Disclosed embodiments relate to a method and device for optimizing compilation of source code. The proposed method receives a first intermediate representation code of a source code and analyses each basic block instruction of the plurality of basic block instructions contained in the first intermediate representation code for blockification. In order to blockify the identical instructions, the one or more groups of basic block instructions are assessed for eligibility of blockification. Upon determining as eligible, the group of basic block instructions are blockified using one of one dimensional SIMD vectorization and two-dimensional SIMD vectorization. The method further generates a second intermediate representation of the source code which is translated to executable target code with more efficient processing capacity.

Embodiments detailed herein relate to systems and methods to zero a tile register pair. In one example, a processor includes decode circuitry to decode a matrix pair zeroing instruction having fields for an opcode and an identifier to identify a destination matrix having a PAIR parameter equal to TRUE; and execution circuitry to execute the decoded matrix pair zeroing instruction to zero every element of a left matrix and a right matrix of the identified destination matrix.

Systems, apparatuses, and methods for implementing a tiling algorithm for a matrix math instruction set are disclosed. A system includes at least a memory, a cache, a processor, and a plurality of compute units. The memory stores a plurality of matrix elements in a linear format, and the processor converts the plurality of matrix elements from the linear format to a tiling format. Each compute unit retrieves a plurality of matrix elements from the memory into the cache. Each compute unit includes a matrix operations unit which loads the plurality of matrix elements of corresponding tile(s) from the cache and performs a matrix operation on the plurality of matrix elements to generate a result in the tiling format. The system generates a classification of a first dataset based on results of the matrix operations.

One embodiment provides for a graphics processing unit (GPU) to accelerate machine learning operations, the GPU comprising an instruction cache to store a first instruction and a second instruction, the first instruction to cause the GPU to perform a floating-point operation, including a multi-dimensional floating-point operation, and the second instruction to cause the GPU to perform an integer operation; and a general-purpose graphics compute unit having a single instruction, multiple thread architecture, the general-purpose graphics compute unit to concurrently execute the first instruction and the second instruction.

An apparatus stores a program including a description of loop processing of iterating a plurality of instructions, and rearranges an execution sequence of the plurality of instructions in the program such that the loop processing is pipelined by software pipeline. The apparatus inserts an instruction to use a register for single instruction multiple data (SIMD) extension instruction, into the description of the loop processing in the program.

Disclosed embodiments relate to systems and methods for performing instructions to transform matrices into a row-interleaved format. In one example, a processor includes fetch and decode circuitry to fetch and decode an instruction having fields to specify an opcode and locations of source and destination matrices, wherein the opcode indicates that the processor is to transform the specified source matrix into the specified destination matrix having the row-interleaved format; and execution circuitry to respond to the decoded instruction by transforming the specified source matrix into the specified RowInt-formatted destination matrix by interleaving J elements of each J-element sub-column of the specified source matrix in either row-major or column-major order into a K-wide submatrix of the specified destination matrix, the K-wide submatrix having K columns and enough rows to hold the J elements.

A processing system executes wavefronts at multiple arithmetic logic unit (ALU) pipelines of a single instruction multiple data (SIMD) unit in a single execution cycle. The ALU pipelines each include a number of ALUs that execute instructions on wavefront operands that are collected from vector general process register (VGPR) banks at a cache and output results of the instructions executed on the wavefronts at a buffer. By storing wavefronts supplied by the VGPR banks at the cache, a greater number of wavefronts can be made available to the SIMD unit without increasing the VGPR bandwidth, enabling multiple ALU pipelines to execute instructions during a single execution cycle.

Apparatuses and methods are disclosed for an FPGA architecture that may improve processing speed and efficiency in processing less complex operands. Some applications may utilize operands that are less complex, such as operands that are 1, 2, or 4 bits, for example. In some examples, the DSP architecture may skip or avoid processing all received operands or may process a common operand more frequently than other operands. An example apparatus may include a first configurable logic unit configured to receive a first operand and a second operand; a second configurable logic unit configured to receive a third operand and the first calculated operand; a first switch configured to receive the first operand and a fourth operand and to output a first selected operand; and a second switch configured to receive the second calculated operand and the first selected operand.

Disclosed embodiments relate to systems and methods for performing instructions to transform matrices into a row-interleaved format. In one example, a processor includes fetch and decode circuitry to fetch and decode an instruction having fields to specify an opcode and locations of source and destination matrices, wherein the opcode indicates that the processor is to transform the specified source matrix into the specified destination matrix having the row-interleaved format; and execution circuitry to respond to the decoded instruction by transforming the specified source matrix into the specified RowInt-formatted destination matrix by interleaving J elements of each J-element sub-column of the specified source matrix in either row-major or column-major order into a K-wide submatrix of the specified destination matrix, the K-wide submatrix having K columns and enough rows to hold the J elements.

The vector data path is divided into smaller vector lanes. A register such as a memory mapped control register stores a vector lane number (VLX) indicating the number of vector lanes to be powered. A decoder converts this VLX into a vector lane control word, each bit controlling the ON of OFF state of the corresponding vector lane. This number of contiguous least significant vector lanes are powered. In the preferred embodiment the stored data VLX indicates that 2.sup.VLX contiguous least significant vector lanes are to be powered. Thus the number of vector lanes powered is limited to an integral power of 2. This manner of coding produces a very compact controlling bit field while obtaining substantially all the power saving advantage of individually controlling the power of all vector lanes.