Abstract: A field programmable gate array (FPGA) including a configurable interconnect fabric connecting a plurality of logic blocks, the configurable interconnect fabric and the logic blocks being configured to implement a data masking circuit configured to: receive input data including data values at a plurality of indices of the input data; select between a data value of the data values and an alternative value using a masking multiplexer to generate masked data, the masking multiplexer being controlled by a mask value of a plurality of mask values at indices corresponding to the indices of the input data; and output the masked data. In some examples, the configurable interconnect fabric and the logic blocks are further configured to implement a mask generation circuit configured to generate the mask values. In some examples, the mask values are received from external memory.

Abstract: Embodiments of the present disclosure include systems and methods for providing hierarchical and shared exponent floating point data types. First and second shared exponent values are determined based on exponent values of a plurality of floating point values. A third shared exponent value is determined based the first shared exponent value and the second shared exponent value. First and second difference values are determined based on the first shared exponent value, the second shared exponent value, and the third shared exponent value. Sign values and mantissa values are determined for the plurality of floating point values. The sign value and the mantissa value for each floating point value in the plurality of floating point values, the third shared exponent value, the first difference value, and the second difference value are stored in a data structure for a shared exponent floating point data type.

Abstract: A field programmable gate array including a configurable interconnect fabric connecting logic blocks implementing a circuit to: receive input data including data values organized into rows and columns, each row having N data values; select R[i] unmasked data values of a row of the input data in accordance with a mask and an index i of the row; select N?[i] unmasked data values of another row of the input data in accordance with the mask and an index of the another row; merge the R[i] unmasked data values of the row and the N?[i] data values of the another row into a combined data vector of N data values; and compute R[i] normalized values based on the R[i] unmasked data values of the combined data vector and N?[i] normalized values based on the N?[i] data values of the combined data vector to generate N normalized data values.

Abstract: Systems and methods related to hardware-assisted gradient optimization using streamed gradients are described. An example method in a system comprising a memory configured to store weights associated with a neural network model comprising L layers, where L is an integer greater than one, a gradient optimizer, and a plurality of workers is described. The method includes during a single burst cycle moving a first set of gradients, received from each of the plurality of workers, from at least one gradient buffer to the gradient optimizer and moving weights from at least one buffer, coupled to the memory, to the gradient optimizer. The method further includes during the single burst cycle writing back the new weights, calculated by the gradient optimizer, to the memory. The method further includes during the single burst cycle transmitting the new weights, from the gradient optimizer, to each of the plurality of workers.

Abstract: Systems and methods related to dual-momentum gradient optimization with reduced memory requirements are described. An example method in a system comprising a gradient optimizer and a memory configured to store momentum values associated with a neural network model comprising L layers is described. The method includes retrieving from the memory a first set of momentum values and a second set of momentum values, corresponding to a layer of the neural network model, having a selected storage format. The method further includes converting the first set of momentum values to a third set of momentum values having a training format associated with the gradient optimizer and converting the second set of momentum values to a fourth set of momentum values having a training format associated with the gradient optimizer. The method further includes performing gradient optimization using the third set of momentum values and the fourth set of momentum values.

Abstract: Embodiments of the present disclosure include systems and methods for fusing operators for neural network hardware accelerators. A plurality of vector multiplication operations in a data path of a mapping function included in a neural network are identified. The plurality of vector multiplication operations are combined into a single vector multiplication operation in the data path of the mapping function.

Abstract: A field programmable gate array (FPGA) including a configurable interconnect fabric connecting a plurality of logic blocks, the configurable interconnect fabric and the logic blocks being configured to implement a data masking circuit configured to: receive input data including data values at a plurality of indices of the input data; select between a data value of the data values and an alternative value using a masking multiplexer to generate masked data, the masking multiplexer being controlled by a mask value of a plurality of mask values at indices corresponding to the indices of the input data; and output the masked data. In some examples, the configurable interconnect fabric and the logic blocks are further configured to implement a mask generation circuit configured to generate the mask values. In some examples, the mask values are received from external memory.

Abstract: Systems and methods related to hardware-assisted gradient optimization using streamed gradients are described. An example method in a system comprising a memory configured to store weights associated with a neural network model comprising L layers, where L is an integer greater than one, a gradient optimizer, and a plurality of workers is described. The method includes during a single burst cycle moving a first set of gradients, received from each of the plurality of workers, from at least one gradient buffer to the gradient optimizer and moving weights from at least one buffer, coupled to the memory, to the gradient optimizer. The method further includes during the single burst cycle writing back the new weights, calculated by the gradient optimizer, to the memory. The method further includes during the single burst cycle transmitting the new weights, from the gradient optimizer, to each of the plurality of workers.

Abstract: Embodiments of the present disclosure include an error recovery method comprising detecting a computing error, restarting a first artificial intelligence processor of a plurality of artificial intelligence processors processing a data set, and loading a model in the artificial intelligence processor, wherein the model corresponds to a same model processed by the plurality of artificial intelligence processors during a previous processing iteration by the plurality of artificial intelligence processors on data from the data set.

Abstract: Systems and methods related to dual-momentum gradient optimization with reduced memory requirements are described. An example method in a system comprising a gradient optimizer and a memory configured to store momentum values associated with a neural network model comprising L layers is described. The method includes retrieving from the memory a first set of momentum values and a second set of momentum values, corresponding to a layer of the neural network model, having a selected storage format. The method further includes converting the first set of momentum values to a third set of momentum values having a training format associated with the gradient optimizer and converting the second set of momentum values to a fourth set of momentum values having a training format associated with the gradient optimizer. The method further includes performing gradient optimization using the third set of momentum values and the fourth set of momentum values.

Abstract: Aspects of embodiments of the present disclosure relate to a field programmable gate array (FPGA) configured to implement an exponential function data path including: an input scaling stage including constant shifters and integer adders to scale a mantissa portion of an input floating-point value by approximately log2 e to compute a scaled mantissa value, where e is Euler's number; and an exponential stage including barrel shifters and an exponential lookup table to: extract an integer portion and a fractional portion from the scaled mantissa value based on the exponent portion of the input floating-point value; apply a bias shift to the integer portion to compute a result exponent portion of a result floating-point value; lookup a result mantissa portion of the result floating-point value in the exponential lookup table based on the fractional portion; and combine the result exponent portion and the result mantissa portion to generate the result floating-point value.

Abstract: Systems, methods, and apparatuses are provided for compressing values. A plurality of parameters may be obtained from a memory, each parameter comprising a floating-point number that is used in a relationship between artificial neurons or nodes in a model. A mantissa value and an exponent value may be extracted from each floating-point number to generate a set of mantissa values and a set of exponent values. The set of mantissa values may be compressed to generate a mantissa lookup table (LUT) and a plurality of mantissa LUT index values. The set of exponent values may be encoded to generate an exponent LUT and a plurality of exponent LUT index values. The mantissa LUT, mantissa LUT index values, exponent LUT, and exponent LUT index values may be provided to one or more processing entities to train the model.

Abstract: Methods, systems, apparatuses, and computer program products are described herein that enable execution of a large AI model on a memory-constrained target device that is communicatively connected to a parameter server, which stores a master copy of the AI model. The AI model may be dissected into smaller portions (e.g., layers or sub-layers), and each portion may be executed as efficiently as possible on the target device. After execution of one portion of the AI model is finished, another portion of the AI model may be downloaded and executed at the target device. To improve efficiency, the input samples may be divided into microbatches, and a plurality of microbatches executing in sequential order may form a minibatch. The size of the group of microbatches or minibatch can be adjusted to reduce the communication overhead. Multi-level parallel parameters reduction may be performed at the parameter server and the target device.

Abstract: Methods, systems, apparatuses, and computer program products are described herein that enable execution of a large AI model on a memory-constrained target device that is communicatively connected to a parameter server, which stores a master copy of the AI model. The AI model may be dissected into smaller portions (e.g., layers or sub-layers), and each portion may be executed as efficiently as possible on the target device. After execution of one portion of the AI model is finished, another portion of the AI model may be downloaded and executed at the target device. To improve efficiency, the input samples may be divided into microbatches, and a plurality of microbatches executing in sequential order may form a minibatch. The size of the group of microbatches or minibatch can be adjusted to reduce the communication overhead. Multi-level parallel parameters reduction may be performed at the parameter server and the target device.

Abstract: Embodiments of the present disclosure include systems and methods for providing hierarchical and shared exponent floating point data types. First and second shared exponent values are determined based on exponent values of a plurality of floating point values. A third shared exponent value is determined based the first shared exponent value and the second shared exponent value. First and second difference values are determined based on the first shared exponent value, the second shared exponent value, and the third shared exponent value. Sign values and mantissa values are determined for the plurality of floating point values. The sign value and the mantissa value for each floating point value in the plurality of floating point values, the third shared exponent value, the first difference value, and the second difference value are stored in a data structure for a shared exponent floating point data type.

Abstract: Methods, systems, apparatuses, and computer program products are described herein that enable execution of a large AI model on a memory-constrained target device that is communicatively connected to a parameter server, which stores a master copy of the AI model. The AI model may be dissected into smaller portions (e.g., layers or sub-layers), and each portion may be executed as efficiently as possible on the target device. After execution of one portion of the AI model is finished, another portion of the AI model may be downloaded and executed at the target device. This paradigm of executing one portion of the AI model at a time allows for dynamic execution of the large AI model.

Abstract: Techniques for shared memory spaces in data and model parallelism are provided to improve memory efficiency and memory access speed. A shared memory space may be established at a host system or in a hardware memory agent. The shared memory may store training data or model parameters for an artificial intelligence model at a memory address in one or more memory circuits. Data for the artificial intelligence model may be processed across a plurality of artificial intelligence accelerators using the training data or the model parameters of the shared memory space. That is, multiple accelerators access one copy of the data from the shared memory space instead of accessing their own separate memory space.

Abstract: Embodiments of the present disclosure include an error recovery method comprising detecting a computing error, restarting a first artificial intelligence processor of a plurality of artificial intelligence processors processing a data set, and loading a model in the artificial intelligence processor, wherein the model corresponds to a same model processed by the plurality of artificial intelligence processors during a previous processing iteration by the plurality of artificial intelligence processors on data from the data set.

Abstract: Embodiments of the present disclosure include an error recovery method comprising detecting a computing error, restarting a first artificial intelligence processor of a plurality of artificial intelligence processors processing a data set, and loading a model in the artificial intelligence processor, wherein the model corresponds to a same model processed by the plurality of artificial intelligence processors during a previous processing iteration by the plurality of artificial intelligence processors on data from the data set.

Abstract: Systems and methods related to dual-momentum gradient optimization with reduced memory requirements are described. An example method in a system comprising a gradient optimizer and a memory configured to store momentum values associated with a neural network model comprising L layers is described. The method includes retrieving from the memory a first set of momentum values and a second set of momentum values, corresponding to a layer of the neural network model, having a selected storage format. The method further includes converting the first set of momentum values to a third set of momentum values having a training format associated with the gradient optimizer and converting the second set of momentum values to a fourth set of momentum values having a training format associated with the gradient optimizer. The method further includes performing gradient optimization using the third set of momentum values and the fourth set of momentum values.