A computer-implemented method of predicting a branch direction of a fetch block in a processor, includes in part, determining a multitude of first counts each associated with a different one of a multitude of branch offsets of a branch direction predictor data associated with the fetch block. Each of the multitude of first counts represents the number of times that the associated branch offset was taken during a multitude of fetch cycles. The computer-implemented method further includes, in part, determining a second count associated with the fetch block. The second count represents the number of times that none of the multitude of branch offsets were taken during the multitude of fetch cycles. The computer-implemented method further includes, in part, computing a confidence level based on the multitude of first counts and the second count, and determining the branch direction of the fetch block in accordance with the computed confidence level.
This document concerns using of an Instruction Accurate reference model of a hardware micro-architecture as the reference to verify a central processing unit (CPU) hardware implementation, include the following. 1) A ‘mirror’ mechanism that enables the VC to maintain an exact copy of the internal architectural state of the DUT. 2) A ‘volatile’ algorithm that allows the VC/RM to adapt its internal state when that state is not architecturally defined, but micro-architecturally (e.g. implementation) defined. 3) A use of ‘speculative execution’ to explore different possible permissible paths through the execution state space of the RM especially in response to asynchronous events and hidden details of the DUT implementations. 4) A technique described as ‘convergence’ which allows the RM to adapt its internal state after a divergence in behaviour/state between the DUT and RM, allowing the verification process to continue.
A memory protection unit (MPU) configuration request may be received, where the MPU configuration request may include a memory protection rule. A first entry in a first MPU circuit may be determined which matches the memory protection rule. A compliance result may be determined based on checking if the memory protection rule complies with the first entry. The memory protection rule may be written in a second MPU circuit based on the compliance result.
Techniques for data manipulation using integer matrix multiplication using pipelining are disclosed. A first integer matrix with dimensions m×k and a second integer matrix with dimensions k×n are obtained for matrix multiplication within a processor. The first and second integer matrices employ a two's complement variable radix point data representation. The first and second integer matrices are distilled into (j×j) submatrices. A first variable radix point format and an initial value for an accumulator register are configured dynamically. A first variable radix point format is configured dynamically for the first integer matrix and a second variable radix point format is configured dynamically for the second integer matrix. Multiply-accumulate operations are executed in a pipelined fashion on the (j×j) submatrices of the first integer matrix and the second integer matrix, where a third variable radix point format is configured for the result.
A request to provide an application with direct memory access to data stored at an external memory address of an external memory is received. Responsive to determining that the external memory address is not registered in a cache, the data is copied from the external memory address to a first internal memory address within the internal memory. A first cache line, within the cache, associated with the external memory address is updated to include a reference to the first internal memory address. The data is provided from the internal memory to the application.
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
G06F 12/0817 - Protocoles de cohérence de mémoire cache à l’aide de méthodes de répertoire
G06F 12/0853 - Mémoire cache avec matrices multiples d’étiquettes ou de données
6.
Integer matrix multiplication engine using pipelining
Techniques for data manipulation using integer matrix multiplication using pipelining are disclosed. A first integer matrix with dimensions m×k and a second integer matrix with dimensions k×n are obtained for matrix multiplication within a processor. The first and second integer matrices employ a two's complement variable radix point data representation. The first and second integer matrices are distilled into (j×j) submatrices. A first variable radix point format and an initial value for an accumulator register are configured dynamically. A first variable radix point format is configured dynamically for the first integer matrix and a second variable radix point format is configured dynamically for the second integer matrix. Multiply-accumulate operations are executed in a pipelined fashion on the (j×j) submatrices of the first integer matrix and the second integer matrix, where a third variable radix point format is configured for the result.
An adaptive hardware trace circuit is presented. The adaptive hardware trace circuit may include one or more trace circuits, a trace port funnel circuit, a trace FIFO buffer, and an adaptation logic circuit. Each trace circuit may be coupled to a processor core and configured to monitor and encode trace data generated by a processor core. The trace buffer may be configured to store the trace data generated by the processor cores. The adaptation circuit may be configured to receive, from a user, one or more buffer capacity thresholds and a priority level assigned to each trace. The adaptation circuit may map ranges of trace buffer capacities to corresponding sets of actions. The adaptation circuit may detect a buffer capacity to determine a set of one or more actions associated with the buffer capacity and execute the set of one or more actions.
A method of detecting an error includes, in part, defining a bit pattern using a first multitude of bits, a second multitude of bits, the bits of an error correction code (ECC), and at least one user selected bit. The method further includes, in part, receiving a first value represented by the first multitude of bits and the at least one user selected bit; receiving a second value represented by the second multitude of bits; receiving a third value represented by the ECC bits. The method further includes, in part, generating a syndrome value from the first, second and third values; and using a subset of the syndrome value bits to detect the error in the first, second or third values. The third value is determined in accordance with the first and second values.
H03M 13/15 - Codes cycliques, c.-à-d. décalages cycliques de mots de code produisant d'autres mots de code, p. ex. codes définis par un générateur polynomial, codes de Bose-Chaudhuri-Hocquenghen [BCH]
H03M 13/00 - Codage, décodage ou conversion de code pour détecter ou corriger des erreursHypothèses de base sur la théorie du codageLimites de codageMéthodes d'évaluation de la probabilité d'erreurModèles de canauxSimulation ou test des codes
G11C 29/42 - Dispositifs de vérification de réponse utilisant des codes correcteurs d'erreurs [ECC] ou un contrôle de parité
G11C 29/52 - Protection du contenu des mémoiresDétection d'erreurs dans le contenu des mémoires
H03M 13/11 - Détection d'erreurs ou correction d'erreurs transmises par redondance dans la représentation des données, c.-à-d. mots de code contenant plus de chiffres que les mots source utilisant un codage par blocs, c.-à-d. un nombre prédéterminé de bits de contrôle ajouté à un nombre prédéterminé de bits d'information utilisant plusieurs bits de parité
H03M 13/45 - Décodage discret, c.-à-d. utilisant l'information de fiabilité des symboles
9.
SYSTEM SIMULATION USING SIMULATION MODELS EXECUTING ON VIRTUAL PROCESSORS
Simulating a system involves running a simulation model on a virtual processor that interfaces to a simulator, the virtual processor being associated with a unique memory space in the memory space used by the simulator. The virtual processor technique gives the developer a simple technique for ensuring the model and simulator interactions are managed effectively, i.e. that no unintentional corruptions of each other's memory space are possible. It also facilitates making multiple instantiations of a C, C++ or other general purpose programming language model. The resulting environment also resolves one of the limitations on the use of third party models within the SystemC environment.
A second virtual address may be received, where the second virtual address is different from a first virtual address. A second hash value may be computed based on the second virtual address. A first comparison result may be determined by comparing the second hash value with a first hash value, where the first hash value is computed based on the first virtual address. The first comparison result may be used to select a selected structure from either a first structure or a second structure. The selected structure may be used to determine predicted aliasing bits which are used to determine an index corresponding to the second virtual address.
G06F 12/08 - Adressage ou affectationRéadressage dans des systèmes de mémoires hiérarchiques, p. ex. des systèmes de mémoire virtuelle
G06F 12/0864 - Adressage d’un niveau de mémoire dans lequel l’accès aux données ou aux blocs de données désirés nécessite des moyens d’adressage associatif, p. ex. mémoires cache utilisant des moyens pseudo-associatifs, p. ex. associatifs d’ensemble ou de hachage
G06F 12/1045 - Traduction d'adresses utilisant des moyens de traduction d’adresse associatifs ou pseudo-associatifs, p. ex. un répertoire de pages actives [TLB] associée à une mémoire cache de données
Techniques for mixed precision quantization of a machine learning (ML) model. The techniques include receiving a target performance relating to the ML model including objects of a first data type represented by a first number of bits, wherein the target performance relates to changing a first portion of the objects to a second data type represented by a second number of bits and changing a second portion of the objects to a third data type represented by a third number of bits. The techniques further include selecting the first portion and the second portion, based on maintaining a performance relating to the ML model at or below the target performance, and changing the first portion of objects from the first data type to the second data type and the second portion of objects from the first data type to the third data type.
Disclosed herein are apparatus, method, and computer-readable storage device embodiments for implementing deconvolution via a set of convolutions. For example, an apparatus includes a memory and at least one processor, communicatively coupled with the memory, to perform operations including obtaining a tensor representing image data, generating a deconvolution segment of a plurality of deconvolution segments by performing a convolution to transform the tensor based at least in part on a convolution filter and a stride value, and generating a complete deconvolution for image processing by assembling each deconvolution segment of the plurality of deconvolution segments.
G06N 3/063 - Réalisation physique, c.-à-d. mise en œuvre matérielle de réseaux neuronaux, de neurones ou de parties de neurone utilisant des moyens électroniques
G06N 3/04 - Architecture, p. ex. topologie d'interconnexion
Techniques for data manipulation using processor cluster address generation are disclosed. One or more processor clusters capable of executing software-initiated work requests are accessed. A plurality of dimensions from a tensor is flattened into a single dimension. A work request address field is parsed, where the address field contains unique address space descriptors for each of the plurality of dimensions, along with a common address space descriptor. A direct memory access (DMA) engine coupled to the one or more processor clusters is configured. Addresses are generated based on the unique address space descriptors and the common address space descriptor. The plurality of dimensions can be summed to generate a single address. Memory is accessed using two or more of the addresses that were generated. The addresses are used to enable DMA access.
G06F 12/06 - Adressage d'un bloc physique de transfert, p. ex. par adresse de base, adressage de modules, extension de l'espace d'adresse, spécialisation de mémoire
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
14.
Varied validity bit placement in tag bits of a memory
A system and method access memory blocks in a memory by receiving a memory transaction request from a processing device. First hash bits of the memory transaction request are compared with second hash bits of a first memory block of a memory. Data associated with the first memory block is output to the processing device based on the comparison of the first hash bits with the second hash bits.
G06F 12/08 - Adressage ou affectationRéadressage dans des systèmes de mémoires hiérarchiques, p. ex. des systèmes de mémoire virtuelle
G06F 12/0802 - Adressage d’un niveau de mémoire dans lequel l’accès aux données ou aux blocs de données désirés nécessite des moyens d’adressage associatif, p. ex. mémoires cache
15.
Memory coherence protocol for communicating data associated with a shared state between processor cores
A system and method service memory transaction requests by receiving a memory transaction request for a first memory line from a first processor core of processor cores of a processing system. A second processor core of the processor cores is determined to include the first memory line in a shared state. Data of the first memory line is communicated from the second processor core to the first processor core based on determining that the second processor core comprises the first memory line in a shared state.
A processing system employs a method to order the elements within a memory. Ordering the elements includes receiving an accessed memory element. The accessed memory element is requested by a processor from a memory. Further, the accessed memory element is compared to stored elements within the memory to generate control signals. Gate control signals from the control signals are generated. The order of the stored elements within the memory is updated based on the gate control signals.
G06F 12/0891 - Adressage d’un niveau de mémoire dans lequel l’accès aux données ou aux blocs de données désirés nécessite des moyens d’adressage associatif, p. ex. mémoires cache utilisant des moyens d’effacement, d’invalidation ou de réinitialisation
17.
Conflict detection and address arbitration for routing scatter and gather transactions for a memory bank
A system and method for mitigating memory transaction conflicts by receiving a first memory transaction from a first processor slice of a processor and a second memory transaction from a second processor slice of the processor. Further, one or more control signals are generated for the first memory transaction and the second memory transaction based on a determination that the first memory transaction and the second memory transaction have a target address associated with a first memory bank of a memory. The first memory transaction is selected to output to the first memory bank based on the one or more control signals.
A system and method mitigates conflicts between clean unique requests by receiving a first clean unique request from a first processor core and a second clean unique request from a second processor core. The first clean unique request and the second clean unique request respectively indicate that the first processor core and second processor core request access to a first address of a memory. The memory is coupled to the first processor core and the second processor core. The first clean unique request and the second clean unique request are determined to be associated with the first address. Further, the second clean unique request is converted into a first read unique request based on determining that the first clean unique request and the second clean unique request are associated with the first address. The first read unique requests indicates that the second processor core requests data.
G06F 12/00 - Accès à, adressage ou affectation dans des systèmes ou des architectures de mémoires
G06F 12/0815 - Protocoles de cohérence de mémoire cache
G06F 12/084 - Systèmes de mémoire cache multi-utilisateurs, multiprocesseurs ou multitraitement avec mémoire cache partagée
G06F 12/0842 - Systèmes de mémoire cache multi-utilisateurs, multiprocesseurs ou multitraitement pour multitraitement ou multitâche
G06F 12/0891 - Adressage d’un niveau de mémoire dans lequel l’accès aux données ou aux blocs de données désirés nécessite des moyens d’adressage associatif, p. ex. mémoires cache utilisant des moyens d’effacement, d’invalidation ou de réinitialisation
Merging store instructions for a memory includes receiving a first store instruction having a first address, and determining a first pattern based on a comparison of the first address and a second address of an entry within a buffer. Further, a size field of the entry is updated based on the first pattern. The first address of the first store instruction is merged with the second address within the entry to generate a merged instruction. The merged store instruction is communicated to the memory.
A system and method for performing a store to load process includes receiving a first store instruction. The first store instruction includes a first target address, a first mask, and a first data structure. Further, the first target address, the first mask, and the first data structure are stored within a first store buffer location of a store buffer. A first entry identification associated with the first store buffer location is stored within an age buffer. The first data structure is output based on an order of entry identifications within the age buffer.
A system and method corrects single bit errors in a memory by detecting a single bit error with a memory. The memory is accessed via data cache stages of a pipeline. Further, based on detecting the single bit error, the data cache stages of the pipeline are stopped from accepting new transactions. A value associated with each address of the memory is read based on stopping the new transactions from being accepted, and the detected single bit errors within the values are corrected.
G06F 11/10 - Détection ou correction d'erreur par introduction de redondance dans la représentation des données, p. ex. en utilisant des codes de contrôle en ajoutant des chiffres binaires ou des symboles particuliers aux données exprimées suivant un code, p. ex. contrôle de parité, exclusion des 9 ou des 11
22.
PROCESSOR GRAPH EXECUTION USING INTERRUPT CONSERVATION
Techniques for data manipulation using processor graph execution using interrupt conservation are disclosed. Processing elements are configured to implement a data flow graph. The processing elements comprise a multilayer graph execution engine. A data engine is loaded with computational parameters for the multilayer graph execution engine. The data engine is coupled to the multilayer graph execution engine, and the computational parameters supply layer-by-layer execution data to the multilayer graph execution engine for data flow graph execution. A first command FIFO is used for loading the data engine with computational parameters, and a second command FIFO is used for loading the multilayer graph execution engine with layer definition data. An input image is provided for a first layer of the multilayer graph execution engine. The data flow graph is executed using the input image and the computational parameters. The executing is controlled by interrupts only when an uncertainty exists within the data flow graph.
A memory address may be received at an m-way set-associative cache, which may store a set of cache entries. The memory address may be partitioned into a tag, an index, and an offset. The m-way set-associative cache may include a first structure to store a first subset of tag bits corresponding to the set of cache entries and a second structure to store a second subset of tag bits corresponding to the set of cache entries. The index may be used to select a first set of entries from the first structure. A first portion of tag bits of the memory address may be matched with the first set of entries. A cache status may be determined based on matching the first portion of tag bits of the memory address with the first set of entries.
G06F 12/0864 - Adressage d’un niveau de mémoire dans lequel l’accès aux données ou aux blocs de données désirés nécessite des moyens d’adressage associatif, p. ex. mémoires cache utilisant des moyens pseudo-associatifs, p. ex. associatifs d’ensemble ou de hachage
G06F 12/0846 - Mémoire cache avec matrices multiples d’étiquettes ou de données accessibles simultanément
A system for executing tensor operations including: a programmable tensor processor; and a memory coupled to the programmable tensor processor, wherein the programmable tensor processor includes: one or more load AGU circuits to generate a first sequence of addresses and read input tensor operands from the memory based on the first sequence of addresses; a datapath circuit to perform the tensor operations on the input tensor operands based on receiving one or more instructions to determine output tensor operands, the one or more instructions being based on a loop iteration count and loop body micro-code instructions defining a loop body of a tensor program stored in the memory, the loop body micro-code instructions being executed in the programmable tensor processor; and a store AGU circuit configured to generate a second sequence of addresses and write the output tensor operands to the memory based on the second sequence of addresses.
G06F 12/1009 - Traduction d'adresses avec tables de pages, p. ex. structures de table de page
G06F 7/544 - Méthodes ou dispositions pour effectuer des calculs en utilisant exclusivement une représentation numérique codée, p. ex. en utilisant une représentation binaire, ternaire, décimale utilisant des dispositifs n'établissant pas de contact, p. ex. tube, dispositif à l'état solideMéthodes ou dispositions pour effectuer des calculs en utilisant exclusivement une représentation numérique codée, p. ex. en utilisant une représentation binaire, ternaire, décimale utilisant des dispositifs non spécifiés pour l'évaluation de fonctions par calcul
G06F 9/30 - Dispositions pour exécuter des instructions machines, p. ex. décodage d'instructions
25.
Efficient look-up table based functions for artificial intelligence (AI) accelerator
A method for approximating an activation function, the method including: receiving an input value of the activation function; determining that the input value is within a range, the range includes a set of non-uniform intervals; determining a selected interval from among the set of non-uniform intervals including the input value; retrieving, by a hardware accelerator, from a look-up table (LUT) associated with a type of the activation function, values of one or more quadratic interpolation parameters associated with the selected interval; performing a quadratic interpolation on the input value to approximate the input value using the values of the one or more quadratic interpolation parameters; and determining a first approximated output of the activation function based on a result of the quadratic interpolation performed on the input value.
A computer-implemented method including: providing a test template for a hardware system-under-test comprising one or more execution threads, wherein the test template comprises a branching instruction to a predetermined shared memory address accessible by at least some of the one or more execution threads; generating and storing, at the predetermined shared memory address, a sequence of instructions which conform to the test template; building, based, at least in part, on the test template, an executable image of a hardware exerciser, wherein the hardware exerciser is adapted to control a test cycle of the hardware system-under-test, and wherein the test cycle comprises at least generation and execution of a test; and executing the executable image of the hardware exerciser by at least a first execution thread of the one or more execution threads of the hardware system-under-test.
The invention relates to methods of simulation of a plurality of processors running on a plurality of cores, to multi-core microprocessor systems in which such methods may be carried out, and to computer program products configured to perform a simulation of a plurality of processors, running on a plurality of cores. According to a first aspect of the invention, there is provided a method of running a plurality of simulated processors on a plurality of cores, in which simulation of the processors is performed in parallel on the plurality of cores.
Embodiments relate to a system, program product, and method for integrating compiler-based testing in post-silicon validation. The method includes generating a test program through a device-under-test (DUT). The method also includes generating a plurality of memory intervals and injecting the plurality of memory intervals into the test program. The method further includes injecting a plurality of compiled test snippets into the test program and executing one or more post-silicon validation tests for the DUT with the test program.
Techniques for data manipulation using integer matrix multiplication using pipelining are disclosed. A first integer matrix with dimensions m×k and a second integer matrix with dimensions k×n are obtained for matrix multiplication within a processor. The first and second integer matrices employ a two's complement variable radix point data representation. The first and second integer matrices are distilled into (j×j) submatrices. A first variable radix point format and an initial value for an accumulator register are configured dynamically. A first variable radix point format is configured dynamically for the first integer matrix and a second variable radix point format is configured dynamically for the second integer matrix. Multiply-accumulate operations are executed in a pipelined fashion on the (j×j) submatrices of the first integer matrix and the second integer matrix, where a third variable radix point format is configured for the result.
Techniques for data manipulation using processor graph execution using interrupt conservation are disclosed. Processing elements are configured to implement a data flow graph. The processing elements comprise a multilayer graph execution engine. A data engine is loaded with computational parameters for the multilayer graph execution engine. The data engine is coupled to the multilayer graph execution engine, and the computational parameters supply layer-by-layer execution data to the multilayer graph execution engine for data flow graph execution. A first command FIFO is used for loading the data engine with computational parameters, and a second command FIFO is used for loading the multilayer graph execution engine with layer definition data. An input image is provided for a first layer of the multilayer graph execution engine. The data flow graph is executed using the input image and the computational parameters. The executing is controlled by interrupts only when an uncertainty exists within the data flow graph.
Embodiments relate to a system, program product, and method for random generation of recoverable errors in the generated instruction stream for post-silicon validation testing. The intentional raising and handling of exceptions in post-silicon validation exercisers randomly creates recoverable errors in a generated instruction test stream. Multiple exceptions may be raised either in a single instruction or in multiple instructions, while the present instruction is permitted to fully execute. The errors responsible for raising the exceptions are automatically repaired.
G01R 31/28 - Test de circuits électroniques, p. ex. à l'aide d'un traceur de signaux
G01R 31/3177 - Tests de fonctionnement logique, p. ex. au moyen d'analyseurs logiques
G01R 31/3183 - Génération de signaux d'entrée de test, p. ex. vecteurs, formes ou séquences de test
G06F 11/36 - Prévention d'erreurs par analyse, par débogage ou par test de logiciel
G06F 11/22 - Détection ou localisation du matériel d'ordinateur défectueux en effectuant des tests pendant les opérations d'attente ou pendant les temps morts, p. ex. essais de mise en route
G06F 11/263 - Génération de signaux d'entrée de test, p. ex. vecteurs, formes ou séquences de test
Techniques for mixed precision quantization of a machine learning (ML) model. A target bandwidth increase is received (302), for the ML model (114) including objects of a first data type represented by a first number of bits. The target bandwidth increase relates to changing a first portion of the objects to a second data type represented by a second number of bits different from the first number of bits (310). The method further includes sorting the objects in the ML model based on bandwidth (304). The method further includes identifying the first portion of the objects to change from the first data type to the second data type, based on the target bandwidth increase and the sorting of the plurality of objects (508). The method further includes changing the first portion of the objects from the first data type to the second data type (508).
Method, apparatus and product for automatically introducing register dependency into tests. A test template represents an abstract test scenario to be utilized for testing a target processor. The abstract test scenario requires that a value be assigned to a register. A test that implements the abstract test scenario is generated. The test is a set of instructions that are executable by the target processor. The generation of the test comprises: determining a memory address to retain the value in a memory that is accessible to the target processor; and adding to the test an instruction to load to the register the value from the memory address, whereby adding a register dependency to the test that is not required by the abstract test scenario. The test can be executed on the target processor or simulation thereof.
G06F 9/38 - Exécution simultanée d'instructions, p. ex. pipeline ou lecture en mémoire
G06F 11/22 - Détection ou localisation du matériel d'ordinateur défectueux en effectuant des tests pendant les opérations d'attente ou pendant les temps morts, p. ex. essais de mise en route
A method, apparatus and product for utilizing address translation structures for testing address translation cache. The method comprises: obtaining a first address translation structure that comprises multiple levels, including a first top level which connects a sub-structure of the first address translation structure using pointers thereto; determining, based on the first address translation structure, a second address translation structure, wherein the second address translation structure comprises a second top level that is determined based on the first top level, wherein the second top level connects the sub-structure of the first address translation structure; executing a test so as to verify operation of an address translation cache of a target processor at least by: adding a plurality of cache lines to the address translation cache, wherein said adding is based on the address translation structures; and verifying the operation of the address translation cache using one or more memory access operations.
G06F 12/00 - Accès à, adressage ou affectation dans des systèmes ou des architectures de mémoires
G06F 12/1045 - Traduction d'adresses utilisant des moyens de traduction d’adresse associatifs ou pseudo-associatifs, p. ex. un répertoire de pages actives [TLB] associée à une mémoire cache de données
G06N 5/00 - Agencements informatiques utilisant des modèles fondés sur la connaissance
35.
Utilizing translation tables for testing processors
A method, apparatus and a product for utilizing translation tables for testing processors. The method is used for testing a target processor that utilizes different translation tables to translate virtual addresses to physical addresses. The method comprises obtaining a test template that comprises directives to be executed in different contexts, during each of which a different translation table is utilized to translate virtual addresses to physical addresses. The translation tables to be used by the target processor in the different contexts are determined, so that the translation tables overlap, at least in part.
G06F 12/1045 - Traduction d'adresses utilisant des moyens de traduction d’adresse associatifs ou pseudo-associatifs, p. ex. un répertoire de pages actives [TLB] associée à une mémoire cache de données
G06F 11/263 - Génération de signaux d'entrée de test, p. ex. vecteurs, formes ou séquences de test
Techniques for data manipulation using processor cluster address generation are disclosed. One or more processor clusters capable of executing software-initiated work requests are accessed. A plurality of dimensions from a tensor is flattened into a single dimension. A work request address field is parsed, where the address field contains unique address space descriptors for each of the plurality of dimensions, along with a common address space descriptor. A direct memory access (DMA) engine coupled to the one or more processor clusters is configured. Addresses are generated based on the unique address space descriptors and the common address space descriptor. The plurality of dimensions can be summed to generate a single address. Memory is accessed using two or more of the addresses that were generated. The addresses are used to enable DMA access.
G06F 12/06 - Adressage d'un bloc physique de transfert, p. ex. par adresse de base, adressage de modules, extension de l'espace d'adresse, spécialisation de mémoire
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
37.
Integer matrix multiplication engine using pipelining
Techniques for data manipulation using integer matrix multiplication using pipelining are disclosed. A first integer matrix with dimensions m×k and a second integer matrix with dimensions k×n are obtained for matrix multiplication within a processor. The first and second integer matrices employ a two's complement variable radix point data representation. The first and second integer matrices are distilled into (j×j) submatrices. A first variable radix point format and an initial value for an accumulator register are configured dynamically. A first variable radix point format is configured dynamically for the first integer matrix and a second variable radix point format is configured dynamically for the second integer matrix. Multiply-accumulate operations are executed in a pipelined fashion on the (j×j) submatrices of the first integer matrix and the second integer matrix, where a third variable radix point format is configured for the result.
Techniques for data manipulation using processor cluster address generation are disclosed. One or more processor clusters capable of executing software-initiated work requests are accessed. A direct memory access (DMA) engine, coupled to the one or more processor clusters, is configured, wherein the DMA engine employs address generation across a plurality of tensor dimensions. A work request address field is parsed, where the address field contains unique address space descriptors for each of the plurality of dimensions, along with a common address space descriptor. DMA addresses are generated based on the unique address space descriptors and the common address space descriptor. Memory using two or more of the DMA addresses that were generated is accessed, where the two or more DMA addresses enable processing within the one or more processor clusters.
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
Techniques for data manipulation using a matrix multiplication engine using pipelining are disclosed. A first and a second matrix are obtained for matrix multiplication. A first matrix multiply-accumulate (MAC) unit is configured, where a first matrix element and a second matrix element are presented to the MAC unit on a first cycle. A second MAC unit is configured in pipelined fashion, where the first element of the first matrix and a second element of the second matrix are presented to the second MAC unit on a second cycle, and where a second element of the first matrix and the first element of the second matrix are presented to the first MAC unit on the second cycle. Additional MAC units are further configured within the processor in pipelined fashion. Multiply-accumulate operations are executed in pipelined fashion on each of n MAC units over additional k sets of m cycles.
Disclosed herein are apparatus, method, and computer-readable storage device embodiments for implementing deconvolution via a set of convolutions. An embodiment includes a convolution processor that includes hardware implementing logic to perform at least one algorithm comprising a convolution algorithm. The at least one convolution processor may be further configured to perform operations including performing a first convolution and outputting a first deconvolution segment as a result of the performing the first convolution. The at least one convolution processor may be further configured to perform a second convolution and output a second deconvolution segment as a result of the performing the second convolution. According to some embodiments, the at least one convolution processor may be further configured to perform at least one further convolution to generate at least one further deconvolution segment, until a number of deconvolution segments output by the convolution processor reaches a deconvolution-size-value.
G06N 3/063 - Réalisation physique, c.-à-d. mise en œuvre matérielle de réseaux neuronaux, de neurones ou de parties de neurone utilisant des moyens électroniques
G06N 3/04 - Architecture, p. ex. topologie d'interconnexion
41.
Data uploading to asynchronous circuitry using circular buffer control
Disclosed embodiments provide an interface circuit for the transfer of data from a synchronous circuit to an asynchronous circuit. Data from the synchronous circuit is received into a memory in the interface circuit. The data in the memory is then sent to the asynchronous circuit based on an instruction in a circular buffer that is part of the interface circuit. Processing elements within the interface circuit execute instructions contained within the circular buffer. The circular buffer rotates to provide new instructions to the processing elements. Flow control paces the data from the synchronous circuit to the asynchronous circuit.
H04L 7/02 - Commande de vitesse ou de phase au moyen des signaux de code reçus, les signaux ne contenant aucune information de synchronisation particulière
H04L 5/24 - Dispositions destinées à permettre l'usage multiple de la voie de transmission utilisant le multiplex à division de temps avec des convertisseurs synchrones marche-arrêt
H04L 12/801 - Commande de flux ou commande de congestion
H04L 12/933 - Cœur de commutateur, p.ex. barres croisées, mémoire partagée ou support partagé
H04L 12/863 - Ordonnancement de file d’attente, p.ex. ordonnancement circulaire
Techniques for a neural network output layer for machine learning are disclosed. A plurality of processing elements within a reconfigurable fabric is configured to implement a data flow graph, where the data flow graph implements a neural network. The data flow graph can include machine learning or deep learning. A layer is implemented, within the neural network, that maps a first vector of real values to a second vector of real values bounded by zero and one, where the second vector sums to a value of one using fixed-point calculations. The layer can include a final layer within the neural network. The layer that maps the first vector includes a Softmax function. Results of the neural network are classified based on a value of the second vector. The classifying can include part of a machine learning or a deep learning process.
Techniques are disclosed for data manipulation within a reconfigurable computing environment for data flow graph computation using exceptions. Processing elements are configured within a reconfigurable fabric to implement a data flow graph. The processing elements are loaded with process agents. Valid data is executed by a first process agent on a first processing element, where the first process agent corresponds to a starting node of the data flow graph. A second processing element detects that an error exception has occurred, where a second process agent is running on the second processing element. A done signal to a third process agent is withheld by the second process agent, where the third process agent is running on a third processing element. The second process agent raises an interrupt request, where the interrupt request is based on the detecting that an error exception has occurred.
Techniques are disclosed for data flow graph node parallel update for machine learning. A first plurality of processing elements is configured to implement a portion of a data flow graph. The nodes include at least one variable node and implement part of a neural network. A second plurality of processing elements is configured to implement a second portion of the data flow graph. These nodes include at least one additional variable node and implement an additional part of the neural network. Training data is issued to the first plurality of processing elements. The training data is used to update variables within the at least one variable node. Additional variables are updated within the at least one additional variable node. The updating includes forwarding training data from the first plurality to the second plurality. The neural network is trained based on the variables that were updated and the additional variables.
An interface circuit is disclosed for the transfer of data from a synchronous circuit, with multiple source elements, to an asynchronous circuit. Data from the synchronous circuit is received into a memory in the interface circuit. The data in the memory is then sent to the asynchronous circuit based on an instruction in a circular buffer that is part of the interface circuit. Processing elements within the interface circuit execute instructions contained within the circular buffer. The circular buffer rotates to provide new instructions to the processing elements. Flow control paces the data from the synchronous circuit to the asynchronous circuit.
H04L 12/28 - Réseaux de données à commutation caractérisés par la configuration des liaisons, p. ex. réseaux locaux [LAN Local Area Networks] ou réseaux étendus [WAN Wide Area Networks]
H04L 12/861 - Mise en mémoire tampon de paquets ou mise en file d’attente; Ordonnancement de file d’attente
H04L 12/863 - Ordonnancement de file d’attente, p.ex. ordonnancement circulaire
H04L 12/801 - Commande de flux ou commande de congestion
H04L 12/859 - Actions liées à la commande de flux basée sur la nature de l’application, p.ex. contrôle de navigation sur l’Internet ou contrôle du trafic de courrier électronique
A method, computer program product, and computer system for physical memory allocation of a computer system, the method including collecting computer system architecture specifications, a configuration, and user requirements, identifying a plurality of memory intervals to be allocated, based on the computer system architecture specification, the configuration, and the user requirements, grouping memory intervals into a plurality of color groups, wherein each memory interval within each of the plurality of color groups comprise identical memory attributes, dividing memory into sets of memory segments, wherein each set of memory segment is assigned a color of the plurality of color groups, allocating a memory interval of the plurality of memory intervals within the set of memory segments of corresponding color, and selecting a page size for a translation of a memory interval of the plurality of memory intervals, depending upon the allocation of the memory interval and the sets of memory segments.
Techniques are disclosed for power conservation. A plurality of processing elements and a plurality of instructions are configured. The plurality of processing elements is controlled by instructions contained in a plurality of circular buffers. The plurality of processing elements can comprise a data flow processor. A first processing element, from the plurality of interconnected processing elements, is set into a sleep state by a first instruction from the plurality of instructions. The first processing element is woken from the sleep state as a result of valid data being presented to the first processing element. A subsection of the plurality of interconnected processing elements is also set into a sleep state based on the first processing element being set into a sleep state. At least one circular buffer from the plurality of circular buffers remains awake while the first processing element is in the sleep state, and the at least one circular buffer provides for data steering through a reconfigurable fabric.
H03K 19/177 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion utilisant des éléments spécifiés utilisant des circuits logiques élémentaires comme composants disposés sous forme matricielle
H01L 25/00 - Ensembles consistant en une pluralité de dispositifs à semi-conducteurs ou d'autres dispositifs à l'état solide
G06F 5/08 - Procédés ou dispositions pour la conversion de données, sans modification de l'ordre ou du contenu des données maniées pour modifier la vitesse de débit des données, c.-à-d. régularisation de la vitesse ayant une séquence d'emplacements d'emmagasinage, les emplacements intermédiaires n'étant pas accessibles pour des opérations soit de mise en file d'attente, soit de retrait de file d'attente, p. ex. utilisant un registre à décalage
G06F 1/3287 - Économie d’énergie caractérisée par l'action entreprise par la mise hors tension d’une unité fonctionnelle individuelle dans un ordinateur
Techniques are disclosed for designing a reconfigurable fabric. The reconfigurable fabric is designed using logical elements, configurable connections between and among the logical elements, and rotating circular buffers. The circular buffers contain configuration instructions. The configuration instructions control connections between and among logical elements. The logical elements change operation based on the instructions that rotate through the circular buffers. Clusters of logical elements are interconnected by a switching fabric. Each cluster contains processing elements, storage elements, and switching elements. A circular buffer within a cluster contains multiple switching instructions to control the flow of data throughout the switching fabric. The circular buffer provides a pipelined execution of switching instructions for the implementation of multiple functions. Each cluster contains multiple processing elements, and each cluster further comprises an additional circular buffer for each processing element. Logical operations are controlled by the circular buffers.
H03K 19/0175 - Dispositions pour le couplageDispositions pour l'interface
H03K 19/177 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion utilisant des éléments spécifiés utilisant des circuits logiques élémentaires comme composants disposés sous forme matricielle
49.
Branchless instruction paging in reconfigurable fabric
Circular buffers containing instructions that enable the execution of operations on logical elements are described where data in the circular buffers is swapped to storage. The instructions comprise a branchless instruction set. Data stored in circular buffers is paged in and out to a second level memory. State information for each logical element is also saved and restored using paging memory. Instructions are provided to logical elements, such as processing elements, via circular buffers. The instructions enable a group of processing elements to perform operations implementing a desired functionality. That functionality is changed by updating the circular buffers with new instructions that are transferred from paging memory. The previous instructions can be saved off in paging memory before the new instructions are copied over to the circular buffers. This enables the hardware to be rapidly reconfigured amongst multiple functions.
G06F 12/00 - Accès à, adressage ou affectation dans des systèmes ou des architectures de mémoires
G06F 12/0875 - Adressage d’un niveau de mémoire dans lequel l’accès aux données ou aux blocs de données désirés nécessite des moyens d’adressage associatif, p. ex. mémoires cache avec mémoire cache dédiée, p. ex. instruction ou pile
G06F 9/38 - Exécution simultanée d'instructions, p. ex. pipeline ou lecture en mémoire
G06F 15/78 - Architectures de calculateurs universels à programmes enregistrés comprenant une seule unité centrale
G06F 3/06 - Entrée numérique à partir de, ou sortie numérique vers des supports d'enregistrement
G06F 12/08 - Adressage ou affectationRéadressage dans des systèmes de mémoires hiérarchiques, p. ex. des systèmes de mémoire virtuelle
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
G06N 3/04 - Architecture, p. ex. topologie d'interconnexion
G06N 3/063 - Réalisation physique, c.-à-d. mise en œuvre matérielle de réseaux neuronaux, de neurones ou de parties de neurone utilisant des moyens électroniques
G06F 9/50 - Allocation de ressources, p. ex. de l'unité centrale de traitement [UCT]
A method, computer program product, and computer system for physical memory allocation of a computer system, the method including collecting computer system architecture specifications, a configuration, and user requirements, identifying a plurality of memory intervals to be allocated, based on the computer system architecture specification, the configuration, and the user requirements, grouping memory intervals into a plurality of color groups, wherein each memory interval within each of the plurality of color groups comprise identical memory attributes, dividing memory into sets of memory segments, wherein each set of memory segment is assigned a color of the plurality of color groups, allocating a memory interval of the plurality of memory intervals within the set of memory segments of corresponding color, and selecting a page size for a translation of a memory interval of the plurality of memory intervals, depending upon the allocation of the memory interval and the sets of memory segments.
Techniques are disclosed for managing data within a reconfigurable computing environment. In a multiple processing element environment, such as a mesh network or other suitable topology, there is an inherent need to pass data between processing elements. Subtasks are divided among multiple processing elements. The output resulting from the subtasks is then merged by a downstream processing element. In such cases, a join operation can be used to combine data from multiple upstream processing elements. A control agent executes on each processing element. A memory buffer is disposed between upstream processing elements and the downstream processing element. The downstream processing element is configured to automatically perform an operation based on the availability of valid data from the upstream processing elements.
H04L 12/861 - Mise en mémoire tampon de paquets ou mise en file d’attente; Ordonnancement de file d’attente
H04L 12/863 - Ordonnancement de file d’attente, p.ex. ordonnancement circulaire
H04L 12/801 - Commande de flux ou commande de congestion
H04L 12/859 - Actions liées à la commande de flux basée sur la nature de l’application, p.ex. contrôle de navigation sur l’Internet ou contrôle du trafic de courrier électronique
H04L 12/879 - Opérations simples sur la mémoire-tampon, p.ex. pointeurs de mémoire-tampon ou descripteurs de mémoire-tampon
H04L 12/939 - Dispositions pour la commutation redondante, p.ex. utilisant des plans de commutation parallèles
H04L 12/933 - Cœur de commutateur, p.ex. barres croisées, mémoire partagée ou support partagé
52.
Reconfigurable processor fabric implementation using satisfiability analysis
Disclosed techniques utilize a satisfiability solver for allocation and/or configuration of resources in a reconfigurable fabric of processing elements. A dataflow graph is an input provided to a toolchain that includes a satisfiability solver. The satisfiability solver operates on subsets of interconnected nodes within a dataflow graph to derive a solution. The solution is trimmed by removing artifacts and unnecessary parts. The solutions of subsets are then used as an input to additional subsets of nodes within the dataflow graph in an iterative process to derive a complete solution. The satisfiability solver technique uses adaptive windowing in both the time dimension and the spatial dimensions of the dataflow graph. Processing elements and routing elements within the reconfigurable fabric are configured based on the complete solution. Data computation is performed based on the dataflow graph using the processing elements and the routing resources.
An apparatus for mathematical manipulation is described allowing the selective combination of shifters to shift binary numbers of various widths. Selective combination allows on-the-fly adjustment of shifters from independent to coordinated shifting operations. Selective combination allows adjustable hardware-based shifting while saving space and resources. Multiple eight-bit shifters can be configured for a variety of operand widths, such as a 32-bit width, a 24-bit width, a 16-bit width, or an eight-bit width. Multiplexers route the appropriate input data to the appropriate shifters. Bidirectional shifting is configured through a selector tree, including both shift left and shift right operations. Opcodes configure the shifters for the desired type of shift and a shifted result is generated.
G06F 5/01 - Procédés ou dispositions pour la conversion de données, sans modification de l'ordre ou du contenu des données maniées pour le décalage, p. ex. la justification, le changement d'échelle, la normalisation
54.
Reconfigurable fabric direct memory access with multiple read or write elements
Techniques are disclosed for data manipulation. Data is obtained from a first switching element where the first switching element is controlled by a first circular buffer. Data is sent to a second switching element where the second switching element is controlled by a second circular buffer. Data is controlled by a third switching element that is controlled by a third circular buffer. The third switching element hierarchically controls the first switching element and the second switching element. Data is routed through a fourth switching element that is controlled by a fourth circular buffer. The circular buffers are statically scheduled. The obtaining data from a first switching element and the sending the data to a second switching element includes a direct memory access (DMA). The switching elements can operate as a master controller or as a slave device. The switching elements can comprise clusters within an asynchronous reconfigurable fabric.
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
G06F 13/16 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus de mémoire
G06F 13/42 - Protocole de transfert pour bus, p. ex. liaisonSynchronisation
55.
Communication between dataflow processing units and memories
A combination of memory units and dataflow processing units is disclosed for computation. A first memory unit is interposed between a first dataflow processing unit and a second dataflow processing unit. Operations for a dataflow graph are allocated across the first dataflow processing unit and the second dataflow processing unit. The first memory unit passes data between the first dataflow processing unit and the second dataflow processing unit to execute the dataflow graph. The first memory unit is a high bandwidth, shared memory device including a hybrid memory cube. The first dataflow processing unit and second dataflow processing unit include a plurality of circular buffers containing instructions for controlling data transfer between the first dataflow processing unit and second dataflow processing unit. Additional dataflow processing units and additional memory units are included for additional functionality and efficiency.
G06F 5/10 - Procédés ou dispositions pour la conversion de données, sans modification de l'ordre ou du contenu des données maniées pour modifier la vitesse de débit des données, c.-à-d. régularisation de la vitesse ayant une séquence d'emplacements d'emmagasinage, chacun étant individuellement accessible à la fois pour des opérations de mise en file d'attente et pour des opérations de retrait de file d'attente, p. ex. utilisant une mémoire à accès aléatoire
G06F 13/16 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus de mémoire
G06F 15/82 - Architectures de calculateurs universels à programmes enregistrés commandés par des données ou à la demande
G11C 19/00 - Mémoires numériques dans lesquelles l'information est déplacée par échelons, p. ex. registres à décalage
An execution pipeline architecture of a microprocessor employs a third-pass functional unit, for example, third-level of arithmetic logic unit (ALU) or third short-latency execution unit to execute instructions with reduced complexity and area cost of out-of-order execution. The third-pass functional unit allows instructions with long latency execution to be moved into a retire queue. The retire queue further includes the third functional unit (e.g., ALU), a reservation station and a graduate buffer. Data dependencies of dependent instructions in the retire queue is handled independently from the main pipeline.
Various embodiments of a microprocessor include a scoreboard implementation that directs the microprocessor to the location of data values. For example, the scoreboard may include individual bits that instruct the microprocessor to retrieve the data from a re-order buffer, retire queue, result bus, or register file. As a first step, the microprocessor receives an instruction indicating a process that requires data from one or more source registers. Instead of automatically retrieving the data from the register file, which is a costly process, the microprocessor may read the scoreboard to determine whether the needed data can be more cost-effectively retrieved from the re-order buffer, retire queue, or result busses. Therefore, the microprocessor can avoid costly data retrieval procedures. Additionally, the scoreboard implementation enables the microprocessor to handle limited out-of-order instructions, which improves overall performance of the microprocessor.
Methods and systems for timing analysis and optimization of asynchronous circuit designs are disclosed. Registration stages are placed between combinational logic circuits. For timing purposes, the registration stages are modified to have a duplicate set of pins. New paths are formed in the circuit for the purposes of timing analysis. The paths are analyzable by timing tools. Once the timing analysis is complete, the paths are reverted to original paths, and new devices are selected for the circuit design based on results of the timing analysis. An updated design is sent for manufacture, based on the timing analysis and optimization of the asynchronous circuit.
Techniques are disclosed for power conservation. A plurality of processing elements and a plurality of instructions are configured. The plurality of processing elements is controlled by instructions contained in a plurality of circular buffers. The plurality of processing elements can comprise a dataflow processor. A first processing element, from the plurality of interconnected processing elements, is set into a sleep state by a first instruction from the plurality of instructions. The first processing element is woken from the sleep state as a result of valid data being presented to the first processing element. A subsection of the plurality of interconnected processing elements is also set into a sleep state based on the first processing element being set into a sleep state. At least one circular buffer from the plurality of circular buffers remains awake while the first processing element is in the sleep state, and the at least one circular buffer provides for data steering through a reconfigurable fabric.
H03K 19/177 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion utilisant des éléments spécifiés utilisant des circuits logiques élémentaires comme composants disposés sous forme matricielle
H01L 25/00 - Ensembles consistant en une pluralité de dispositifs à semi-conducteurs ou d'autres dispositifs à l'état solide
G06F 7/38 - Méthodes ou dispositions pour effectuer des calculs en utilisant exclusivement une représentation numérique codée, p. ex. en utilisant une représentation binaire, ternaire, décimale
G06F 5/08 - Procédés ou dispositions pour la conversion de données, sans modification de l'ordre ou du contenu des données maniées pour modifier la vitesse de débit des données, c.-à-d. régularisation de la vitesse ayant une séquence d'emplacements d'emmagasinage, les emplacements intermédiaires n'étant pas accessibles pour des opérations soit de mise en file d'attente, soit de retrait de file d'attente, p. ex. utilisant un registre à décalage
Certain embodiments of the present disclosure support a method and apparatus for efficient multithreading on a single core microprocessor. Thread switching in the single core microprocessor presented herein is based on a reserved space in a memory allocated to each thread for storing and restoring of registers in a register file. The thread switching is achieved without full save and restore of the register file, and only those registers referenced in the memory are saved and restored during thread switching.
Embodiments of the present disclosure support hardware based thread switching in a multithreading environment. The thread switching is implemented on a multithread microprocessor by utilizing thread mailbox registers and other auxiliary registers that can be pre-programmed for hardware based thread switching. A set of mailbox registers can be allocated to each thread of a plurality of threads that can be executed in the microprocessor. A mailbox register in the set of mailbox registers comprises an identifier of a next thread of the plurality of threads to which an active thread switches based on a thread switch condition further indicated in the mailbox register. The auxiliary registers in the microprocessor can be used to configure a number of threads for simultaneous execution in the microprocessor, a priority for thread switching, and to store a program counter of each thread and states of registers of each thread.
Embodiments of the present disclosure support implementation of a Level-1 (L1) cache in a microprocessor based on independently accessed data and tag arrays. Presented implementations of L1 cache do not require any stall pipeline mechanism for stalling execution of instructions, leading to improved microprocessor performance. A data array in the cache is interfaced with one or more data index queues that comprise, upon occurrence of a conflict between at least one instruction requesting access to the data array and at least one other instruction that accessed the data array, at least one data index for accessing the data array associated with the at least one instruction. A tag array in the cache is interfaced with a tag queue that stores one or more tag entries associated with one or more data outputs read from the data array based on accessing the data array.
G06F 12/0875 - Adressage d’un niveau de mémoire dans lequel l’accès aux données ou aux blocs de données désirés nécessite des moyens d’adressage associatif, p. ex. mémoires cache avec mémoire cache dédiée, p. ex. instruction ou pile
G06F 12/084 - Systèmes de mémoire cache multi-utilisateurs, multiprocesseurs ou multitraitement avec mémoire cache partagée
G06F 12/0895 - Mémoires cache caractérisées par leur organisation ou leur structure de parties de mémoires cache, p. ex. répertoire ou matrice d’étiquettes
A code protection scheme for controlling access to a memory region in an integrated circuit includes a processor with an instruction pipeline that includes multiple processing stages. A first processing stage receives one or more instructions. A second processing stage receives address information identifying a protected memory region of the memory from the first processing stage and protection information for an identified protected memory region. The protection information indicates a protection state assigned to each protected memory region. Based on the instruction type of the received instruction and the protection information associated with a particular protected memory region, the second processing stage determines whether to enable or disable access to the particular protected memory region by the processor or other external host.
G06F 12/14 - Protection contre l'utilisation non autorisée de mémoire
G06F 21/52 - Contrôle des utilisateurs, des programmes ou des dispositifs de préservation de l’intégrité des plates-formes, p. ex. des processeurs, des micrologiciels ou des systèmes d’exploitation au stade de l’exécution du programme, p. ex. intégrité de la pile, débordement de tampon ou prévention d'effacement involontaire de données
G06F 21/71 - Protection de composants spécifiques internes ou périphériques, où la protection d'un composant mène à la protection de tout le calculateur pour assurer la sécurité du calcul ou du traitement de l’information
G06F 21/78 - Protection de composants spécifiques internes ou périphériques, où la protection d'un composant mène à la protection de tout le calculateur pour assurer la sécurité du stockage de données
G06F 21/79 - Protection de composants spécifiques internes ou périphériques, où la protection d'un composant mène à la protection de tout le calculateur pour assurer la sécurité du stockage de données dans les supports de stockage à semi-conducteurs, p. ex. les mémoires adressables directement
G06F 9/30 - Dispositions pour exécuter des instructions machines, p. ex. décodage d'instructions
G06F 9/38 - Exécution simultanée d'instructions, p. ex. pipeline ou lecture en mémoire
G06F 21/62 - Protection de l’accès à des données via une plate-forme, p. ex. par clés ou règles de contrôle de l’accès
Disclosed embodiments select a proper hum frequency reference by utilizing one or more functional logic circuits within a cluster. The slowest logic circuit is determined, and an instance of that logic circuit is used in timing circuitry for the cluster. Multiple logic circuits with similar characteristics are incorporated into the timing circuit. Each cluster is interconnected to a second level timing circuit. Each cluster inputs timing information into the second level timing circuit. The second level timing circuit then determines when the next cycle, or tic, of the self-generated clock starts, and the process repeats, providing a self-generated clock signal.
H03K 19/20 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion caractérisés par la fonction logique, p. ex. circuits ET, OU, NI, NON
H03K 5/135 - Dispositions ayant une sortie unique et transformant les signaux d'entrée en impulsions délivrées à des intervalles de temps désirés par l'utilisation de signaux de référence de temps, p. ex. des signaux d'horloge
65.
Configurable and programmable multi-core architecture with a specialized instruction set for embedded application based on neural networks
A programmable architecture specialized for convolutional neural networks (CNNs) processing such that different applications of CNNs may be supported by the presently disclosed method and apparatus by reprogramming the processing elements therein. The architecture may include an optimized architecture that provides a low-area or footprint and low-power solution desired for embedded applications while still providing the computational capabilities required for CNN applications that may be computationally intensive, requiring a huge number of convolution operations per second to process inputs such as video streams in real time.
G06N 3/063 - Réalisation physique, c.-à-d. mise en œuvre matérielle de réseaux neuronaux, de neurones ou de parties de neurone utilisant des moyens électroniques
G06N 3/04 - Architecture, p. ex. topologie d'interconnexion
66.
Techniques for creating and using a hierarchical data structure
Techniques for creating and using a hierarchical data structure, in accordance with embodiments of the present invention include storing received data as records in a first level of the hierarchical data structure. One or more parameters for each block of records in the first level are summarized and stored in a second level of the hierarchical data structure. The techniques may also include querying a given level of a hierarchical data structure. One or more blocks of records one level below the given level are accessed. Each of the accessed blocks correspond to records that are summarized by records in the given level that satisfies the query.
Clusters of logical elements are interconnected by a switching fabric. Each cluster contains processing elements, storage elements, and switching elements. A circular buffer within a cluster contains multiple switching instructions to control the flow of data throughout the switching fabric. The circular buffer provides a pipelined execution of switching instructions for the implementation of multiple functions. Each cluster contains multiple processing elements, and each cluster further comprises an additional circular buffer for each processing element. Logical operations are controlled by the circular buffers.
H03K 19/0175 - Dispositions pour le couplageDispositions pour l'interface
H03K 19/177 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion utilisant des éléments spécifiés utilisant des circuits logiques élémentaires comme composants disposés sous forme matricielle
A plurality of software programmable processors is disclosed. The software programmable processors are controlled by rotating circular buffers. A first processor and a second processor within the plurality of software programmable processors are individually programmable. The first processor within the plurality of software programmable processors is coupled to neighbor processors within the plurality of software programmable processors. The first processor sends and receives data from the neighbor processors. The first processor and the second processor are configured to operate on a common instruction cycle. An output of the first processor from a first instruction cycle is an input to the second processor on a subsequent instruction cycle.
G06F 9/38 - Exécution simultanée d'instructions, p. ex. pipeline ou lecture en mémoire
G06F 15/173 - Communication entre processeurs utilisant un réseau d'interconnexion, p. ex. matriciel, de réarrangement, pyramidal, en étoile ou ramifié
G06F 15/82 - Architectures de calculateurs universels à programmes enregistrés commandés par des données ou à la demande
G06F 1/324 - Économie d’énergie caractérisée par l'action entreprise par réduction de la fréquence d’horloge
Techniques for modifying a circuit are described herein. In some examples, a method includes generating a set of testing data and detecting a predetermined modification to a translation path corresponding to a memory address mapping, the predetermined modification to change a physical memory address of the testing data associated with a virtual memory address to a second physical memory address of the testing data. The method can also include generating a test template comprising a first instruction to implement the predetermined modification and a second instruction comprising the second physical memory address in the translation path and transmitting the test template to the circuit for each of a plurality of software instruction threads. Furthermore, the method can include detecting a defect in the execution of the test template by the circuit and modifying the circuit to prevent the defect during execution of the test template.
G06F 9/44 - Dispositions pour exécuter des programmes spécifiques
G06F 13/10 - Commande par programme pour dispositifs périphériques
G06F 11/00 - Détection d'erreursCorrection d'erreursContrôle de fonctionnement
G01R 31/28 - Test de circuits électroniques, p. ex. à l'aide d'un traceur de signaux
G01R 27/28 - Mesure de l'atténuation, du gain, du déphasage ou des caractéristiques qui en dérivent dans des réseaux électriques quadripoles, c.-à-d. des réseaux à double entréeMesure d'une réponse transitoire
G06F 12/1027 - Traduction d'adresses utilisant des moyens de traduction d’adresse associatifs ou pseudo-associatifs, p. ex. un répertoire de pages actives [TLB]
G06F 17/30 - Recherche documentaire; Structures de bases de données à cet effet
G06F 9/455 - ÉmulationInterprétationSimulation de logiciel, p. ex. virtualisation ou émulation des moteurs d’exécution d’applications ou de systèmes d’exploitation
70.
Adaptive content dependent intra prediction mode coding
A system and a method are disclosed for encoding and decoding a video frame using spatial prediction. The video frame is separated into a plurality of image blocks, and a plurality of spatial predictors is created for an image block using methods well-known in the art. The set of predictors is reduced to a set containing fewer spatial predictors before continuing the coding process for the block. The reduction of spatial predictors involves comparing a plurality of spatial predictors in a predetermined sequence and grouping or processing a subset of the spatial predictors to a set of representative spatial predictors.
H04N 19/11 - Sélection du mode de codage ou du mode de prédiction parmi plusieurs modes de codage prédictif spatial
H04N 19/176 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant le codage adaptatif caractérisés par l’unité de codage, c.-à-d. la partie structurelle ou sémantique du signal vidéo étant l’objet ou le sujet du codage adaptatif l’unité étant une zone de l'image, p. ex. un objet la zone étant un bloc, p. ex. un macrobloc
H04N 19/159 - Type de prédiction, p. ex. prédiction intra-trame, inter-trame ou de trame bidirectionnelle
H04N 19/593 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant le codage prédictif mettant en œuvre des techniques de prédiction spatiale
H04N 19/147 - Débit ou quantité de données codées à la sortie du codeur selon des critères de débit-distorsion
A code protection scheme for controlling access to a memory region in an integrated circuit includes a processor with an instruction pipeline that includes multiple processing stages. A first processing stage receives one or more instructions. A second processing stage receives address information identifying a protected memory region of the memory from the first processing stage and protection information for an identified protected memory region. The protection information indicates a protection state assigned to each protected memory region. Based on the instruction type of the received instruction and the protection information associated with a particular protected memory region, the second processing stage determines whether to enable or disable access to the particular protected memory region by the processor or other external host.
G06F 12/14 - Protection contre l'utilisation non autorisée de mémoire
G06F 21/52 - Contrôle des utilisateurs, des programmes ou des dispositifs de préservation de l’intégrité des plates-formes, p. ex. des processeurs, des micrologiciels ou des systèmes d’exploitation au stade de l’exécution du programme, p. ex. intégrité de la pile, débordement de tampon ou prévention d'effacement involontaire de données
G06F 21/71 - Protection de composants spécifiques internes ou périphériques, où la protection d'un composant mène à la protection de tout le calculateur pour assurer la sécurité du calcul ou du traitement de l’information
G06F 21/78 - Protection de composants spécifiques internes ou périphériques, où la protection d'un composant mène à la protection de tout le calculateur pour assurer la sécurité du stockage de données
G06F 21/79 - Protection de composants spécifiques internes ou périphériques, où la protection d'un composant mène à la protection de tout le calculateur pour assurer la sécurité du stockage de données dans les supports de stockage à semi-conducteurs, p. ex. les mémoires adressables directement
G06F 9/30 - Dispositions pour exécuter des instructions machines, p. ex. décodage d'instructions
G06F 9/38 - Exécution simultanée d'instructions, p. ex. pipeline ou lecture en mémoire
G06F 21/62 - Protection de l’accès à des données via une plate-forme, p. ex. par clés ou règles de contrôle de l’accès
72.
System and method of debugging multi-threaded processes
A system and method of debugging a multi-threaded process with at least one running thread and at least one suspended thread is disclosed. Embodiments utilize a blocking function to block the thread of a process while other threads are allowed to run. The blocking function may be executed in a suspended thread by a debugger under control of a thread blocking controller. The other threads may implement interprocess communication channels for enabling communication between the process and another application. A simulated user interface (UI) of a debugger enables interaction with users while a hardware simulation thread is blocked, where blocking of the hardware simulation thread may be implemented by a thread blocking component implemented externally to the debugger. Where a thread blocking controller is implemented within the debugger, a debugger UI may interact with a user while the hardware simulation thread is blocked and interprocess communication threads are running.
The invention relates to methods of simulation of a plurality of processors running on a plurality of cores, to multi-core microprocessor systems in which such methods may be carried out, and to computer program products configured to perform a simulation of a plurality of processors, running on a plurality of cores. According to a first aspect of the invention, there is provided a method of running a plurality of simulated processors on a plurality of cores, in which simulation of the processors is performed in parallel on the plurality of cores.
Circular buffers containing instructions that enable the execution of operations on logical elements are described where data in the circular buffers is swapped to storage. Data stored in circular buffers is paged in and out to a second level memory. State information for each logical element is also saved and restored using paging memory. Logical elements such as processing elements are provided instructions via circular buffers. The instructions enable a group of processing elements to perform operations implementing a desired functionality. That functionality is changed by updating the circular buffers with new instructions that are transferred from paging memory. The previous instructions can be saved off in paging memory before the new instructions are copied over to the circular buffers. This enables the hardware to be rapidly reconfigured amongst multiple functions.
G06F 12/00 - Accès à, adressage ou affectation dans des systèmes ou des architectures de mémoires
G06F 12/0802 - Adressage d’un niveau de mémoire dans lequel l’accès aux données ou aux blocs de données désirés nécessite des moyens d’adressage associatif, p. ex. mémoires cache
G06F 12/0868 - Transfert de données entre une mémoire cache et d'autres sous-systèmes, p. ex. des dispositifs de stockage ou des systèmes hôtes
75.
Visual quality measure for real-time video processing
A measure of visual quality of processed images relative to unprocessed images is generated in real-time. The measure of visual quality closely correlates with a human's actual perception of the processed image relative to the original image. The measure of visual quality is computed based on a measure of discrepancy (e.g., mean square errors) between the processed and unprocessed images and the variance of each image in the pixel domain or the transform domain may be determined. If the processed image is unavailable, a prediction of the processed image may be used in place of the processed image. The prediction of a processed image may involve predicting the variance values for processed image blocks. The visual quality measure may be used in a feedback loop to improve processing or encoding.
H04N 7/12 - Systèmes dans lesquels le signal de télévision est transmis par un canal ou une pluralité de canaux parallèles, la bande passante de chaque canal étant inférieure à la largeur de bande du signal de télévision
H04N 19/154 - Qualité visuelle après décodage mesurée ou estimée de façon subjective, p. ex. mesure de la distorsion
H04N 19/176 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant le codage adaptatif caractérisés par l’unité de codage, c.-à-d. la partie structurelle ou sémantique du signal vidéo étant l’objet ou le sujet du codage adaptatif l’unité étant une zone de l'image, p. ex. un objet la zone étant un bloc, p. ex. un macrobloc
H04N 19/147 - Débit ou quantité de données codées à la sortie du codeur selon des critères de débit-distorsion
H04N 19/103 - Sélection du mode de codage ou du mode de prédiction
H04N 19/60 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant un codage par transformée
76.
Compact logic evaluation gates using null convention
Compact logic evaluation gates are built using null convention logic (NCL) circuits. The inputs to a null convention circuit include a NCL true input and a NCL complement input. The NCL circuit includes a gate coupled to the pair of inputs, where the gate comprises a plurality of transistors. The transistors allow for logical signal capture, provide a pair of cross-coupled inverters for data storage, and include a first and second pull-down device. The first pull-down device causes a first side of the pair of cross-coupled inverters to go to a “0” state when a “1” is applied to the NCL true input, and the second pull-down device causes a second side of the pair of cross-coupled inverters to go to a “0” state when a “1” is applied to the NCL complement input.
H03K 19/00 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion
H03K 19/094 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion utilisant des éléments spécifiés utilisant des dispositifs à semi-conducteurs utilisant des transistors à effet de champ
77.
Computing resource allocation based on flow graph translation
Systems and methods are disclosed for computing resource allocation based on flow graph translation. First, a high-level description of logic circuitry is obtained and translated to generate a flow graph representing sequential operations. Using the flow graph, similar processing elements in an array are interchangeably allocated to perform computational, communication, and storage tasks as needed. The sequential operations are executed using the array of interchangeable processing elements. Data is provided from the storage elements through the communication elements to the computational elements. Computational results are stored in the storage elements. Outputs from some of the computational elements provide inputs to other computational elements. Execution of the instructions can be controlled with time stepping. The processors are reallocated as needed, based on changes to the flow graph.
A system for simulating hardware is disclosed wherein stopping the simulating for investigating a state of the simulating is constrained to occur at points of simulation synchronization. A delta cycle may be inserted into a simulation to provide a point of simulation synchronization. The delta cycle may be inserted at the beginning of a simulation loop.
Generating a description of compiler code selector rules from an architecture description. A method comprises accessing a target architecture model written in an architecture description language (ADL) and extracting semantic information therefrom to generate a plurality of semantic statements. Rules that map from source code operations to semantic patterns are accessed. The semantic statements are searched for matches for the semantic patterns to generate mappings that serve as a description of compiler code selector rules.
Multi-threshold flash Null Convention Logic (NCL) includes one or more high threshold voltage transistors within a flash NCL gate to reduce power consumption due to current leakage by transistors of the NCL gate. High-threshold voltage transistors may be added and/or may be used in place of one or more lower voltage threshold transistors of the NCL gate. A high-Vt device is included in the pull-up path to reduce power when the flash NCL logic gate is in the null state.
H03K 19/00 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion
H03K 19/094 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion utilisant des éléments spécifiés utilisant des dispositifs à semi-conducteurs utilisant des transistors à effet de champ
H03K 19/23 - Circuits de majorité ou de minorité, c.-à-d. donnant un signal de sortie dont l'état est celui de la majorité ou de la minorité des signaux d'entrée
H03K 23/40 - Signaux d'ouverture de porte ou d'horloge appliqués à tous les étages, c.-à-d. compteurs synchrones
H03K 23/58 - Signaux d'ouverture de porte ou d'horloge non appliqués à tous les étages, c.-à-d. compteurs asynchrones
A code protection scheme for controlling access to a memory region in an integrated circuit includes a processor with an instruction pipeline that includes multiple processing stages. A first processing stage receives one or more instructions. A second processing stage receives address information identifying a protected memory region of the memory from the first processing stage and protection information for an identified protected memory region. The protection information indicates a protection state assigned to each protected memory region. Based on the instruction type of the received instruction and the protection information associated with a particular protected memory region, the second processing stage determines whether to enable or disable access to the particular protected memory region by the processor or other external host.
G06F 12/14 - Protection contre l'utilisation non autorisée de mémoire
G06F 21/52 - Contrôle des utilisateurs, des programmes ou des dispositifs de préservation de l’intégrité des plates-formes, p. ex. des processeurs, des micrologiciels ou des systèmes d’exploitation au stade de l’exécution du programme, p. ex. intégrité de la pile, débordement de tampon ou prévention d'effacement involontaire de données
G06F 21/71 - Protection de composants spécifiques internes ou périphériques, où la protection d'un composant mène à la protection de tout le calculateur pour assurer la sécurité du calcul ou du traitement de l’information
G06F 21/78 - Protection de composants spécifiques internes ou périphériques, où la protection d'un composant mène à la protection de tout le calculateur pour assurer la sécurité du stockage de données
G06F 21/79 - Protection de composants spécifiques internes ou périphériques, où la protection d'un composant mène à la protection de tout le calculateur pour assurer la sécurité du stockage de données dans les supports de stockage à semi-conducteurs, p. ex. les mémoires adressables directement
G06F 9/30 - Dispositions pour exécuter des instructions machines, p. ex. décodage d'instructions
G06F 9/38 - Exécution simultanée d'instructions, p. ex. pipeline ou lecture en mémoire
G06F 21/62 - Protection de l’accès à des données via une plate-forme, p. ex. par clés ou règles de contrôle de l’accès
System and method for generating hardware accelerators and processor offloads. System for hardware acceleration. System and method for implementing an asynchronous offload. Method of automatically creating a hardware accelerator. Computerized method for automatically creating a test harness for a hardware accelerator from a software program. System and method for interconnecting hardware accelerators and processors. System and method for interconnecting a processor and a hardware accelerator. Computer implemented method of generating a hardware circuit logic block design for a hardware accelerator automatically from software. Computer program and computer program product stored on tangible media implementing the methods and procedures of the invention.
Systems and methods of user interface for facilitation of high level generation of processor extensions. In accordance with a method embodiment of the present invention, an instruction format is accessed at a graphical user interface. A programming language description of a computation element for an execution unit of the processor extension is accessed. A representation of a hardware design for the processor extension comprising the instruction format and the computation element is generated.
Clusters of logical elements are interconnected by a switching fabric. Each cluster contains processing elements, storage elements, and switching elements. A circular buffer within a cluster contains multiple switching instructions to control the flow of data throughout the switching fabric. The circular buffer provides a pipelined execution of switching instructions. Each cluster contains multiple processing elements, and each cluster further comprises an additional circular buffer for each processing element. Logical operations are controlled by the circular buffers.
H03K 19/0175 - Dispositions pour le couplageDispositions pour l'interface
H03K 19/177 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion utilisant des éléments spécifiés utilisant des circuits logiques élémentaires comme composants disposés sous forme matricielle
85.
Self-timed user-extension instructions for a processing device
A processor for executing configurable instructions and a method of configuring the processor are disclosed. In one embodiment, the processor includes (i) a processor core to execute preconfigured instructions and (ii) a processor core extension to execute user-defined extension instructions that are configurable instructions. The user-defined extension instructions may include an autonomous instruction with varying execution cycles based on source data and an operation performed. The processor core extension employs extension interface signals as a handshake protocol to operate together with the processor core without knowing any priori knowledge of how many processor clock cycles that the autonomous instruction will take to complete.
G06F 15/76 - Architectures de calculateurs universels à programmes enregistrés
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
G06F 9/30 - Dispositions pour exécuter des instructions machines, p. ex. décodage d'instructions
G06F 3/0484 - Techniques d’interaction fondées sur les interfaces utilisateur graphiques [GUI] pour la commande de fonctions ou d’opérations spécifiques, p. ex. sélection ou transformation d’un objet, d’une image ou d’un élément de texte affiché, détermination d’une valeur de paramètre ou sélection d’une plage de valeurs
A processor includes an instruction fetch unit and an execution unit. The instruction fetch unit retrieves instructions from memory to be executed by the execution unit. The instruction fetch unit includes a branch prediction unit which is configured to predict whether a branch instruction is likely to be executed. The memory includes an instruction cache comprising a portion of the fetch blocks available in the memory. The instruction fetch unit may use a combination of way prediction and serialized access to retrieve instructions from the instruction cache. The instruction fetch unit initially accesses the instruction cache to retrieve the predicted fetch block associated with a way prediction. The instruction fetch unit compares a cache tag associated with the way prediction with the address of the cache line that includes the predicted fetch block. If the tag matches, then the way prediction is correct and the retrieved fetch block is valid.
A processor uses a prediction unit to predict subsequent instructions of a program to be executed by the processor. Many implementations or combinations of implementations may be used to predict the subsequent instruction of the program. In one embodiment, a branch cache is used to store branch information. A prediction table is used to store prediction information based on the branch. A prediction logic module determines whether a branch is taken or not taken based on the branch information stored in the branch cache and the prediction information stored in the prediction table.
G06F 7/38 - Méthodes ou dispositions pour effectuer des calculs en utilisant exclusivement une représentation numérique codée, p. ex. en utilisant une représentation binaire, ternaire, décimale
G06F 9/30 - Dispositions pour exécuter des instructions machines, p. ex. décodage d'instructions
G06F 9/38 - Exécution simultanée d'instructions, p. ex. pipeline ou lecture en mémoire
G06F 12/08 - Adressage ou affectationRéadressage dans des systèmes de mémoires hiérarchiques, p. ex. des systèmes de mémoire virtuelle
88.
Method and processor for reducing code and latency of TLB maintenance operations in a configurable processor
A memory management unit (MMU) is disclosed for storing mappings between virtual addresses and physical addresses. The MMU includes a translation look-aside buffer (TLB) and a memory management unit controller. The TLB stores mappings between a virtual address and a physical address. The MMU controller receives a request to insert an entry into the TLB and performs a set of operations based on the received request. The MMU controller determines whether an entry stored in the TLB is associated with the virtual address of the request, removes the entry stored in the TLB that is associated with the virtual address and inserts the requested entry into the TLB.
Multiple threshold voltage circuitry based on silicon-on-insulator (SOI) technology is disclosed which utilizes N-wells and/or P-wells underneath the insulator in SOI FETs. The well under a FET is biased to influence the threshold voltage of the FET. A PFET and an NFET share a common buried P-well or N-well. Various types of logic can be fabricated in silicon-on-insulator (SOI) technology using multiple threshold voltage FETs. Embodiments provide circuits including the advantageous properties of both low-leakage transistors and high-speed transistors.
H01L 27/10 - Dispositifs consistant en une pluralité de composants semi-conducteurs ou d'autres composants à l'état solide formés dans ou sur un substrat commun comprenant des éléments de circuit passif intégrés avec au moins une barrière de potentiel ou une barrière de surface le substrat étant un corps semi-conducteur comprenant une pluralité de composants individuels dans une configuration répétitive
H03K 19/0948 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion utilisant des éléments spécifiés utilisant des dispositifs à semi-conducteurs utilisant des transistors à effet de champ utilisant des transistors MOSFET utilisant des dispositifs CMOS
H01L 27/12 - Dispositifs consistant en une pluralité de composants semi-conducteurs ou d'autres composants à l'état solide formés dans ou sur un substrat commun comprenant des éléments de circuit passif intégrés avec au moins une barrière de potentiel ou une barrière de surface le substrat étant autre qu'un corps semi-conducteur, p.ex. un corps isolant
H01L 21/84 - Fabrication ou traitement de dispositifs consistant en une pluralité de composants à l'état solide ou de circuits intégrés formés dans ou sur un substrat commun avec une division ultérieure du substrat en plusieurs dispositifs individuels pour produire des dispositifs, p.ex. des circuits intégrés, consistant chacun en une pluralité de composants le substrat étant autre chose qu'un corps semi-conducteur, p.ex. étant un corps isolant
H03K 19/08 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion utilisant des éléments spécifiés utilisant des dispositifs à semi-conducteurs
H03K 19/20 - Circuits logiques, c.-à-d. ayant au moins deux entrées agissant sur une sortieCircuits d'inversion caractérisés par la fonction logique, p. ex. circuits ET, OU, NI, NON
90.
Method and system for instruction set simulation with concurrent attachment of multiple debuggers
An ISS (instruction set simulator) debugging method. The method includes executing a plurality of instructions of an ISS (instruction set simulator). For each instruction, an ISS simulation state is synchronized to a simulation time when the execution of the instruction is finished. At least two debugger components are executed in conjunction with the execution of the ISS, wherein the debugger components are each coupled to control the ISS.
A system and method of debugging a multi-threaded process with at least one running thread and at least one suspended thread is disclosed. Embodiments utilize a blocking function to block the thread of a process while other threads are allowed to run. The blocking function may be executed in a suspended thread by a debugger under control of a thread blocking controller. The other threads may implement interprocess communication channels for enabling communication between the process and another application. A simulated user interface (UI) of a debugger enables interaction with users while a hardware simulation thread is blocked, where blocking of the hardware simulation thread may be implemented by a thread blocking component implemented externally to the debugger. Where a thread blocking controller is implemented within the debugger, a debugger UI may interact with a user while the hardware simulation thread is blocked and interprocess communication threads are running.
A processing device is provided. A cluster includes a plurality of groups of processing elements. A multi-word device is connected to the processing elements within the groups. Each processing element in a particular group is in communication with all other processing elements within the particular group, and only one of the processing elements within other groups in the cluster. Each processing element is limited to operations in which input bits can be processed and an output obtained without reference to other bits. The multi-word device is configured to cooperate with at least two other processing elements to perform processing that requires reference to other bits to obtain a result.
G06F 9/38 - Exécution simultanée d'instructions, p. ex. pipeline ou lecture en mémoire
G06F 9/30 - Dispositions pour exécuter des instructions machines, p. ex. décodage d'instructions
G06F 15/173 - Communication entre processeurs utilisant un réseau d'interconnexion, p. ex. matriciel, de réarrangement, pyramidal, en étoile ou ramifié
93.
Structured block transfer module, system architecture, and method for transferring
Structured block transfer module, system architecture, and method for transferring content or data. Circuit allowing content in one memory to be shifted, moved, or copied to another memory with no direction from a host, the circuit comprising: a connection manager, at least one copy engine, and a connection between the connection manager and the copy engine. Method for transferring the contents of one of a number of blocks of source memory to one of a number of possible destination memories comprising: selecting source memory; selecting available destination memory; marking the selected destination as no longer available; copying contents of selected source memory into selected destination memory; and marking selected source as available.
G06F 3/06 - Entrée numérique à partir de, ou sortie numérique vers des supports d'enregistrement
G06F 13/362 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus ou au système à bus communs avec commande d'accès centralisée
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
An apparatus for mathematical manipulation is described allowing the selective combination of shifters to shift binary numbers of various widths. Selective combination allows on-the-fly adjustment of shifters from independent to coordinated shifting operations. Selective combination allows adjustable hardware-based shifting while saving space and resources. Multiple eight-bit shifters can be configured for a variety of operand widths, such as a 32-bit width, a 24-bit width, a 16-bit width, or an eight-bit width. Multiplexers route the appropriate input data to the appropriate shifters. Opcodes configure the shifters for the desired type of shift and a shifted result is generated.
G06F 5/01 - Procédés ou dispositions pour la conversion de données, sans modification de l'ordre ou du contenu des données maniées pour le décalage, p. ex. la justification, le changement d'échelle, la normalisation
95.
Structured block transfer module, system architecture, and method for transferring
Structured block transfer module, system architecture, and method for transferring content or data. Circuit allowing content in one memory to be shifted, moved, or copied to another memory with no direction from a host., the circuit comprising: a connection manager, at least one copy engine, and a connection between the connection manager and the copy engine. Method for transferring the contents of one of a number of blocks of source memory to one of a number of possible destination memories comprising: selecting source memory; selecting available destination memory; marking the selected destination as no longer available; copying contents of selected source memory into selected destination memory; and marking selected source as available.
G06F 13/362 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus ou au système à bus communs avec commande d'accès centralisée
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
G06F 3/06 - Entrée numérique à partir de, ou sortie numérique vers des supports d'enregistrement
An extensible iterative multiplier design is provided. Embodiments provide cascaded 8-bit multipliers for simplifying the performance of multi-byte multiplications. Booth encoding is performed in the lowest order multiplier, with the result of the Booth encoding then provided to higher order multipliers. Additionally, multiply-add operations can be performed by initializing a partial product sum register. Configurable connections between the multipliers facilitate a variety of possible multiplication options, including the possibility of varying the width of the operands.
Structured block transfer module, system architecture, and method for transferring content or data. Circuit allowing content in one memory to be shifted, moved, or copied to another memory with no direction from a host, the circuit comprising: a connection manager, at least one copy engine, and a connection between the connection manager and the copy engine. Method for transferring the contents of one of a number of blocks of source memory to one of a number of possible destination memories comprising: selecting source memory; selecting available destination memory; marking the selected destination as no longer available; copying contents of selected source memory into selected destination memory; and marking selected source as available.
G06F 12/00 - Accès à, adressage ou affectation dans des systèmes ou des architectures de mémoires
G06F 13/00 - Interconnexion ou transfert d'information ou d'autres signaux entre mémoires, dispositifs d'entrée/sortie ou unités de traitement
G06F 13/28 - Gestion de demandes d'interconnexion ou de transfert pour l'accès au bus d'entrée/sortie utilisant le transfert par rafale, p. ex. acces direct à la mémoire, vol de cycle
A method of automatically extracting information from an architecture description. A memory resident directed acyclic graph data structure comprising nodes representing instructions and edges whose weights represent dependencies between pairs of instructions is constructed. A list of ready nodes are maintained in the directed acyclic graph. A list of nodes not scheduled is maintained. And, it is determined whether the next instruction to be scheduled is to be taken from the list of ready nodes or from the list of nodes not yet scheduled.
A system and method of debugging a multi-threaded process with at least one running thread and at least one suspended thread is disclosed. Embodiments utilize a blocking function to block the thread of a process while other threads are allowed to run. The blocking function may be executed in a suspended thread by a debugger under control of a thread blocking controller. The other threads may implement interprocess communication channels for enabling communication between the process and another application. A simulated user interface (UI) of a debugger enables interaction with users while a hardware simulation thread is blocked, where blocking of the hardware simulation thread may be implemented by a thread blocking component implemented externally to the debugger. Where a thread blocking controller is implemented within the debugger, a debugger UI may interact with a user while the hardware simulation thread is blocked and interprocess communication threads are running.
A computer implemented method of generating a compiler description from an architecture description. Information is automatically extracted from an architecture description that is usable in a description of an architecture described by the architecture description. The extracted information is imported into a program comprising a graphical user interface that accepts user provided additional information that is usable in the compiler description. User provided additional information is accessed that is usable in the compiler description. A compiler description is automatically generated for the architecture described by the architecture description, based on the automatically extracted information and the accessed user provided additional information.
brevets