing the data signal into one or more components to be modeled as separate inputs;modeling the data frame as a combination of fixed length blocks of waveforms drawn from a reference buffer to determine model parameters, the reference buffer comprising all but the most recent data point of the current frame;evaluating qualities of a plurality of models to select an optimal model; andprocessing the computed model parameters, said processing being capable of storing, transmitting, analyzing, and classifying the data signal.2. The method as recited in claim 1 , further comprising the step of storing and making available for future computations at least one of: the waveform ...

METHODS AND COMPUTER PROGRAM PRODUCTS FOR COMPRESSION OF SEQUENCING DATA

A compression method includes measuring a waveform associated with a chemical event occurring on a sensor array, wherein the waveform comprises at least one region associated with expected measured values and at least one region associated with unpredictable measured values; applying a first compression process to the waveform, the first compression process including an averaging of one or more frames in one or more portions of the waveform; and applying a second compression process to the waveform, the second compression process including a truncating of data corresponding to a portion of the waveform that is not related to a nucleotide incorporation component of the waveform. 1. A compression method , comprising:measuring a waveform associated with a chemical event occurring on a sensor array, wherein the waveform comprises at least one region associated with expected measured values and at least one region associated with unpredictable measured values;applying a first compression process to the waveform, the first compression process including an averaging of one or more frames in one or more portions of the waveform; andapplying a second compression process to the waveform, the second compression process including a truncating of data corresponding to a portion of the waveform that is not related to a nucleotide incorporation component of the waveform.2. The method of claim 1 , wherein the truncating of data comprises determining claim 1 , for each of a plurality of sensors in the sensor array claim 1 , a cut-off time point for the waveform for that sensor defining a data range to be truncated.3. The method of claim 2 , wherein each cut-off time point is determined by mining a plurality of past analysis runs for a given sensor array geometry.4. The method of claim 2 , wherein each cut-off time point is determined prior to every run or during a calibration procedure.5. The method of claim 2 , wherein each cut-off time point is factory pre-determined for a given ...

Delta-Sigma D/A Converter

A delta-sigma D/A converter, by which a digital valued, input signal is convertible into a binary, clock signal time discrete, output signal. By forming an average value of the output signal over a number of clock signal cycles, an analog value of the input signal can be displayed. The delta-sigma D/A converter is embodied in such a manner that, in use, it provides the output signal by serial arrangement of signal patterns of a set of signal patterns, wherein the signal patterns of the set are, in each case, binary, clock signal time discrete and extend over a signal pattern cycles total of a plurality of clock cycles. At least two signal patterns of the set have mutually different signal pattern average values, which are formed over the respective signal pattern cycles total, and all signal patterns of the set have, in each case, essentially the same number, especially exactly the same number, of edges.

SIGNAL PROCESSOR, WINDOW PROVIDER, ENCODED MEDIA SIGNAL, METHOD FOR PROCESSING A SIGNAL AND METHOD FOR PROVIDING A WINDOW

A signal processor for providing a processed version of an input signal in dependence on the input signal includes a windower configured to window a portion of the input signal, or of a pre-processed version thereof, in dependence on a signal processing window described by signal processing window values for a plurality of window value index values, in order to obtain the processed version of the input signal. The signal processor also includes a window provider for providing the signal processing window values for a plurality of window value index values in dependence on one or more window shape parameters. 5. A non-transitory computer readable medium including a computer program for performing the method according to when the computer program runs on a computer.6. A non-transitory computer readable medium including a computer program for performing the method according to when the computer program runs on a computer. This application is a continuation of copending International Application No. PCT/EP2011/053484, filed Mar. 8, 2011, which is incorporated herein by reference in its entirety, and additionally claims priority from U.S. Application No. 61/312,775, filed Mar. 11, 2010, and European Application EP 10175142.8, filed Sep. 2, 2010, which are all incorporated herein by reference in their entirety.Embodiments according to the invention are related to a signal processor for providing a processed version of an input signal in dependence on the input signal, to a window provider for providing signal processing window values, to an encoded media signal, to a method for processing a signal and to a method for providing signal processing window values.An embodiment according to the invention is related to an apparatus for encoding or decoding an audio or video signal using variable window functions. Another embodiment according to the invention is related to a method for encoding or decoding an audio or video signal using variable window functions.Embodiments ...

SYSTEM AND METHOD FOR MULTI-RESIDUE MULTIVARIATE DATA COMPRESSION

A method for encoding data comprising generating a table having N moduli, where N is a positive integer equal to two or more, where each of a plurality of integers has a unique set of residue values associated with the moduli. Storing or transmitting the first data field value of a sequence of L data fields values, where L is an integer equal to or greater than 2. Storing or transmitting a set of K residue values, where K Подробнее

Compression and decompression system, compression apparatus, decompression apparatus and compression and decompression method

In a compression and decompression system that performs data compression and decompression, the decompression of compressed data is performed in a way that a compression apparatus generates a byte code string as compressed data, and a decompression apparatus executes the byte code string. The byte code includes an 8-byte-unit copy instruction and direct data processing instruction, and the compression apparatus determines whether to use the 8-byte-unit copy instruction and direct data processing instruction or a byte-unit copy instruction and direct data processing instruction upon decompression, and generates the byte code.

BYPASS BINS FOR REFERENCE INDEX CODING IN VIDEO CODING

In an example, aspects of this disclosure relate to a method for decoding a reference index syntax element in a video decoding process that includes decoding at least one bin of a reference index value with a context coding mode of a context-adaptive binary arithmetic coding (CABAC) process. The method also includes decoding, when the reference index value comprises more bins than the at least one bin coded with the context coded mode, at least another bin of the reference index value with a bypass coding mode of the CABAC process, and binarizing the reference index value. 1. A method for encoding a reference index syntax element in a video encoding process , the method comprising:binarizing a reference index value;encoding at least one bin of the binarized reference index value with a context coding mode of a context-adaptive binary arithmetic coding (CABAC) process; andencoding, when the binarized reference index value comprises more bins than the at least one bin coded with the context coded mode, at least another bin of the binarized reference index value with a bypass coding mode of the CABAC process.2. The method of claim 1 , encoding a first bin (bin0) of the binarized reference index value with a first context (ctx0),', 'encoding a second bin (bin1) of the binarized reference index with a second context (ctx1), and, 'wherein encoding at least one bin of the binarized reference index value with the context coding mode comprises 'encoding a third bin (bin2) and all remaining bins after the third bin (bin2) with the bypass coding mode.', 'wherein encoding at least another bin of the binarized reference index value with the bypass coding mode comprises3. The method of claim 2 , wherein binarizing the reference index value comprises binarizing the reference index value using a combined truncated unary and exponential Golomb code.4. The method of claim 1 , wherein encoding the at least another bin of the binarized reference index value with the bypass coding mode ...

METHOD FOR COMPRESSING DIGITAL VALUES OF IMAGE, AUDIO AND/OR VIDEO FILES

A method for differential compression of a sequence of digital values adapted to avoid error propagation during restoration of the values. 1. A method for compressing a sequence (S) of initial digital values (Va , . . . , Vp , Vq , Vr , . . . , Vz) into a compressed sequence (SC) of compressed values (Va , VCp , VCq , VCr , . . . , VCz) , intending to restore these values into a decompressed sequence (S2 ,S4) of decompressed values (Va , . . . , VDp , VDq , VDr , . . . , VDz) , wherein for a first initial value (Va) of the sequence , the compressed value (VCa=Va) of the first initial value is equal to said first initial value (Va) and the decompressed value (VDa=VCa=Va) of said compressed value (VCa) of the first initial value (Va) is equal to said first initial value (Va) , then in that , for each current initial value (Vq) , the following steps are carried out:calculating the difference (Dq=Vq−VDp) between the current initial value (Vq) and the decompressed value (VDp) of the initial value immediately preceding the current value (Vq); then,calculating a compressed value (VCq=f[Dq]) of said difference (Dq) using a complementary compression function (f); then,{'sup': '1', 'calculating the decompressed value (VDq=f[VCq]+VDp) corresponding to said current initial value (Vq); then,'}applying the three preceding steps to the immediately following value (Vr) if there is one; and,constituting the compressed sequence (SC) of the compressed values (Va, . . . , VCp, VCq, VCr, . . . , VCz), each corresponding to a respective initial value (Va, . . . , Vp, Vq, Vr, . . . , Vz).2. The method according to claim 1 , in which the compressed value is rounded to the nearest whole number.3. The method according to claim 1 , further comprising keeping claim 1 , on the compressed value claim 1 , the sign of the difference (if DqO then VCq>O).4. The method according to claim 1 , in which the complementary compression function (f) is an nth root function claim 1 , ...

EFFICIENT TECHNIQUES FOR ALIGNED FIXED-LENGTH COMPRESSION

Systems and methods for performing compression of data. A data buffer is separated into equal-sized segments of data. A frequency count is performed to determine how often each segment of data appears in the data buffer. Frequently occurring segments are encoded with unique compression codes, while all other infrequently occurring segments are encoded with a common compression code. The compressed data buffer includes the compression codes, which are all of the same bit-length, and the uncompressed segments. The compression codes and the uncompressed segments are stored in the compressed data buffer in the order in which the corresponding segments appear in the original data buffer. 1. A method comprising:partitioning a data buffer into a plurality of segments;determining how often each distinct segment value occurs within the data buffer; assigning a unique compression code for each value which occurs at least a given number of times; and', 'assigning a common compression code to represent values which occur fewer than the given number of times;, 'generating a fixed-length compression code for each segment of the plurality of segments, wherein said generating comprisesstoring the generated compression codes and values which occur fewer than the given number of times consecutively in a compressed data buffer.2. The method as recited in claim 1 , wherein the compressed data buffer further comprises a table including an identification of at least those values occurring more than the given number of times and corresponding compression codes.3. The method as recited in claim 1 , wherein the compressed data buffer further comprises a header claim 1 , wherein the header comprises:a compression algorithm type indicator;an uncompressed length of the data buffer; anda table identifying at least those values occurring more than the given number of times.4. The method as recited in claim 1 , wherein generated compression codes are placed in the compressed data buffer in an ...

COMPRESSION MATCH ENUMERATION

In embodiments of compression match enumeration, a trie structure can be searched to enumerate data sequence matches after the trie structure is generated. The trie structure can be generated by overwriting a suffix array of data sequences stored in a memory buffer, where the suffix array includes an array of buffer positions sorted alphabetically by a data string that begins a data sequence at a respective buffer position. The trie structure can then be traversed to determine the data sequence matches from nodes in the trie structure. The trie structure can also be updated while being traversed to bypass multiple enumerations of the data sequence matches that have the same match length. 1. A method , comprising:searching a trie structure to enumerate data sequence matches after the trie structure is generated by overwriting a suffix array of data sequences stored in a memory buffer, the suffix array including an array of buffer positions sorted alphabetically by a data string that begins a data sequence at a respective buffer position; andtraversing the trie structure to determine the data sequence matches from nodes in the trie structure.2. A method as recited in claim 1 , further comprising determining the data sequence matches that have the least compressed data based on a number of matching data bytes in a data sequence match represented by a node of the trie structure.3. A method as recited in claim 1 , further comprising updating the trie structure while said traversing the trie structure to bypass multiple enumerations of the data sequence matches that have the same match length.4. A method as recited in claim 1 , wherein the trie structure is generated by incrementally updating the trie structure from consecutive suffixes of the suffix array as the trie structure is generated in-place of the suffix array in the memory buffer.5. A method as recited in claim 1 , wherein the trie structure comprises nodes that each represent one or more suffixes of the suffix ...

COMPRESSING AND DECOMPRESSING SIGNAL DATA

The present invention relates to data compression and decompression, and particularly relates to a method and an apparatus for compressing and decompressing signal data. In an embodiment of the present invention, there is disclosed a method for compressing signal data, comprising: obtaining signal data; determining block lengths of a plurality of data blocks into which the signal data are divided, and determining exponents of the data blocks; forming a mantissa sequence of the data block by using the exponent of the data block to compress signal data contained in the data block; and forming a compressed data block by using the block length, the exponent and the mantissa sequence of the data block. By constructing a variable-length data block adapted to dynamic characteristics of signal data, the method for compressing signal data of the present invention increases the compression ratio of signal data. 1. A method for compressing signal data , the method comprising:obtaining signal data;determining block lengths of a plurality of data blocks into which the signal data are divided and determining exponents of the data blocks;forming a mantissa sequence of a data block by using the exponent of the data block to compress signal data contained in the data block; andforming a compressed data block from the block length, the exponent and the mantissa sequence of the data block.3. The method according to claim 1 , wherein the forming a mantissa sequence of the data block by using the exponent of the data block to compress signal data contained in the data block comprises: compressing signal data contained in the data block from the next bit of a sign bit claim 1 , wherein the number of bits by which the signal data is compressed are determined by the exponent of the data block claim 1 , and the compressed data block forms a mantissa sequence of the data block.4. The method according to claim 3 , wherein the forming a compressed data block from the block length claim 3 , the ...

COMPRESSING AND DECOMPRESSING SIGNAL DATA

The present invention relates to data compression and decompression, and particularly relates to a method and an apparatus for compressing and decompressing signal data. In an embodiment of the present invention, there is disclosed a method for compressing signal data, comprising: obtaining signal data; determining block lengths of a plurality of data blocks into which the signal data are divided, and determining exponents of the data blocks; forming a mantissa sequence of the data block by using the exponent of the data block to compress signal data contained in the data block; and forming a compressed data block by using the block length, the exponent and the mantissa sequence of the data block. By constructing a variable-length data block adapted to dynamic characteristics of signal data, the method for compressing signal data of the present invention increases the compression ratio of signal data. 1. An apparatus for compressing signal data , comprising:an obtainment module configured to obtain signal data;a block length and exponent determination module configured to determine block lengths of a plurality of data blocks into which the signal data are divided and determine exponents of the data blocks;a mantissa sequence formation module configured to form a mantissa sequence of the data block by using the exponent of the data block to compress signal data contained in the data block; anda compressed data block formation module configured to form a compressed data block from the block length, the exponent and the mantissa sequence of the data block.2. The apparatus according to claim 1 , wherein the block length and exponent determination module comprises:an initial data block division module configured to initially divide the signal data into a plurality of data blocks; {'br': None, 'i': E', 'B', 'e', 'd, 'sub': d', {'sub2': 'j'}, '∈B', 'j, '()=min(())'}, 'an exponent calculation module configured to calculate an exponent of a data block according to exponents ...

METHODS FOR ENCODING AND DECODING DATA

The transmission of broadcast data, such as financial data and news feeds, is accelerated over a communication channel using data compression and decompression to provide secure transmission and transparent multiplication of communication bandwidth, as well as reduce the latency. Broadcast data may include packets having fields. Encoders associated with particular fields may be selected to compress those particular fields. 1. A method for compressing data , wherein one or more types of encoding are applied to one or more data blocks containing financial data , with packet independent data encoding being performed , at least in part , based on data residing within data fields in the one or more data blocks , the method comprising:recognizing at least one of any characteristic, attribute, or parameter of the data residing within the data field in the one or more data blocks;selecting an encoder associated with the at least one of the characteristic, attribute, or parameter of the data field;encoding data in the data field utilizing the selected encoder; andassociating a recognizable data token with the encoded data, the recognizable data token being capable of identifying the selected encoder.2. The method of claim 1 , wherein the recognizable data token is associated with one or more encoded data blocks.3. The method of claim 1 , wherein the recognizable data token is associated with a payload portion of a data packet.4. The method of claim 1 , wherein the recognizable data token is appended to an encoded data block.5. The method of claim 1 , wherein the selecting the encoder associated with the recognized any characteristic claim 1 , attribute claim 1 , or parameter of the data residing within the data field is based on a packet type of a data packet.6. The method of claim 1 , further comprising:utilizing numerical difference encoding to encode relative numeric values.7. The method of claim 1 , further comprising:utilizing difference encoding based on an absolute ...

DATA TRANSLATION SYSTEM AND METHOD

A data translation system and method. This invention provides a reverse approach to implement a M bit input to N bit output cumulative/monotonic transfer function, (where M>N) by a (2**N)×M bit memory instead of the conventional (2**M)×N bit memory. The invention offers substantial circuit size savings without compromising on transfer function resolution and is independent of transfer function mapping algorithms. The M bit memory content of the reverse LUT contains input video group information for each output level and the (2**N) addresses of the reverse LUT represent the corresponding transfer function output levels. This data to address representation of the input to output relationship is exactly opposite to the conventional address to data format. Search and compare methods are employed to locate the input video group that the incoming video belongs to and the associated address of the reverse LUT represents the output. 131-. (canceled)32. A system for translating an M-bit input to an N-bit output , where M>N , using a cumulative-monotonic input-output transfer function , comprising:{'sup': 'N', 'memory means for storing 2M-bit digital words defining breakpoints of the input-output transfer function and'}means for comparing; said M-bit input to said breakpoints and providing an output in response thereto.33. The invention of wherein the means for comparing is a hybrid parallel pipeline.34. The invention of wherein said pipeline has jbits determined at each stage and wherein the sum of all of the j's equals N and more than one bit of output is determined in at least one stage.35. The invention of wherein said means for comparing includes comparators.36. The invention of wherein the comparators are disposed in parallel.37. The invention of wherein said means for comparing is comprised of an N-stage pipeline.38. The invention of wherein said pipeline has one bit determined at each stage.39. The invention of wherein said means for comparing includes iterating ...

METHODS FOR ENCODING AND DECODING DATA

The transmission of broadcast data, such as financial data and news feeds, is accelerated over a communication channel using data compression and decompression to provide secure transmission and transparent multiplication of communication bandwidth, as well as reduce the latency. Broadcast data may include packets having fields. Encoders associated with particular fields may be selected to compress those particular fields. 1. A method of decoding , on a packet independent basis using a data decoding engine , one or more encoded messages residing in data blocks of a data packet , the encoded messages including financial information transmitted in sequence from a source , internal or external to the decoding engine , wherein at least one of a plurality of decompression techniques are applied to the one or more encoded messages , the method comprising:identifying one or more recognizable data tokens in an encoded message, the one or more recognizable data tokens being associated with the data packet and including any recognizable data token indicative of any characteristic, attribute, or parameter of data residing in the one or more data blocks, and wherein the encoders were selected based on analyses of the content of the data block;applying one or more decoders to a data block from among the one or more data blocks associated with the encoded message based on the one or more recognizable data tokens and a description table, the description table including the characteristics, attributes, or parameters of the data residing in the one or more data blocks and their associated decoders;decoding the data block with an associated decoder when the recognizable data token indicates that the data block was encoded utilizing content dependent data compression; anddecoding the data block with an associated decoder when the recognizable data token indicates that the data block was encoded utilizing content independent data compression.2. The method of claim 1 , wherein the ...

ACCELEROMETER DATA COMPRESSION

A method of compressing data output from an acceleration measurement means configured to be transported, carried or worn by a user is provided. Acceleration values indicative of the movement of the user are measured at a first frequency and values representative of the measured acceleration values are generated at a second frequency, which is lower than the first frequency. The step of generating comprises: defining a plurality of time windows, each time window containing a plurality of measured acceleration values; and applying a transformation to the measured acceleration values within each time window to generate a plurality of transformed values. For each time window, storing at least one of said plurality of transformed values and/or one or more parameters associated therewith. 1. A method of compressing data output from an acceleration measurement means configured to be transported , carried or worn by a user , comprising the steps of:measuring acceleration values indicative of the movement of the user at a first frequency; defining a plurality of time windows, each time window containing a plurality of measured acceleration values; and', 'applying a transformation to the measured acceleration values within each time window to generate a plurality of transformed values; and,, 'generating values representative of the measured acceleration values at a second frequency, which is lower than the first frequency, said step of generating comprisingfor each time window, storing at least one of said plurality of transformed values and/or one or more parameters associated therewith.2. A method as claimed in claim 1 , wherein the acceleration values are measured using one or more accelerometers.3. A method as claimed in claim 1 , wherein the first frequency is between 1 Hz and 20 Hz.4. A method as claimed in claim 1 , wherein the acceleration values are stored in a data buffer.5. A method as claimed in claim 1 , further comprising determining an absolute acceleration ...

Data compression utilizing longest common subsequence template

In response to receipt of an input string, an attempt is made to identify, in a template store, a closely matching template for use as a compression template. In response to identification of a closely matching template that can be used as a compression template, the input string is compressed into a compressed string by reference to a longest common subsequence compression template. Compressing the input string includes encoding, in a compressed string, an identifier of the compression template, encoding substrings of the input string not having commonality with the compression template of at least a predetermined length as literals, and encoding substrings of the input string having commonality with the compression template of at least the predetermined length as a jump distance without reference to a base location in the compression template. The compressed string is then output.

DATA COMPRESSION UTILIZING LONGEST COMMON SUBSEQUENCE TEMPLATE

In response to receipt of an input string, an attempt is made to identify, in a template store, a closely matching template for use as a compression template. In response to identification of a closely matching template that can be used as a compression template, the input string is compressed into a compressed string by reference to a longest common subsequence compression template. Compressing the input string includes encoding, in a compressed string, an identifier of the compression template, encoding substrings of the input string not having commonality with the compression template of at least a predetermined length as literals, and encoding substrings of the input string having commonality with the compression template of at least the predetermined length as a jump distance without reference to a base location in the compression template. The compressed string is then output. 1. A method of data processing system in a data processing system , comprising:in response to receipt of an input string, an integrated circuit attempting to identify, in a template store, a closely matching template for use as a compression template; encoding, in a compressed string, an identifier of the compression template;', 'encoding, in the compressed string, first substrings of the input string not having commonality with the compression template of at least a predetermined length as literals;', 'in response to identifying first substrings, encoding, in the compressed string, jump lengths in the compression template until a next match between the input string and the compression template;', 'encoding, in the compressed string, second substrings of the input string having commonality with the compression template of at least the predetermined length as a template reference without reference to a base location in the compression template; and, 'in response to identification of the closely matching template as the compression template, the integrated circuit compressing the input ...

DATA COMPRESSION USING DICTIONARY ENCODING

Embodiments relate to data compression using dictionary encoding. An aspect includes subdividing a table of uncompressed data into a first block and a second block of complete rows. Another aspect includes determining information about a frequency of occurrence of different values for each column of the first block. Another aspect includes selecting a row of the first block to be removed out of the first block using frequency of occurrence-information. Another aspect includes removing the a row out of the first block to form an updated first block and determining information about a frequency of occurrence of different values for each column of the updated first block. Another aspect includes deriving a dictionary containing code-words for encoding the values of the updated first block. Another aspect includes encoding the values of the updated first block based on the code-words. Another aspect includes adding the removed row to the second block. 1. A computer-implemented method for data compression using dictionary encoding , the method comprising:providing at least one table of uncompressed data with values arranged in columns and rows;subdividing the table into at least a first block and a second block of complete rows;selecting the first block of rows to be processed to perform the data compression;determining information about a frequency of occurrence of different values for each column of the first block;evaluating and selecting at least one row of the first block to be removed out of the first block using frequency of occurrence-information to reduce code-word length;removing the at least one row out of the first block to form an updated first block;determining information about a frequency of occurrence of different values for each column of the updated first block;deriving at least one dictionary containing code-words for encoding the values of the updated first block;encoding the values of the updated first block based on the code-words of the at least ...

DATA COMPRESSION

An exemplary encoder device includes a data storage comprising an indication of a coordinate system and a set of symbols, each having at least one coordinate position in the coordinate system. A processor associated with the data storage is configured to encode at least one symbol within information comprising a plurality of symbols using the coordinate position of the at least one symbol relative to the coordinate position of another symbol within the information. 1. (canceled)2. An encoder device , comprising:a data storage comprising an indication of a coordinate system and a set of symbols, each having at least one coordinate position in the coordinate system; and determine a primary coordinate position of a selected one of the symbols of the information, and', 'determine a sequence of movements within the coordinate system corresponding to a sequence of the plurality of symbols of the information, wherein at least a beginning of the sequence of movements is determined based on the primary coordinate position., 'a processor associated with the data storage, the processor being configured to encode at least one symbol within information comprising a plurality of symbols using the coordinate position of the at least one symbol relative to the coordinate position of another symbol within the information, wherein the processor is configured to'}3. The device of claim 2 , wherein the sequence of movements comprises:a first movement from the primary coordinate position of the selected one of the symbols to the coordinate position of a sequentially adjacent one of the symbols of the information; anda plurality of next movements to the coordinate position of a next sequentially adjacent one of the symbols from the coordinate position of the previous adjacent one of the symbols.4. An encoder device claim 2 , comprising:a data storage comprising an indication of a coordinate system and a set of symbols, each having at least one coordinate position in the coordinate system ...

COMPRESSION DEVICE, COMPRESSION METHOD, DECOMPRESSION DEVICE, DECOMPRESSION METHOD, INFORMATION PROCESSING SYSTEM, AND RECORDING MEDIUM

A compression device includes a processor configured to execute a process. The process includes: storing, in a storage, a first compressed code in association with a first element, the first compressed code corresponding to a combination of a first element and a first delimiter, the first element being one of a plurality of elements constituting input data, the first delimiter being one of delimiters delimiting the plurality of elements and succeeding the first element in the input data; acquiring, from the storage, the first compressed code in response to reading a sequence of the first element and the first delimiter from the input data; and writing the first compressed code into a storage area that stores therein compressed data of the input data. 1. A compression device comprising:a processor configured to execute a process including:storing, in a storage, a first compressed code in association with a first element, the first compressed code corresponding to a combination of a first element and a first delimiter, the first element being one of a plurality of elements constituting input data, the first delimiter being one of delimiters delimiting the plurality of elements and succeeding the first element in the input data;acquiring, from the storage, the first compressed code in response to reading a sequence of the first element and the first delimiter from the input data; andwriting the first compressed code into a storage area that stores therein compressed data of the input data.2. The compression device according to claim 1 , whereinthe storing includes storing, in the storage, a first control code corresponding to a combination of a control symbol indicating an instruction to delete the first delimiter and the first delimiter,the acquiring includes acquiring, from the storage, the first compressed code and the first control code in response to reading the first element from the input data, andthe writing includes writing the first control code at a position ...

Apparatuses, methods and systems for efficient ad-hoc querying of distributed data

The APPARATUSES, METHODS AND SYSTEMS FOR EFFICIENT AD-HOC QUERYING OF DISTRIBUTED DATA (“RTC”) provides a platform that, in various embodiments, is configurable to provide fast ad-hoc querying against large volumes of data. In one embodiment, the RTC is configurable to select a subset of fields from raw data in association with a domain and compact the corresponding data. Such packed records may be distributed to one or more worker nodes, which maintain the records and associated indexes. A master server facilitates query processing across the worker nodes.

EFFICIENT STORAGE AND RETRIEVAL OF TEXTUAL DATA

A method of and system of efficient storage of data entries containing textual data is disclosed. The method may include accessing a plurality of data entries in a dataset, arranging the plurality of data entries in the dataset in a lexical order, placing a predetermined number of the plurality of data entries in each of a plurality of subblocks, performing data compression on the plurality of data entries in each of the plurality of subblocks to reduce redundancy in the plurality of data entries and create compressed data entries, placing one or more subblocks in each of a plurality of page blocks, and storing each of the plurality of page blocks in a storage device to provide efficient searching and improved functionality for the dataset. 1. A data processing system comprising:a processor; and accessing a plurality of data entries in a dataset;', 'arranging the plurality of data entries in the dataset in a lexical order;', 'placing a predetermined number of the plurality of data entries in each of a plurality of subblocks;', 'performing data compression on the plurality of data entries in each of the plurality of subblocks to reduce redundancy in the plurality of data entries and create compressed data entries;', 'placing one or more subblocks in each of a plurality of page blocks; and', 'storing each of the plurality of page blocks in a storage device to provide efficient searching and improved functionality for the dataset., 'a memory in communication with the processor, the memory comprising executable instructions that, when executed by the processor, cause the data processing system to perform functions of2. The data processing system of claim 1 , wherein the plurality of subblocks are in a lexical order.3. The data processing system of claim 2 , wherein the plurality of page blocks are in a lexical order.4. The data processing system of claim 1 , wherein performing data compression on the plurality of data entries in each of the plurality of subblocks to ...

Data compression apparatus, data compression method, and memory system including the data compression apparatus

Provided are data compression method, data compression apparatus, and memory system. The data compression method includes receiving input data and generating a hash key for the input data, searching a hash table with the generated hash key, and if it is determined that the input data is a hash hit, compressing the input data using the hash table; and searching a cache memory with the input data, and if it is determined that the input data is a cache hit, compressing the input data using the cache memory.

TECHNIQUES TO PROVIDE A MULTI-LEVEL MEMORY ARCHITECTURE VIA INTERCONNECTS

Various embodiments are generally directed to an apparatus, method and other techniques to enable memory interfaces to communicate read request, write requests, and data via an interconnect. Embodiments, include processing write requests to write data into memory coupled via an interconnect and processing read requests to read data from memory coupled via an interconnect. In embodiments, the data may be compressed data based on a compression mechanism and communicated in a fabric packet including a compression mechanism indicator, the compressed data, and an address, the compression mechanism indicator to indicate which compression mechanism is applied to the data. 1. An apparatus , comprising: generate compressed data based on a compression mechanism to be applied to data, and', 'generate a fabric packet including a compression mechanism indicator, the compressed data, and an address to write the data in at least one of a second level memory and a third level memory, the compression mechanism indicator to indicate which compression mechanism is applied to the data., 'a memory interface coupled to a memory controller, the memory interface to2. The apparatus of claim 1 , the compression mechanism indicator to indicate the compression mechanism is one of a byte pattern repeat claim 1 , a word pattern repeat claim 1 , a double word (Dword) pattern repeat claim 1 , a quad word (Qword) pattern repeat claim 1 , and a compression alogrithm.3. The apparatus of claim 1 , the second level memory comprising one or more volatile memory devices and the third level memory comprising one or more byte addressable write-in-place non-volatile memory devices.4. The apparatus of claim 1 , the memory interface to send the compressed data to the at one of the second level memory and the third level memory via an interconnect comprising one of an optical interconnect and an electrical interconnect.5. The apparatus of claim 1 , the memory interface to:receive another fabric packet ...

GRAPH-BASED COMPRESSION OF DATA RECORDS

In general, embodiments of the present invention provide systems, methods and computer readable media for data record compression using graph-based techniques. 130-. (canceled)31. A computer-implemented method for generating a compressed list of impression data records , the method comprising:receiving, by a processor, a set of impression data records describing a sequence of consumer behavior instances collected during a time window, wherein the set of impression data records is associated with a particular consumer;generating, by the processor, a directed link graph representing the consumer behavior instances; andgenerating, by the processor, the compressed list of the impression data records based at least in part on properties of the directed link graph.32. The method of claim 31 , wherein each impression data record of the set of impression data records comprises a plurality of data components.33. The method of claim 31 , wherein the graph nodes respectively represent the consumer behavior instances and each of the graph edges connecting a pair of the nodes represents a hyperlink between the nodes.34. The method of claim 31 , further comprising:determining a first component of the set of impression data records to be an index component;generating a sorted list of the set of impression data records by ordering the set of impression data records using a respective value of the index component within each impression data record;identifying a set of unique index component values within the sorted list of the set of impression data records; andgenerating an ordered list of the set of unique index component values using the sorted list of the set of impression data records.35. The method of claim 34 , further comprising generating an index value list by replacing each impression data record of the set of impression data records with its respective position in the ordered list of the set of unique index component values.36. The method of claim 35 , further comprising ...

DATA COMPRESSION DEVICE AND METHOD

A data compression device including a processor to perform a procedure comprising: obtaining data of a predetermined number (Z) of digits in a time series; and performing a compression process on the data. The data is obtained by encoding a vibration state of a measurement target. The compression process includes: deleting upper digits when the upper digits do not include significant information; and adding a unique code to a top of the upper digits when the upper digits include significant information. A digit number (X) of the upper digits is smaller than the predetermined number (Z). 1. A data compression device including a processor to perform a procedure , the procedure comprising:obtaining data of a predetermined number (Z) of digits in a time series, the data being obtained by encoding a vibration state of a measurement target; andperforming a compression process on the data,wherein the compression process includes:deleting upper digits when the upper digits do not include significant information, a digit number (X) of the upper digits being smaller than the predetermined number (Z); andadding a unique code to a top of the upper digits when the upper digits include significant information.2. The data compression device according to claim 1 , wherein the processor performs the compression process on a difference between items of the data adjacent in the time series.3. The data compression device according to claim 1 , wherein the processor sets the digit number (X) based on a variance value of the data in the time series.4. The data compression device according to claim 2 , wherein the processor sets the digit number (X) based on a variance value of the data in the time series.5. The data compression device according to claim 3 , wherein the processor estimates the variance value based on a number of times that the difference between the items of the data adjacent in the time series is less than a predetermined value.6. The data compression device according to ...

METHOD FOR DATA COMPRESSION

A data compression system includes: (a) a data compression module that receives a sequence of input vectors and that provides a sequence of compressed vectors; (b) a data decompression module that receives the compressed vectors to provide a sequence of output vectors; and (c) a parameter update module that receives the sequence of input vectors and the sequence of output vectors, and which learns the data compression module and data decompression module based on evaluating a loss function of the input vectors, the output vectors, and the parameters controlling the compression module and the decompression module. Each input vector and its corresponding output vector may represent digitized time-domain signals (e.g., speech, audio or video signals) over a predetermined time period. The loss function may be evaluated for each of a sequence of predetermined time periods. 1. A data compression system , comprising:a data compression module that receives a sequence of input vectors and that provides a sequence of compressed vectors, wherein the total number of data bits representing the input vectors is greater than the number of data bits representing the compressed vectors and wherein the data compression module implements a computational model characterized by a first set of parameters;a data decompression module that receives the compressed vectors to provide a sequence of output vectors, wherein the total number of data bits representing the output vectors is greater than the total number of data bits representing the compressed vectors and wherein the data decompression module implements a computational model characterized by a second set of parameters; anda parameter update module that receives the sequence of input vectors and the sequence of output vectors, wherein the parameter update module updates values of the first set of parameters and the second set of parameters based on a loss function of the input vectors, the output vectors, the first set of parameters ...

Graph-Based Compression Of Data Records

In general, embodiments of the present invention provide systems, methods and computer readable media for data record compression using graph-based techniques. 130-. (canceled)31. A computer-implemented method for generating a compressed list of impression data records , the method comprising:receiving, by a processor, a set of impression data records describing a sequence of consumer behavior instances collected during a time window, wherein the set of impression data records is associated with a particular consumer;generating, by the processor, a directed link graph representing the consumer behavior instances;generating, by the processor, the compressed list of the impression data records based at least in part on properties of the directed link graph;determining, by the processor, a first component of the set of impression data records to be an index component;generating, by the processor, a sorted list of the set of impression data records by ordering the set of impression data records using a respective value of the index component within each impression data record;identifying, by the processor, a set of unique index component values within the sorted list of the set of impression data records; andgenerating, by the processor, an ordered list of the set of unique index component values using the sorted list of the set of impression data records.32. The method of claim 31 , wherein each impression data record of the set of impression data records comprises a plurality of data components.33. The method of claim 31 , wherein the graph nodes respectively represent the consumer behavior instances and each of the graph edges connecting a pair of the nodes represents a hyperlink between the nodes.34. The method of claim 31 , further comprising generating claim 31 , by the processor claim 31 , an index value list by replacing each impression data record of the set of impression data records with its respective position in the ordered list of the set of unique index ...

DYNAMIC DATA COMPRESSION SELECTION

Aspects of dynamic data compression selection are presented. In an example method, as uncompressed data chunks of a data stream are compressed, at least one performance factor affecting selection of one of multiple compression algorithms for the uncompressed data chunks of the data stream may be determined. Each of the multiple compression algorithms may facilitate a different expected compression ratio. One of the multiple compression algorithms may be selected separately for each uncompressed data chunk of the data stream based on the at least one performance factor. Each uncompressed data chunk may be compressed using the selected one of the multiple compression algorithms for the uncompressed data chunk. 1. A method for data compression , the method comprising:receiving uncompressed source data at a data compression system from a data storage device, the uncompressed source data comprising a set of data chunks; and identifying a compression algorithm to use to compress the data chunk based on at least one performance factor, the compression algorithm being of a plurality of compression algorithms;', 'compressing the data chunk using the compression algorithm, wherein different compression algorithms are used to compress at least two of the set of data chunks; and', 'routing the compressed data chunk to the data target., 'streaming source data to a data target by, for each data chunk of at least some of the set of data chunks2. The method of claim 1 , wherein each of at least two of the different compression ratios are associated with different expected compression ratios.3. The method of claim 1 , wherein at least one of the different compression ratios includes a lossless data compression algorithm.4. The method of claim 1 , wherein streaming the source data includes transmitting a data file claim 1 , data file folder claim 1 , data container or data structure over a period of time.5. The method of claim 1 , wherein the identification of the compression ...

GRAPH-BASED COMPRESSION OF DATA RECORDS

In general, embodiments of the present invention provide systems, methods and computer readable media for data record compression using graph-based techniques. 1. A computer-implemented method for generating a compressed list of impression data records , the method comprising:receiving a consumer impression data records set describing consumer behavior instances collected during a time window, wherein each impression data record of the impression data records set is a compound data record including data components;generating a directed link graph representing the consumer behavior instances, the directed link graph having graph nodes and graph edges, wherein the graph nodes respectively represent the consumer behavior instances and each of the graph edges connecting a pair of the nodes represents a hyperlink between the pair of nodes; andgenerating, by a processor, the compressed list of the consumer impression data records based at least in part on properties of the directed link graph.2. The method of claim 1 , wherein generating the compressed list of the consumer impression data records comprises:determining a first component of a consumer impression data record to be an index component;generating a sorted list of the consumer impression data records by ordering the consumer impression data records using a respective value of the index component in each consumer impression data record;identifying a unique index component values set within the sorted list of consumer impression data records;generating an ordered list of the unique index component values set using the sorted list of the consumer impression data records;determining whether there are duplicates of at least one unique index component value within the sorted list of consumer impression data records;in an instance in which there are duplicates of at least one unique second component value within the sorted list of consumer impression data records,associating, for each duplicated unique index component ...

FAST DATA COMPRESSION FOR SYSTEMS WITH LARGE DYNAMIC RANGES OR OTHER SYSTEMS

A method includes defining a lookup table having multiple compressed data values and, for each compressed data value, a range of one or more raw data values associated with the compressed data value. The method also includes compressing raw data using the lookup table. Each raw data value is associated with one of the compressed data values in the lookup table based on a distance ratio involving (i) a distance of the raw data value from a minimum value and (ii) a distance of the raw data value from a maximum value. 1. A method comprising:defining a lookup table comprising multiple compressed data values and, for each compressed data value, a range of one or more raw data values associated with the compressed data value; andcompressing raw data using the lookup table;wherein each raw data value is associated with one of the compressed data values in the lookup table based on a distance ratio involving (i) a distance of the raw data value from a minimum value and (ii) a distance of the raw data value from a maximum value.2. The method of claim 1 , wherein claim 1 , for each raw data value claim 1 , the distance ratio involves (i) a distance of the raw data value from a global minimum value identifying a smallest possible value of the raw data and (ii) a distance of the raw data value from a global maximum value identifying a largest possible value of the raw data.3. The method of claim 2 , wherein:the raw data being compressed spans a range from a local minimum value to a local maximum value, the range not extending from the global minimum value to the global maximum value; andthe method further comprises remapping at least one of the raw data and the lookup table so that an increased number of compressed values in the lookup table span the range from the local minimum value to the local maximum value.4. The method of claim 1 , wherein each raw data value is associated with one of the compressed data values in the lookup table based on a logarithm of the distance ...

SIGNAL TRANSMISSION

An approach for bandwidth-efficient transmission of signals is provided, e.g. as an apparatus, as a method or as a computer program. According to an example embodiment, the approach comprising storing, in one or more data buffers, information descriptive of data transmitted from respective one or more data sources to a destination, identifying reference data in said one or more data buffers on basis of data to be transmitted by an apparatus and on basis of information stored in said one or more data buffers, determining residual data descriptive of the data to be transmitted by the apparatus in relation to the said reference data, and transmitting the residual data to the destination. 180-. (canceled)81. An apparatus comprising at least one processor and at least one memory including computer program code for one or more programs , the at least one memory and the computer program code configured to , with the at least one processor , cause the apparatus to perform at least the following:store, in one or more data buffers, information descriptive of data transmitted from respective one or more data sources to a destination,identify reference data in said one or more data buffers on basis of data to be transmitted by the apparatus and on basis of information stored in said one or more data buffers,determine residual data descriptive of the data to be transmitted by the apparatus in relation to the said reference data, andtransmit the residual data to the destination.82. An apparatus according to claim 81 , wherein identifying reference data comprises identifying a reference source and to identify the reference data in the data buffer corresponding to the reference source.83. An apparatus according to claim 82 , wherein identifying reference data comprises identifying the reference source on basis of data transmitted from the apparatus and on basis of information stored in said one or more data buffers.84. An apparatus according to claim 82 , wherein identifying ...

SYSTEMS AND METHODS FOR PERFORMING MEMORY COMPRESSION

Systems, apparatuses, and methods for efficiently moving data for storage and processing are described. In various embodiments, a compression unit within a processor includes multiple hardware lanes, selects two or more input words to compress, and for assigns them to two or more of the multiple hardware lanes. As each assigned input word is processed, each word is compared to an entry of a plurality of entries of a table. If it is determined that each of the assigned input words indexes the same entry of the table, the hardware lane with the oldest input word generates a single read request for the table entry and the hardware lane with the youngest input word generates a single write request for updating the table entry upon completing compression. Each hardware lane generates a compressed packet based on its assigned input word. 1. An apparatus comprising:a first buffer configured to store a plurality of input words;a table comprising a plurality of entries; and assign a first group of two or more input words from the plurality of input words to the first plurality of hardware lanes;', 'determine at least two of the first group of two or more input words correspond to a same entry of the table;', determine whether to update the table with any of the input words of the first group; and', 'assign a second group of input words from the plurality of input words to the first plurality of hardware lanes, wherein the second group is different from the first group;, 'prior to determining dependencies between input words of the first group, 'generate a compression packet that stores data corresponding to the plurality of input words., 'compression circuitry comprising a first plurality of hardware lanes, wherein in response to receiving an indication of a compression instruction, the compression circuitry is configured to2. The apparatus as recited in claim 1 , wherein the compression circuitry is further configured to:determine dependencies between input words of the ...

TYPE CONVERSION USING FLOATING-POINT UNIT

Techniques are disclosed relating to type conversion using a floating-point unit. In one embodiment, to convert a floating-point value to a normalized integer format, a floating-point unit is configured to perform an operation to generate a result having a significant portion and an exponent portion, where the operation includes multiplying the floating-point value by a constant. In one embodiment, the apparatus is further configured to add a value to the exponent portion of the result, and set a rounding mode to round to nearest. The constant may be a greatest value less than one that can be represented using the particular number of unsigned bits. The value added to the initial exponent may be equal to the number of unsigned bits of the normalized integer format. The apparatus may perform this conversion in response to a pack instruction. 1. A method , comprising: performing an operation, by a floating-point unit, to generate a result having a significand portion and an exponent portion, wherein the operation includes multiplying the floating-point number by a number equal to a greatest value less than one that can be represented using the particular number of unsigned bits;', 'changing the exponent portion of the result by adding the particular number to the exponent portion; and', 'rounding the result, by the floating-point unit, after the adding., 'converting a floating-point number to a normalized integer output format having a particular number of unsigned bits, including2. The method of claim 1 , wherein the rounding includes rounding using a round to nearest even mode.3. The method of claim 1 ,wherein the integer output format is a 16-bit signed integer output format;wherein the particular number of unsigned bits is 15; andwherein the number equal to a greatest value less than one that can be represented using the particular number of unsigned bits is represented in binary by 0.111111111111111.4. The method of claim 1 , further comprising:saturating an ...

PARALLELIZED ROUNDING FOR DECIMAL FLOATING POINT TO BINARY CODED DECIMAL CONVERSION

A computer-implemented method includes: receiving, using a processor, a decimal floating point number; and using a floating point unit within the processor to convert the decimal floating point number into a binary coded decimal number, wherein the floating point unit starts a conversion loop subsequent to a rounding loop starting, wherein the rounding loop and the conversion loop run in parallel once started. 1. A computer-implemented method comprising:receiving, using a processor, a decimal floating point number; andusing a floating point unit within the processor to convert the decimal floating point number into a binary coded decimal number, wherein the floating point unit starts a conversion loop subsequent to a rounding loop starting, wherein the rounding loop and the conversion loop run in parallel once started.2. The method according to claim 1 , wherein the floating point unit finishes the rounding loop and outputs a result of the rounding loop.3. The method according to claim 1 , wherein the floating point unit uses a pipeline having a first plurality of stages as part of the conversion loop.4. The method according to claim 3 , wherein the pipeline further includes a first plurality of shift registers as part of the conversion loop.5. The method according to claim 4 , wherein the pipeline further includes a second plurality of stages as part of the rounding loop.6. The method according to claim 5 , wherein the pipeline further includes a second plurality of shift registers as part of the rounding loop.7. The method according to claim 1 , wherein the conversion loop uses a conversion register which outputs a result which is added to the rounding loop.8. A system comprising:a receiving module configured to receive a decimal floating point number; anda floating point unit configured to convert the decimal floating point number into a binary coded decimal number, wherein the floating point unit starts a conversion loop subsequent to a rounding loop starting, ...

PARALLEL ROUNDING FOR CONVERSION FROM BINARY FLOATING POINT TO BINARY CODED DECIMAL

Embodiments of the invention are directed to a computer-implemented method of for parallel conversion to binary coded decimal format. The method includes receiving, by a floating point unit (FPU), a value in binary floating point (BFP) format. The BFP value includes an integer part and a fractional part. The FPU converts the BFP value to a binary coded decimal (BCD) value. In parallel to converting the BFP value to a BCD value, the FPU performs a rounding operation on the BFP value. The FPU receives the rounding information and operates on the BCD value accordingly. 1. A computer-implemented method comprising:receiving, by a floating point unit (FPU), a binary floating point (BFP) value, wherein the BFP value comprises a significand, and wherein the significand comprises an integer part and a fractional part;converting the integer part of the BFP value to a binary coded decimal (BCD) value;determining, in parallel to the converting, whether to round the BFP value based on the fractional part; andperforming a rounding operation on the BCD value based on the rounding determination.2. The computer-implemented method of further comprising:prior to converting, in response to the BFP value being a negative number, modifying the BFP value so that it is a positive BFP number; andsubsequent to the converting and in response to modifying the BFP value so that it is a positive number, modifying the BCD value so that it is a negative number.3. The computer-implemented method of claim 1 , wherein the BFP value further comprises a base and an exponent.4. The computer-implemented method of further comprising performing a bit shifting operation based on the exponent of the BFP value.5. The computer-implemented method of claim 1 , wherein converting the integer part of the BFP value to a BCD value comprises:converting the BFP value to an intermediate value; andconverting the intermediate value to a BCD value.6. The computer-implemented method of claim 1 , wherein the BFP value ...

System and method to concurrently match any type, combination and/or range of characters defined by a regular expression on a hardware implementation using a single detection logic

Method to simultaneously detect all possible types, combinations and/or ranges of characters of a Regular Expression search pattern in a search space by re-encoding the search pattern and the search space in single set bit representations, employing a single detection logic. The search space n-bit representation is re-encoded into a single set bit representation of 2n bits while single/multiple search pattern(s) n-bit representation(s) is (are) re-encoded and combined into a sole single/multiple set bit representation of 2n bits. The hardware implementation of the method detects the presence of any character existent from the search pattern in the search space.

MEMORY COMPRESSION OPERABLE FOR NON-CONTIGUOUS WRITE/READ ADDRESSES

Disclosed embodiments include a system having a first memory for storing a plurality of data quantities, each data quantity consisting of a first number of bits, and a second memory for storing a plurality of compressed data quantities, each compressed data quantity consisting of a second number of bits that is less than the first number of bits. The system includes circuitry for reading data quantities from the first memory and for writing compressed data quantities, corresponding to respective read data quantities, to non-sequential addresses in the second memory. The circuitry for reading data quantities from the first memory is for reading along a read orientation selected from one of row-orientation or column-orientation from the first memory, and the circuitry for writing compressed data quantities in the second memory is for writing along a write orientation in the second memory that differs from the read orientation. 1. A digital data storage and retrieval system , comprising:a first memory for storing a plurality of data quantities, wherein each data quantity, in the plurality of data quantities, consists of a first number of bits;a second memory for storing a plurality of compressed data quantities, wherein each compressed data quantity, in the plurality of compressed data quantities, consists of a second number of bits that is less than the first number of bits;circuitry for reading data quantities from the first memory; andcircuitry for writing compressed data quantities, corresponding to respective read data quantities, to non-sequential addresses in the second memory;wherein the circuitry for reading data quantities from the first memory is for reading along a read orientation selected from one of row-orientation or column-orientation from the first memory, and wherein the circuitry for writing compressed data quantities in the second memory is for writing along a write orientation in the second memory that differs from the read orientation.2. The ...

THREE-DIMENSIONAL DATA ENCODING METHOD, THREE-DIMENSIONAL DATA DECODING METHOD, THREE-DIMENSIONAL DATA ENCODING DEVICE, AND THREE-DIMENSIONAL DATA DECODING DEVICE

A three-dimensional data encoding method includes: encoding, for each of three-dimensional points included in point cloud data, geometry information and one or more items of attribute information to generate a bitstream. In the encoding, a normal vector of each of the three-dimensional points is encoded as an item of attribute information included in the one or more items of attribute information of the three-dimensional point. 1. A three-dimensional data encoding method , comprising:encoding, for each of three-dimensional points included in point cloud data, geometry information and one or more items of attribute information to generate a bitstream, whereinin the encoding, a normal vector of each of the three-dimensional points is encoded as an item of attribute information included in the one or more items of attribute information of the three-dimensional point.2. The three-dimensional data encoding method according to claim 1 , wherein transforming the normal vector represented by a floating point to an integer; and', 'encoding the integer., 'the encoding includes3. The three-dimensional data encoding method according to claim 1 , whereinthe bitstream includes control information shared by the geometry information and the one or more items of attribute information, information indicating that the item of attribute information included in the one or more items of attribute information indicates the normal vector; or', 'information indicating that the normal vector is data including three elements for each of the three-dimensional points., 'the control information includes at least one of'}4. A three-dimensional data decoding method claim 1 , comprising:obtaining a bitstream, the bitstream being generated by encoding, for each of three-dimensional points included in point cloud data, geometry information and one or more items of attribute information, the bitstream including, for each of the three-dimensional points, a normal vector encoded as an item of attribute ...

CONVERSION OF UNORM INTEGER VALUES TO FLOATING-POINT VALUES IN LOW POWER

Methods and apparatus relating to conversion of an unsigned normalized (unorm) integer values to floating-point (float) values in low power are described. In an embodiment, conversion logic converts a unorm integer value to a floating-point value based on detection of whether the unorm integer matches one of three cases, wherein the unorm integer value comprises n bits. Memory stores a count value corresponding to n−1 bits of the unorm integer value after detection of a leading 1 in the unorm integer value. The three cases include: a first case with all zeros, a second case with all ones, and a third case with a combination of one or more zeros and one or more ones. Other embodiments are also disclosed and claimed. 1. An apparatus comprising:conversion logic to convert an unsigned normalized (unorm) integer value to a floating-point value based on detection of whether the unorm integer matches one of three cases, wherein the unorm integer value comprises n bits; andmemory to store a count value corresponding to n−1 bits of the unorm integer value after detection of a leading 1 in the unorm integer value,wherein the three cases comprise: a first case with all zeros, a second case with all ones, and a third case with a combination of one or more zeros and one or more ones.2. The apparatus of claim 1 , wherein for the first case and the second case claim 1 , an output of the conversion logic is a constant.3. The apparatus of claim 1 , wherein for the third case claim 1 , leading one detection claim 1 , shift claim 1 , and rounding operations on the unorm integer value are limited to n bits.4. The apparatus of claim 3 , wherein the shift operation is a left shift operation by up to n bits.5. The apparatus of claim 3 , comprising logic to replicate data shifted by the shift operation.6. The apparatus of claim 1 , wherein the conversion logic is to extract a least significant bit from a mantissa of the unorm integer value.7. The apparatus of claim 6 , wherein the ...

Universal floating-point instruction set architecture for computing directly with decimal character sequences and binary formats in any combination

A universal floating-point Instruction Set Architecture (ISA) implemented entirely in hardware. Using a single instruction, the universal floating-point ISA has the ability, in hardware, to compute directly with dual decimal character sequences up to IEEE 754-2008 “H=20” in length, without first having to explicitly perform a conversion-to-binary-format process in software before computing with these human-readable floating-point or integer representations. The ISA does not employ opcodes, but rather pushes and pulls “gobs” of data without the encumbering opcode fetch, decode, and execute bottleneck. Instead, the ISA employs stand-alone, memory-mapped operators, complete with their own pipeline that is completely decoupled from the processor's primary push-pull pipeline. The ISA employs special three-port, 1024-bit wide SRAMS; a special dual asymmetric system stack; memory-mapped stand-alone hardware operators with private result buffers having simultaneously readable side-A and side-B read ports; and dual hardware H=20 convertFromDecimalCharacter conversion operators.

SYSTEM FOR COMPRESSING FLOATING POINT DATA

A processor comprises a first memory to store data elements that are encoded according to a floating point format including a sign field, an exponent field, and a significand field; and a compression engine comprising circuitry, the compression engine to generate a compressed data block that is to include a tag type per data element, wherein responsive to a determination that a first data element includes a value in its exponent field that does not match a value of any entry in a dictionary, a first tag type and an uncompressed value of the data element are included in the compressed data block; and responsive to a determination that a second data element includes a value in its exponent field that matches a value of a first entry in the dictionary, a second tag type and a compressed value of the data element are included in the compressed data block. 1. A processor comprising:a first memory to store a plurality of data elements that are encoded according to a floating point format including a sign field, an exponent field, and a significand field; and responsive to a determination that a first data element includes a value in its exponent field that does not match a value of any entry in a dictionary, a first tag type and an uncompressed value of the first data element are included in the compressed data block; and', 'responsive to a determination that a second data element includes a value in its exponent field that matches a value of a first entry in the dictionary, a second tag type and a compressed value of the second data element are included in the compressed data block., 'a compression engine comprising circuitry, the compression engine to generate a compressed data block that is to include a tag type per data element of the plurality of data elements, wherein2. The processor of claim 1 , wherein responsive to a determination that a third data element includes a value in its exponent field that matches a value of a second entry in the dictionary claim 1 , a ...

METHOD AND APPARATUS FOR DATA PROCESSING OPERATION

A system includes a fixed-point accumulator for storing numbers in an anchored fixed-point number format, a data interface arranged to receive a plurality of weight values and a plurality of data values represented in a floating-point number format, and logic circuitry. The logic circuitry is configured to: determine an anchor value indicative of a value of a lowest significant bit of the anchored fixed-point number format; convert at least a portion of the plurality of data values to the anchored fixed-point number format; perform MAC operations between the converted at least portion and respective weight values, using fixed-point arithmetic, to generate an accumulation value in the anchored fixed-point number format; and determine an output element of a later of a neural network in dependence on the accumulation value. 1. A system arranged to perform a sequence of multiply-accumulate (MAC) operations between a plurality of weight values and a plurality of data values of a data array , to determine an output element of a layer of a neural network , the system comprising:a fixed-point accumulator for storing numbers in an anchored fixed-point number format;a data interface arranged to receive the plurality of weight values and the plurality of data values, wherein the data values are represented in a floating-point number format; and determine an anchor value indicative of a value of a lowest significant bit of the anchored fixed-point number format;', 'convert at least a portion of the plurality of data values from the floating-point number format to the anchored fixed-point number format;', 'perform MAC operations between the converted at least portion of the plurality of data values and respective weight values of the plurality of weight values, using fixed-point arithmetic, to generate an accumulation value in the fixed-point accumulator in the anchored fixed-point number format; and', 'determine the output element in dependence on the accumulation value., ' ...

Accumulation-Based Data Object Processing

A system includes first logic to obtain a fixed constituent data object and a variable constituent data object for each data object of a set of objects, second logic to analyze the fixed constituent data objects to allocate each fixed constituent data object to one of a plurality of fixed groups, third logic to analyze the variable constituent data objects to allocate each variable constituent data object to one of a plurality of variable groups, fourth logic to determine a net magnitude for each fixed group and for each variable group, fifth logic to determine a historical component for each variable constituent data object, sixth logic to determine a net historical magnitude for each variable group based on the historical components, and seventh logic to define at least one representative data object to represent the fixed and variable constituent data objects in each pair of fixed and variable groups having a matching common set of properties, the at least one representative data object maintaining the net magnitudes and maintaining the net historical magnitude. 1. A system for processing a set of data objects , each data object of the set of data objects specifying a fixed constituent data object and a variable constituent data object , the system comprising:a processor;a memory coupled with the processor;first logic stored in the memory and executable by the processor to cause the processor to obtain the fixed constituent data object and the variable constituent data object for each data object of the set of objects;second logic stored in the memory and executable by the processor to cause the processor to analyze the fixed constituent data objects to allocate each fixed constituent data object to one of a plurality of fixed groups, the fixed constituent data objects in each fixed group of the plurality of fixed groups having a common set of properties;third logic stored in the memory and executable by the processor to cause the processor to analyze the ...

METHODS AND SYSTEMS FOR CONVERTING WEIGHTS OF A DEEP NEURAL NETWORK FROM A FIRST NUMBER FORMAT TO A SECOND NUMBER FORMAT

Methods and system for converting a plurality of weights of a filter of a Deep Neural Network (DNN) in a first number format to a second number format, the second number format having less precision than the first number format, to enable the DNN to be implemented in hardware logic. The method comprising: determining, for each of the plurality of weights, a quantisation error associated with quantising that weight to the second number format in accordance with a first quantisation method; determining a total quantisation error for the plurality of weights based on the quantisation errors for the plurality of weights; identifying a subset of the plurality of weights to be quantised to the second number format in accordance with a second quantisation method based on the total quantisation error for the plurality of weights; and generating a set of quantised weights representing the plurality of weights in the second number format, the quantised weight for each weight in the subset of the plurality of weights based on quantising that weight to the second number format in accordance with the second quantisation method and the quantised weight for each of the remaining weights of the plurality of weights based on quantising that weight to the second number format in accordance with the first quantisation method. 1. A computer-implemented method of converting a plurality of weights of a Deep Neural Network (DNN) from a set of first number formats to a set of second number formats to enable the DNN to be implemented in hardware logic , the plurality of weights of the DNN being divisible into a plurality of filters , each filter being an input to a layer of the DNN and associated with a first number format of the set of first number formats and a second number format of the set of second number formats , the second number format associated with a filter having less precision than the first number format associated with the filter , the method comprising , for each filter of ...

DATA STORAGE BASED ON ENCODED DNA SEQUENCES

Devices, methods, and systems for encoding data as DNA are provided. An encoder device can include an encoder engine configured to encode a data file having a bit sequence encoding data and further configured to generate a virtual DNA (VDNA) sequence of virtual nucleotide bases (Vnb) that reversibly encodes the bit sequence of the data file, divide the VDNA sequence into a plurality of VDNA fragments, associate each VDNA fragment with an archive library sequence (Arc_SEQ), and generate a read instruction (READ) sequence of differences between each VDNA fragment and each associated Arc_SEQ including sufficient instruction to facilitate regeneration of each VDNA fragment from each associated Arc_SEQ. A codeword sequence (Code_SEQ) is additionally generated for each VDNA fragment comprising a codename identifying the associated Arc_SEQ, the READ sequence associated with the VDNA fragment, and an index sequence (Idx_SEQ) including an index mapping of the VDNA fragment in the VDNA sequence. 1. An encoder device , comprising: generate a virtual deoxyribonucleic acid (VDNA) sequence of virtual nucleotide bases (Vnb) that reversibly encodes the bit sequence of the data file;', 'divide the VDNA sequence into a plurality of VDNA fragments;', 'associate each VDNA fragment with an archive library sequence (Arc_SEQ);', 'generate a read instruction (READ) sequence of differences between each VDNA fragment and each associated Arc_SEQ including sufficient instruction to facilitate regeneration of each VDNA fragment from each associated Arc_SEQ; and', a codename identifying the associated Arc_SEQ;', 'the READ sequence associated with the VDNA fragment; and', 'an index sequence (Idx_SEQ) including an index mapping of the VDNA fragment in the VDNA sequence., 'generate a codeword sequence (Code_SEQ) for each VDNA fragment comprising], 'an encoder engine configured to receive instructions to encode a data file having a bit sequence of binary bits encoding data, the encoder engine ...

EFFICIENT TRANSFER OF IQ SAMPLE DATA

In mobile communications networks, requirements on signal distortion may be fulfilled at a lower bit rate, or alternatively quantization noise be reduced for a given bit rate, by including fractional exponent bits in a block floating point format. One or more fractional exponent bits may apply to all samples in the block. Alternatively, fractional bits may apply to sub-blocks within the block. The optimal number of fractional bits depends on the number of samples in the block. 1. A method for fronthaul data transfer performed in a first network node of a wireless communications network comprising:sending IQ sample pairs or beam coefficients in a block floating point format to a second network node of the wireless communications network via a fronthaul link, the blocks having a size number (Ns) equal to or greater than 4, and including a number of integer exponent bits (Ne) and a number of fractional exponent bits (Nf), where Nf being at least 1.2. The method according to claim 1 , wherein the Ne integer exponent bits of a block are applicable to all samples in the block and two or more disjunct sets of fractional exponent bits out of a full set of Nf fractional exponent bits are applicable to corresponding disjunct subsets of samples out of a full set of Ns samples in the block.3. The method according to claim 2 , wherein a block of size Ns=24 comprises 12 IQ sample pairs claim 2 , the number of integer exponent bits Ne is 4 and the number of fractional exponent bits Nf is 4 claim 2 , each fractional exponent bit being applicable to 3 IQ sample pairs.4. The method according to claim 2 , wherein a block of size Ns=24 comprises 12 IQ sample pairs claim 2 , the number of integer exponent bits Ne is 2 and the number of fractional exponent bits Nf is 6 claim 2 , each fractional exponent bit being applicable to 2 IQ sample pairs.5. The method according to claim 1 , wherein the block size Ns is 8 or greater and Nf is at least 2.6. The method according to claim 1 , wherein ...

Method and apparatus for compressing a data set using incremental deltas and a variable reference value

In one embodiment, an apparatus comprises a processor to receive a plurality of values of a data set, the data set comprising a first value, a second value, and a third value; calculate and store a first delta corresponding to the first value, wherein the first delta is equal to the difference between the first value and the second value; and calculate and store a second delta corresponding to the second value, wherein the second delta is equal to the difference between the second value and the third value.

CHECKSUM ADDER

Embodiments relate to a hardware circuit that is operable as a fixed point adder and a checksum adder. An aspect includes a driving of a multifunction compression tree disposed on a circuit path based on a control bit to execute one of first and second schemes of vector input addition and a driving of a multifunction adder disposed on the circuit path based on the control bit to perform the one of the first and second schemes of vector input addition. 1a computer readable storage medium having program instructions embodied therewith, wherein the computer readable storage medium is not a signal, the program instructions readable by a processor to cause the processor to perform a method comprising:{'sub': 0', 'n−1, 'driving, by the processor, a multifunction compression tree disposed on a circuit path and comprising n bit circuits arranged in a sequence, wherein n≧3, a lowest numbered bit circuit is bitand a highest numbered bit circuit is bitbased on a control bit to execute one of fixed point vector input addition and checksum vector input addition, wherein the driving, by the processor, of the multifunction compression tree comprisesreceiving at a zeroth level compressor of each bit circuit an input of three vector inputs and outputting from the zeroth level compressor a zeroth level carryout and a zeroth level sum;receiving at a first level compressor of each bit circuit an input of two vector inputs and a first level carry-in derived from a zeroth level carry-out of an adjacent, higher numbered bit circuit and outputting from the first level compressor a first level carry-out and a first level sum;receiving at a second level compressor of each bit circuit an input of the zeroth level and first level sums and a second level carry-in derived from a first level carry-out of the adjacent, higher numbered bit circuit and outputting from the second level compressor a carry term and a sum term;{'sub': 0', 'n−1, 'ANDing the control bit along with the zeroth level carry- ...

Systems, media and methods for virtualized data compression

Systems, media, and methods for virtualized data compression are provided. For example, a stream of numbered sequences may be generated by transforming an input stream into a sequence of samples each having a predetermined amplitude and a predetermined length. An error correction protocol may be applied through an analysis of a number of bits. A compression routine may be implemented by choosing a number of threads and determining compression depth. The stream analyze may be analyzed utilizing independent component analysis. A specified compression routine may be executed. An output file size may be determined.

Compression device, compression method, decompression device, decompression method, and computer-readable recording medium

A non-transitory computer-readable recording medium has stored therein a compression program that causes a computer to execute a process. The process includes: extracting a numeric character string from input data, the numeric character string indicating a numerical value and including therein at least one numeral; and converting the numeric character string to a compressed code utilizing a dictionary that associates numeric character strings and compressed codes.

Method and System for Compression of Radar Signals

A radar system is provided that includes a compression component configured to compress blocks of range values to generate compressed blocks of range values, and a radar data memory configured to store compressed blocks of range values generated by the compression component. 1. A radar system comprising:a compression component configured to compress blocks of range values to generate compressed blocks of range values; anda radar data memory configured to store compressed blocks of range values generated by the compression component.2. The radar system of claim 1 , in which the compression component is configured to compress a block of range values using block floating point (BFP) compression.3. The radar system of claim 1 , in which the compression component is configured to compress a block of range values using order k exponential Golomb (EG) compression.4. The radar system of claim 3 , in which the compression component is configured to select an optimal Golomb parameter value for compressing a block of range values from a plurality of specified candidate Golomb parameter values.5. The radar system of claim 3 , in which the compression component is configured determine a scale factor for a block of range values and to use the scale factor to truncate each range value in the block of range values claim 3 , the scale factor determined as a minimum number of bits to be dropped from each range value in order to generate a corresponding compressed block of range values of a size less than or equal to a specified size.6. The radar system of claim 3 , in which the compression component is configured to estimate an encoded size of the block of range values based on a number of leading consecutive zero bits and a number of consecutive one bits following the leading consecutive zero bits in each range value.7. The radar system of claim 1 , in which the compression component is configured to provide at least two types compression.8. The radar system of claim 7 , in which ...

COMPUTER AND METHOD OF CREATING GRAPH DATA

Disclosed is a computer configured to create graph data having a vertex corresponding to a single index, an edge that links a pair of the vertices having a correlation, and an edge weight as a value of the element from the correlation matrix data having correlation values between a plurality of indices as elements, in which the correlation matrix data is acquired from the storage unit, elements of a spanning tree formed by linking vertices corresponding to indices included in the acquired correlation matrix data and an element having a value equal to or greater than a predetermined threshold value are detected, and the graph data is created on the basis of the detected elements. 1. A computer provided with a processor and a memory connected to the processor and configured to execute a processing using correlation matrix data having correlation values between a plurality of indices as elements , the computer comprising:a graph processing unit configured to create first graph data having a list structure provided with a vertex corresponding to a single index, an edge that links a pair of the vertices having a correlation, and an edge weight as a value of the element from the correlation matrix data acquired from a storage unit,the graph processing unit havinga control factor calculation section configured to calculate a maximum number of edges that can be contained in the first graph data in order to complete the processing using the correlation matrix data within a predetermined time,a spanning tree creating section configured to convert the correlation matrix data into second graph data having a list structure and create third graph data as a spanning tree provided with all vertices and a part of the edges of the second graph data, anda graph data creating section configured to create the first graph data on the basis of the second and third graph data using the maximum number of edges.2. The computer according to claim 1 , wherein the spanning tree is a maximum ...

Measuring device and electrical controller

A measuring device includes a delta-sigma modulator configured to take an analog signal as a basis for generating a bit stream, and an evaluation unit that receives the bit stream from the delta-sigma modulator and evaluates the received bit stream. The measuring device has a single data transmission line, wherein the delta-sigma modulator is configured to transmit the bit stream to the evaluation unit via the single data transmission line using a transmit clock, and wherein the evaluation unit is configured to reconstruct the transmit clock and/or a phase of bits within the bit stream from the received bit stream and to extract the bits from the received bit stream based on the reconstructed transmit clock and/or based on the reconstructed phase.

SELECTIVE DATA COMPRESSION FOR IN-MEMORY DATABASES

Techniques are provided for maintaining data persistently in one format, but making that data available to a database server in more than one format. For example, one of the formats in which the data is made available for query processing is based on the on-disk format, while another of the formats in which the data is made available for query processing is independent of the on-disk format. Data that is in the format that is independent of the disk format may be maintained exclusively in volatile memory to reduce the overhead associated with keeping the data in sync with the on-disk format copies of the data. Selection of data to be maintained in the volatile memory may be based on various factors. Once selected the data may also be compressed to save space in the volatile memory. The compression level may depend on one or more factors that are evaluated for the selected data. The factors for the selection and compression level of data may be periodically evaluated, and based on the evaluation, the selected data may be removed from the volatile memory or its compression level changed accordingly. 1. A method comprising:maintaining, by a database server, a database on persistent storage; a first data portion, that has been designated as eligible for mirroring within volatile memory; and', 'a second data portion that has been also designated as eligible for mirroring within the volatile memory;, 'wherein the database includesstoring first eligibility data that corresponds to the first data portion, wherein the first eligibility data reflects an estimated benefit of having the first data portion mirrored in the volatile memory;storing second eligibility data that corresponds to the second data portion, wherein the second eligibility data reflects an estimated benefit of having the second data portion mirrored in the volatile memory;wherein the first eligibility data and the second eligibility data are determined by the database server based on one or more factors; ...

Systems and Methods for Weighted Quantization

Generally, the present disclosure is directed to systems and methods of quantizing a database with respect to a novel loss or quantization error function which applies a weight to an error measurement of quantized elements respectively corresponding to the datapoints in the database. The weight is determined based on the magnitude of an inner product between the respective datapoints and a query compared therewith. In contrast to previous work, embodiments of the proposed loss function are responsive to the expected magnitude of an inner product between the respective datapoints and a query compared therewith and can prioritize error reduction for higher-ranked pairings of the query and the datapoints. Thus, the systems and methods of the present disclosure provide solutions to some of the problems with traditional quantization approaches, which regard all error as equally impactful. 1. A computer-implemented method of quantizing a dataset , the method comprising:obtaining, by one or more computing devices, a dataset containing a plurality of data elements; andcompressing, by the one or more computing devices, the dataset into a quantized dataset that has a reduced memory size, wherein compressing, by the one or more computing devices, the data into the quantized dataset comprises determining, by the one or more computing devices, a plurality of quantized elements of the quantized dataset that respectively correspond to the plurality of data elements of the dataset;wherein each of the plurality of quantized elements has a respective quantization error; andwherein the respective quantization error for each quantized element is weighted by a respective weight value having a weight magnitude that is positively correlated with a magnitude of an inner product between the corresponding data element for such quantized element and a query.2. The computer-implemented method of claim 1 , wherein the respective quantization error for each quantized element comprises:a parallel ...

Constrained System Endec

Various embodiments of the present invention provide apparatuses and methods for encoding and decoding data for constrained systems with reduced or eliminated need for hardware and time intensive arithmetic operations such as multiplication and division. 1. A method of generating an encoder comprising:generating a first directed graph characterizing a constraint set for a constrained system;calculating a first approximate eigenvector for the first directed graph;calculating a second approximate eigenvector as an approximation of the first approximate eigenvector;performing a first state splitting operation on the first directed graph using the second approximate eigenvector to yield a second directed graph with a third approximate eigenvector;performing a second state splitting operation on the second directed graph with the third approximate eigenvector to yield a third directed graph with a fourth approximate eigenvector; andgenerating the encoder based on the third directed graph.2. The method of claim 1 , wherein a connectivity matrix for the first directed graph multiplied by the first approximate eigenvector is at least equal to the first approximate eigenvector multiplied by a power of 2.3. The method of claim 1 , wherein a number of ones in a binary representation of the second approximate eigenvector does not exceed a maximum number of states into which the first directed graph is split in the first state splitting operation for all states in the second directed graph.4. The method of claim 3 , further comprising upper bounding to a number K a number of states produced from each state during the first state splitting operation.5. The method of claim 4 , wherein the number K is selected from a group consisting of: seven and eight.6. The method of claim 3 , further comprising selecting the maximum number of states in order that a desired code rate can be achieved when generating the encoder after discarding at least one edge in the first directed graph during ...

DATA PROCESSING APPARATUS AND METHOD

A data processing apparatus and a data processing method thereof are provided. The data processing apparatus includes a register and a processor electrically connected to the register. The register is stored with a plurality of data. The plurality of data each includes a first sub-datum and a second sub-datum. The plurality of first sub-data corresponds to a first column and the plurality of second sub-data corresponds to a second column. The processor compresses the first sub-data by a first compression algorithm according to a first characteristic of the plurality of first sub-data and compresses the second sub-data by a second compression algorithm according to a second characteristic of the plurality of second sub-data. 1. A data processing apparatus , comprising:a register, being configured to store a plurality of data, each of the data comprises a first sub-datum and a second sub-datum, the plurality of first sub-data corresponding to a first column, and the plurality of second sub-data corresponding to a second column; anda processor electrically connected to the register, being configured to compress the plurality of first sub-data by a first compression algorithm according to a first characteristic of the plurality of first sub-data and compress the plurality of second sub-data by a second compression algorithm according to a second characteristic of the plurality of second sub-data.2. The data processing apparatus as claimed in claim 1 , wherein the first compression algorithm and the second compression algorithm are different from each other.3. The data processing apparatus as claimed in claim 1 , wherein each of the plurality of data further comprises an identification code claim 1 , the processor compresses those of the plurality of first sub-data that have a same identification code into a first compressed sub-datum individually by the first compression algorithm claim 1 , and the processor compresses those of the plurality of second sub-data that have ...

Method and device for encoding an orientation vector of a connected component, corresponding decoding method and device and storage medium carrying such encoded data

The invention is made in the field of encoding and decoding at least one orientation vector of a connected component. When quantizing vector components for encoding, an acceptable quantization deviation of encoded vector components sometimes leads to unacceptable deviations of calculated vector components. Therefore, a method is proposed which comprises quantizing and de-quantizing a first and a second component of the vector, and encoding the quantized first and second component and a bit signalling the sign of a third component of said vector, using the pre-determined length and the de-quantized first and second component for determining whether a calculated absolute of an approximation of the third component of said vector is smaller than a first threshold, and, if the calculated absolute is smaller than the first threshold, determining, quantizing and encoding a residual between the calculated absolute of the third component and the absolute of the third component.

SIGNAL COMPRESSION FOR SERIALIZED SIGNAL BANDWIDTH REDUCTION

Signal compression for serialized data bandwidth reduction based on decomposition of a data signal into separate signal components with different SQNR or dynamic range requirements, and quantizing the signal components with different bit precisions. Compression logic decomposes the input data signal into the first/second signal components, quantizes the first component with a pre-defined first bit precision to provide a first quantized data signal, quantizes the second component with a pre-defined second bit precision to provide a second quantized data signal, the second bit precision less than the first bit precision, the first and second quantized data signals bit packed into a compressed digital data signal. At the receive-end, decompression logic bit unpacks the compressed digital data signal into the first/second quantized data signals, and filters/combines the first/second quantized data signals into a decompressed data signal corresponding to the input data signal including the first and second signal components. 1. A communications link , including a transmit end and a receive end , comprising: a data input to receive an input data signal with at least first and second signal components with respectively a higher and a lower SQNR (signal to quantization noise ratio) requirement or lower dynamic range requirement;', decomposition logic to decompose the input data signal into the first and second signal components,', 'quantization logic to quantize the first signal component with a pre-defined first bit precision to provide a first quantized data signal, and to quantize the second signal component with a pre-defined second bit precision to provide a second quantized data signal, the second bit precision less than the first bit precision, and', 'bit-packing logic to assemble the first and second quantized data signals into a compressed digital data signal;, 'compression logic including], 'at the transmit end, a transmit interface including a com input to ...

DYNAMIC DATA COMPRESSION SELECTION

Aspects of dynamic data compression selection are presented. In an example method, as uncompressed data chunks of a data stream are compressed, at least one performance factor affecting selection of one of multiple compression algorithms for the uncompressed data chunks of the data stream may be determined. Each of the multiple compression algorithms may facilitate a different expected compression ratio. One of the multiple compression algorithms may be selected separately for each uncompressed data chunk of the data stream based on the at least one performance factor. Each uncompressed data chunk may be compressed using the selected one of the multiple compression algorithms for the uncompressed data chunk. 1. A method for data compression , the method comprising:receiving a data stream at data-target system from a single data source;detecting each of a set of compressed data chunks within the data stream; andfor each compressed data chunk of at least some of the set of compressed data chunks:identifying a compression algorithm was used to compress the compressed data chunk, wherein the compression algorithm identified to have been used to compress a first compressed data chunk of the at least some of the set of compressed data chunks is a first compression algorithm that differs from a second compression algorithm identified to have been used to compress a second compressed data chunk of the at least some of the set of compressed data chunks;selecting a decompression algorithm based on the identified compression algorithm; anddecompressing the compressed data chunk using the decompression algorithm.2. The method of claim 1 , wherein an identity of the compression algorithm for data compression is correlated with at least one performance factor.3. The method of claim 1 , wherein the first compression algorithm corresponds to a first compression ratio that is different than a second compression ratio to which the second compression ratio corresponds.4. The method of ...

DATA COMPRESSION DEVICES, OPERATING METHODS THEREOF, AND DATA PROCESSING APPARATUSES INCLUDING THE SAME

A method of operating a data compression device includes analyzing data using an analyzer and generating a result of the analysis, while the data is buffered by an input buffer, and selectively compressing the buffered data according to the result of the analysis. A data compression device includes a data pattern analyzer configured to analyze data transmitted to an input buffer, and generate an analysis code based on the analysis of the data; and a data compression manager configured to selectively compress the data in the input buffer based on the analysis code. 1. A method of compressing data , comprising:analyzing, by an analyzer, the data to determine whether to compress the data;generating a result of the analysis, at least one of the analyzing and the generating being performed while the data is buffered in an input buffer; andselectively compressing the buffered data based on the result of the analysis.2. The method of claim 1 , wherein the analyzing the data comprises:analyzing an indication bit included in a header of the data, the indication bit indicating whether to compress the data; whereinthe result of the analysis is generated based on the analysis of the indication bit.3. The method of claim 1 , wherein the analyzing the data comprises:counting a frequency of each of a plurality of symbols included in the data; andcomparing a reference value with one of a maximum and minimum of the counted frequencies; whereinthe result of the analysis is generated based on the comparison.4. The method of claim 1 , wherein the analyzing the data comprises:counting a frequency of each of a plurality of symbols included in the data;assigning a codeword to each of the plurality of symbols;calculating an estimated size of the data if compressed based on the frequency of each of the plurality of symbols and a number of bits of the codeword assigned to each of the plurality of symbols; andcomparing the estimated size with a reference value; whereinthe result of the ...

Generating data pattern information

A data storage system stores at least one dataset including a plurality of records. A data processing system, coupled to the data storage system, processes the plurality of records to produce codes representing data patterns in the records, the processing including: for each of multiple records in the plurality of records, associating with the record a code encoding one or more elements, wherein each element represents a state or property of a corresponding field or combination of fields as one of a set of element values, and, for at least one element of at least a first code, the number of element values in the set is smaller than the total number of data values that occur in the corresponding field or combination of fields over all of the plurality of records in the dataset.

METHOD FOR LOSSLESS COMPRESSION AND REGENERATION OF DIGITAL DESIGN DATA

Disclosed herein is a method for lossless compression and regeneration of digital design data in a manner maintaining the native formats outputted by modeling software used with prime focus on reduction in file size, portability, interchangeability of file storage format and providing database management functions while being implemented as a plug-and-play add-on utility to existing modeling software. Feature-based extraction of design attributes serves as a core of this inventive method and software utility based thereon. 2. The secure method for lossless compression and regeneration of native computer aided design (CAD) file representing a design according to claim 1 , wherein the severable first routine further comprises:a. a first subroutine the execution of which determines geometries present in the input native CAD file and thereafter their measurements including Length, Diameter, Area, Parallelism, Perpendicular, center distance between bodies;b. a second subroutine which when executed subsequent to the first subroutine results in extraction of feature definitions as applicable for each geometry mapped;c. a third subroutine which when executed subsequent to the second subroutine results indetermination of body parameters including loft, rib and their ParentChild relationships for each feature defined;d. a fourth subroutine which when executed post the third subroutine, provides for finalization of extracted data by a discrete function each for obtaining the CAD model or CAD Part file and it's feature details, obtaining the string split by a special character, obtaining the edge name if any else set a new edge name, obtaining Edge details like Start, End, Closed for a particular edge, obtaining the edges for a particular feature based on a user-defined criteria; ande. a fifth subroutine configured to provide an user interface for allowing a user to initialize the first routine and thereafter actuate among discrete functions for running the program in file mode ...

DATA TRANSMISSION METHOD AND APPARATUS

The present application discloses a data transmission method and apparatus. A specific implementation of the method includes: receiving to-be-transmitted data sent from an information sending end, and determining a sending coding type of the to-be-transmitted data; determining a receiving coding type of an information receiving end receiving the to-be-transmitted data; converting the to-be-transmitted data from the sending coding type to the receiving coding type using a preset transcoding model, to obtain transcoded transmission data, the transcoding model representing a corresponding relationship between the sending coding type and the receiving coding type; and sending the transcoded transmission data to the information receiving end. This implementation improves the data transmission efficiency. 1. A data transmission method , comprising:receiving to-be-transmitted data sent from an information sending end, and determining a sending coding type of the to-be-transmitted data;determining a receiving coding type of an information receiving end receiving the to-be-transmitted data;when an existing transcoding model exists, converting the to-be-transmitted data from the sending coding type to the receiving coding type using a preset transcoding model corresponding to the existing transcoding model, to obtain transcoded transmission data, the transcoding model representing a corresponding relationship between the sending coding type and the receiving coding type; and acquiring sending coding type information from information sending ends to form a sending coding type information set and acquiring receiving coding type information from information receiving ends to form a receiving coding type information set;', 'determining a decoding module for a sending coding type corresponding to the sending coding type information, the decoding module decoding to-be-transmitted data of the sending coding type into data contents of a designated type, the data contents comprising ...

Signal sampling with joint training of learnable priors for sampling operator and decoder

A method of sampling and decoding of a signal of interest x comprising, at a training stage: acquiring a set of training signals {x i } i=1 M , providing a sampling operator P Ω and a decoder g θg (.), training operator P Ω on signals {x i } i=1 M to obtain a learned sampling operator P {circumflex over (Ω)} ; and, at a sampling stage: applying P {circumflex over (Ω)} in a transform domain Ψ If to signal x, resulting in observation signal y; applying the decoder g θg (.) to y, to produce an estimate {circumflex over (x)} of signal x to decode and/or, decide about, the signal. Decoder g θg (.) is trained jointly with P Ω on signals {x i } i=1 M , to obtain a learned decoder g θg , by jointly determining, during a cost minimization step, sampling parameters Ω and decoding parameters θ g according to a cost function, and wherein the step of applying the decoder g θg (.) uses decoding parameters θ g , such that estimate {circumflex over (x)} is produced by the learned decoder g {circumflex over (θ)}g .

PATH ENCODING AND DECODING

This invention relates to a system, method and computer program product for encoding an input string of binary characters including: a cellular data structure definition including a starting empty cell; one or more path definitions defining paths through the data structure; a character reading and writing engine for writing a binary character to an empty cell with a predefined initial position; a next cell determination engine for determining a next empty cell by methodically checking cells along one of the paths in the data structure until an empty cell is located; a loop facilitator for looping back to the writing next character step and the determining next cell step until there are no more data characters or a next empty cell is not determined; and a serialization deserialization engine for methodically serializing the data structure into a one dimensional binary string of characters representing an encoded string of alphanumeric characters. 1. A system for encoding an input string of characters , comprising:a definition of a data structure, the data structure definition including a starting empty cell;a path definition defining a path through the data structure;a character reading and writing engine for writing a next character to an empty cell with a predefined initial position;a next cell determination engine for determining a next empty cell by checking cells along the path in the data structure until an empty cell is located;a loop facilitator for looping back to the writing next character step and the determining next empty cell step until there are no more data characters or a next empty cell is not determined; anda serialization deserialization engine for serializing the data structure into a string of characters representing an encoded string of alphanumeric characters.2. A system according to claim 1 , wherein checking cells along the path comprises traversing cells from the current cell in a current direction along the path to a new cell and wherein ...

ADAPTIVE COMPRESSION SUPPORTING OUTPUT SIZE THRESHOLDS

Methods and systems for adaptive compression include compressing input data according to a first compression ratio; pausing compression after a predetermined amount of input data is compressed; estimating which of a set of ranges a compressed output size will fall within using current settings; and performing compression on a remainder of the input data according to a second compression ratio based on the estimated range. 1. A method for adaptive compression , comprising:compressing input data according to a first compression ratio;pausing compression after a predetermined amount of input data is compressed;estimating which of a plurality of ranges a compressed output size will fall within using current settings; andperforming compression on a remainder of the input data according to a second compression ratio based on the estimated range.2. The method of claim 1 , wherein the plurality of ranges correspond to integer multiples of a storage block size.3. The method of claim 1 , wherein said step of performing compression comprises using a second compression ratio that is higher than the first compression ratio if the higher second compression ratio will produce an estimated compressed output size that is below a closest lower range threshold.4. The method of claim 1 , wherein said step of performing compression comprises using a second compression ratio that is lower than the first compression ratio if a higher compression second compression ratio would not produce an estimated compressed output size that is below a closest lower range threshold.5. The method of claim 4 , wherein the second compression ratio is set to produce a compressed output size that is below a closest higher range threshold.6. The method of claim 1 , further comprising aborting compression if a current output size at pausing exceeds a maximum range threshold.7. The method of claim 1 , wherein performing compression on a remainder of the input data comprises repeating said steps of pausing ...

DATA INSPECTION FOR COMPRESSION/DECOMPRESSION CONFIGURATION AND DATA TYPE DETERMINATION

Distribution of data in a neural network data set is used to determine an optimal compressor configuration for compressing the neural network data set and/or the underlying data type of the neural network data set. By using a generalizable optimization of examining the data prior to compressor invocation, the example non-limiting technology herein makes it possible to tune a compressor to better target the incoming data. For sparse data compression, this step may involve examining the distribution of data (e.g., in one example, zeros in the data). For other algorithms, it may involve other types of inspection. This changes the fundamental behavior of the compressor itself. By inspecting the distribution of data (e.g., zeros in the data), it also possible to very accurately predict the data width of the underlying data. This is useful because this data type is not always known a priori, and lossy compression algorithms useful for deep learning depend on knowing the true data type to achieve good compression rates. 1. A data compressor including a processing circuit configured to:inspect data distribution within a data set to detect a granularity; andcompress the data set using the detected granularity to establish correspondence between compressed data elements and data width within the data set.2. The data compressor of wherein the data set comprises neural network data.3. The data compressor of claim 1 , wherein the data set includes binary data and the granularity is detected based on distribution of zeros in a subset of the data set.4. The data compressor of claim 1 , wherein compressing the data set includes:selectively omitting elements of the data width;generating a mask comprising data that stands in for each omitted element; andoutputting the mask and the non-omitted elements.5. A data compressor including a processing circuit configured to:inspect a subset of a data set to detect a distribution of data;infer an element size of the data set based on the ...

GRAPH DATA PROCESSING

Implementations of the present disclosure relate to methods, systems, and computer program products for graph data processing. In one implementation, a computer-implemented method is disclosed. In the method, an adjacency graph of a source graph may be determined by traversing the source graph based on a deep first search rule. Subgraphs may be extracted from the determined adjacency graph based on a predefined shape. Respective subgraph nodes may be built based on nodes in the respective subgraphs. Then the adjacency graph may be updated based on the respective subgraph nodes. In other implementations, a computer-implemented system and a computer program product for graph data processing are disclosed. 1. A method for graph data processing , the method comprising:determining an adjacency graph of a source graph by traversing the source graph based on a deep first search rule;extracting one or more subgraphs from the determined adjacency graph based on a predefined shape;building a plurality of subgraph nodes based on one or more nodes in the extracted one or more subgraphs; andupdating the adjacency graph based on the built plurality of subgraph nodes.2. The method of claim 1 , further comprising:determining a relationship among a plurality of subgraph edges associated with the extracted one or more subgraphs based on one or more shapes associated with the extracted one or more subgraphs;determining the plurality of subgraph edges satisfying the determined relationship from the determined adjacency graph; andutilizing the determined plurality of subgraph edges to extract the one or more subgraphs.3. The method of claim 1 , wherein updating the adjacency graph based on the built plurality of subgraph nodes claim 1 , further comprises:replacing the extracted one or more subgraphs in the determined adjacency graph with the built plurality of subgraph nodes; andupdating the determined plurality of subgraph edges associated with the replaced adjacency graph based on the ...

Lossless machine learning activation value compression

Techniques are disclosed for compressing data. The techniques include identifying, in data to be compressed, a first set of values, wherein the first set of values include a first number of two or more consecutive identical non-zero values; including, in compressed data, a first control value indicating the first number of non-zero values and a first data item corresponding to the consecutive identical non-zero values; identifying, in the data to be compressed, a second value having an exponent value included in a defined set of exponent values; including, in the compressed data, a second control value indicating the exponent value and a second data item corresponding to a portion of the second value other than the exponent value; and including, in the compressed data, a third control value indicating a third set of one or more consecutive zero values in the data to be compressed.

SYSTEMS AND METHODS FOR SEQUENCE ENCODING, STORAGE, AND COMPRESSION

Genomic data is written to disk in a compact format by dividing the data into segments and encoding each segment with the smallest number of bits per character necessary for whatever alphabet of characters appears in that segment. A computer system dynamically chooses the segment boundaries for maximum space savings. A first one of the segments may use a different number of bits per character than a second one of the segments. In one embodiment, dividing the data into segments comprises scanning the data and keeping track of a number of unique characters, noting positions in the sequence where the number increases to a power of two, calculating a compression that would be obtained by dividing the genomic data into one of the plurality of segments at ones of the noted positions, and dividing the genomic data into the plurality of segments at the positions that yield the best compression. 1. A method for encoding genomic data , the method comprising:dividing genomic data into a plurality of segments, andencoding characters within each of the plurality of segments with a smallest number of bits per character needed to represent all unique characters present in each segment, wherein an identity and number of each character is encoded.2. The method of claim 1 , wherein a first one of the segments uses a different number of bits per character than a second one of the segments.3. The method of claim 1 , wherein claim 1 , within a segment claim 1 , every character is encoded using the same number of bits.4. The method of claim 1 , wherein dividing the genomic data into the plurality of segments comprises:scanning the genomic data and keeping track of a number of unique characters scanned;noting positions in the sequence where the number increases to a power of two;calculating a compression that would be obtained by dividing the genomic data into one of the plurality of segments at ones of the noted positions; anddividing the genomic data into the plurality of segments at ...

HARDWARE MODULE FOR CONVERTING NUMBERS

A hardware module comprising circuity configured to: store a sequence of n bits in a register of the hardware module; generate a signed integer comprising a magnitude component and a sign bit by: if the most significant bit of the sequence of n bits is equal to one: set each of the n−1 of the most significant bits of the magnitude component to be equal to the corresponding bit of the n−1 least significant bits of the sequence of n bits; and set the sign bit to be zero; if the most significant bit of the sequence of n bits is equal to zero: set each of the n−1 of the most significant bits of the magnitude component to be equal to the inverse of the corresponding bit of the n−1 least significant bits of the sequence of n bits; and set the sign bit to be one. 1. A hardware module comprising circuity configured to:store a sequence of n bits in a register of the hardware module; and detecting a binary state of the most significant bit of the sequence of n bits; setting the sign bit as the inverse the of most significant bit of the sequence of n bits; and', 'setting each of the n−1 of the most significant bits of the magnitude component to an output binary state in dependence on the detected binary state of the most significant bit of the sequence of n bits wherein responsive to detecting that the binary state of the most significant bit of the sequence of n bits is one, the output binary state equals the corresponding bit of the n−1 least significant bits of the sequence of n bits., 'generate a signed integer comprising a magnitude component and a sign bit by2. (canceled)3. A hardware module as claimed in claim 1 , further comprising a NOT gate configured to receive the most significant bit of the sequence of n bits and to output the sign bit.4. A hardware module as claimed in claim 1 , further comprising a plurality of XNOR gates claim 1 , wherein each XNOR gate is configured to:receive as an input, the most significant bit of the sequence of n bits;receive as an input, ...

INFORMATION PROCESSING APPARATUS, NON-TRANSITORY COMPUTER-READABLE STORAGE MEDIUM HAVING STORED THEREIN INFORMATION PROCESSING PROGRAM, INFORMATION PROCESSING SYSTEM, AND INFORMATION PROCESSING METHOD

An example of a controller calculates the orientation of the controller based on an output from an inertial sensor, and when a value representing the calculated orientation satisfies a first condition, compresses data in a mode and when the first condition is not satisfied, but a second condition is satisfied, compresses the data in a mode and when the second condition is not satisfied, compresses the data in a mode Then, the controller transmits the compressed data to another apparatus. 1. An information processing apparatus including an inertial sensor , the information processing apparatus comprising at least one processor configured to:based on an output from the inertial sensor at a first timing, acquire a first value regarding an orientation or a position of the information processing apparatus;based on an output from the inertial sensor at a second timing different from the first timing, acquire a second value regarding the orientation or the position of the information processing apparatus;based on at least one of the first value and the second value, determine whether or not a first condition is satisfied;when it is determined that the first condition is satisfied, compress at least one of the first value and the second value by a first method, thereby generating first compressed data including data related to the first value and data related to the second value; andwhen it is determined that the first condition is not satisfied, compress at least one of the first value and the second value by a second method different from the first method, thereby generating second compressed data including data related to the first value and data related to the second value.2. The information processing apparatus according to claim 1 , whereinthe at least one processor is further configured to transmit the first compressed data or the second compressed data to another apparatus.3. The information processing apparatus according to claim 1 , whereinbased on at least a ...

Data Compression/Decompression Device

When compressing an arrangement of fixed-length records in a columnar direction, a data compression device carries out data compression aligned with the performance of a data decompression device by computing a number of rows processed with one columnar compression from the performance on the decompression device side, such as the memory cache capacity of the decompression device or the capacity of a primary storage device which may be used by an application, and the size of one record. Thus, while improving compression ratios of large volumes of data, including an alignment of a plurality of fixed-length records, decompression performance is improved. 1. A data decompression device that decompresses compressed data including a plurality of compressed fixed-length records , wherein:the compressed data includes a result acquired by dividing compression object data in units of predetermined compression object block size and compressing the data every compression object block; 'one or more compression object blocks are acquired based upon the compressed data; and', 'when the compression object block size is determined based upon the size of the fixed-length record and the specification information of the data decompression device, the size of the compression object block is determined based upon the size of the fixed-length record and the specification information of the data decompression device;'}a columnar data decompression unit that decompresses each compressed columnar data as a result of compressing the same columnar data of the plurality of fixed-length records included in the compression object block every column and restores the plurality of fixed-length records is provided.2. The data decompression device according to claim 1 , further comprising:a software execution unit that executes software, wherein:the compressed data includes one or more blocks each of which is a result of compression every compression object block and each of which includes one or ...

COMPRESSION OF TIME-VARYING SIMULATION DATA

A method, executed by at least one processor, for compressing time-varying scientific data, includes receiving time-varying data corresponding to a physical phenomenon within a domain comprising one or more spatial dimensions, conducting a proper orthogonal decomposition of the time-varying data to provide basis vectors for the time-varying data, generating a set of expansion coefficients corresponding to the basis vectors that are most prominent in the time-varying data, conducting an image compression algorithm on the expansion coefficients to provide a compressed representation of the time-varying data, and storing the compressed representation of the time-varying data. The time-varying data may be numeric data generated from a physical simulation or from experimentation. In some embodiments, the time-varying data corresponds to one or more sub-domains within a larger dataset. The sub-domains may be coherent sub-domains that have similar modes. A corresponding computer-program product and computing system are also disclosed herein. 1. A method , executed by at least one processor , for compressing time-varying scientific data , the method comprising:receiving time-varying data corresponding to a physical phenomenon within a domain comprising one or more spatial dimensions;conducting a proper orthogonal decomposition of the time-varying data to provide basis vectors for the time-varying data;generating a set of expansion coefficients corresponding to the basis vectors that are most prominent in the time-varying data;conducting an image compression algorithm on the expansion coefficients to provide a compressed representation of the time-varying data; andstoring the compressed representation of the time-varying data.2. The method of claim 1 , further comprising conducting the image compression algorithm on the basis vectors that are most prominent to provide compressed basis vectors and including the compressed basis vectors in the compressed representation.3. The ...

ENCODING METHOD AND DEVICE, DECODING METHOD AND DEVICE, AND DISPLAY DEVICE

This application relates to an encoding method and device, a decoding method and device, and a signal transmission system. The encoding method includes: encoding 8-bit data corresponding to a to-be-encoded byte of to-be-transmitted data into 9-bit data, the to-be-transmitted data comprising at least one to-be-encoded byte; detecting the first digit of data of the 9-bit data and a previous digit of data adjacent to the first digit of data, when the to-be-encoded byte is not the first byte of the to-be-transmitted data; inverting the 9-bit data and then adding a tenth digit of data for indicating that the inverted 9-bit data has undergone an inversion operation behind the inverted 9-bit data to obtain 10-bit data, when the numerical value of the first digit of data is the same as that of the previous digit of data; and adding a tenth digit of data for indicating that the 9-bit data has not undergone an inversion operation behind the 9-bit data to obtain 10-bit data, when the numerical value of the first digit of data is different from that of the previous digit of data. 1. An encoding method , comprising:encoding 8-bit data corresponding to a to-be-encoded byte of to-be-transmitted data into 9-bit data, the to-be-transmitted data comprising at least one to-be-encoded byte;detecting the first digit of data of the 9-bit data and a previous digit of data adjacent to the first digit of data, when the to-be-encoded byte is not the first byte of the to-be-transmitted data;inverting the 9-bit data and then adding a tenth digit of data for indicating that the inverted 9-bit data has undergone an inversion operation behind the inverted 9-bit data to obtain 10-bit data, when the numerical value of the first digit of data is the same as that of the previous digit of data; andadding a tenth digit of data for indicating that the 9-bit data has not undergone an inversion operation behind the 9-bit data to obtain 10-bit data, when the numerical value of the first digit of data is ...