In one embodiment, a method for tensor data distribution using a direct-memory access agent includes generating, by a first controller, source addresses indicating locations in a source memory where portions of a source tensor are stored. A second controller may generate destination addresses indicating locations in a destination memory where portions of a destination tensor are to be stored. The direct-memory access agent receives a source address generated by the first controller and a destination address generated by the second controller and determines a burst size. The direct-memory access agent may issue a read request comprising the source address and the burst size to read tensor data from the source memory and may store the tensor data into an alignment buffer. The direct-memory access agent then issues a write request comprising the destination address and the burst size to write data from the alignment buffer into the destination memory.

Various examples are directed to systems and methods in which a first flow controller of a first synchronous flow may receive an instruction to execute a first loop using the first synchronous flow. The first flow controller may determine a first iteration index for a first iteration of the first loop. The first flow controller may send, to a first compute element of the first synchronous flow, a first synchronous message to initiate a first synchronous flow thread for executing the first iteration of the first loop. The first synchronous message may comprise the iteration index. The first compute element may execute an input/output operation at a first location of a first compute element memory indicated by the first iteration index.

A streaming engine employed in a digital signal processor specifies a fixed read only data stream. Once fetched the data stream is stored in two head registers for presentation to functional units in the fixed order. Data use by the functional unit is preferably controlled using the input operand fields of the corresponding instruction. A first read only operand coding supplies data from the first head register. A first read/advance operand coding supplies data from the first head register and also advances the stream to the next sequential data elements. Corresponding second read only operand coding and second read/advance operand coding operate similarly with the second head register. A third read only operand coding supplies double width data from both head registers.

Described herein is a graphics processing unit (GPU) comprising a first processing cluster to perform parallel processing operations, the parallel processing operations including a ray tracing operation and a matrix multiply operation; and a second processing cluster coupled to the first processing cluster, wherein the first processing cluster includes a floating-point unit to perform floating point operations, the floating-point unit is configured to process an instruction using a bfloat16 (BF16) format with a multiplier to multiply second and third source operands while an accumulator adds a first source operand with output from the multiplier.

A computing unit is disclosed, comprising a first memory bank for storing input activations and a second memory bank for storing parameters used in performing computations. The computing unit includes at least one cell comprising at least one multiply accumulate (“MAC”) operator that receives parameters from the second memory bank and performs computations. The computing unit further includes a first traversal unit that provides a control signal to the first memory bank to cause an input activation to be provided to a data bus accessible by the MAC operator. The computing unit performs one or more computations associated with at least one element of a data array, the one or more computations being performed by the MAC operator and comprising, in part, a multiply operation of the input activation received from the data bus and a parameter received from the second memory bank.

A processor includes a plurality of execution units. At least one of the execution units is configured to repeatedly execute a first instruction based on a first field of the first instruction indicating that the first instruction is to be iteratively executed.

An SIMD instruction generation and processing method and a related device are provided. The method may include: obtaining a length of each loop dimension of a first tensor formula; selecting, from a plurality of groups of information about a first SIMD instruction model based on the length of each loop dimension of a first tensor formula, information about a second SIMD instruction model matching the first tensor formula; generating, based on a length of at least one loop dimension of the first tensor formula and the second SIMD instruction model, a first SIMD instruction obtained after the first tensor formula is converted. The information about a second SIMD instruction model is selected from the plurality of groups of information about a first SIMD instruction model based on the length of each loop dimension of the tensor formula.

A method is provided that includes performing, by a processor in response to a vector sort instruction, sorting of values stored in lanes of the vector to generate a sorted vector, wherein the values in a first portion of the lanes are sorted in a first order indicated by the vector sort instruction and the values in a second portion of the lanes are sorted in a second order indicated by the vector sort instruction; and storing the sorted vector in a storage location.

Stateful microbranch instructions, including: generating, based on an instruction, a first one or more microinstructions including a stateful microbranch instruction, wherein the stateful microbranch instruction includes: an address of a next instruction after the instruction; a branch target address; one or more microcode attributes; and executing the first one or more microinstructions.

Various examples are directed to systems and methods in which a flow controller of a first synchronous flow may receive an instruction to execute a first loop using the first synchronous flow. The flow controller may determine a first iteration index for a first iteration of the first loop. The flow controller may send, to a first compute element of the first synchronous flow, a first synchronous message to initiate a first synchronous flow thread for executing the first iteration of the first loop. The first synchronous message may comprise the iteration index. The first compute element may execute an input/output operation at a first location of a first compute element memory indicated by the first iteration index.