Intel Corporation

Darin Starkey

Subramaniam M Maiyuran

Jorge E Parra

Chandra S Gurram

Ashutosh Garg

Supratim Pal

Guei-Yuan Lueh

Varghese George

Shubra Marwaha

Kaiyu Chen

Durgesh Borkar

Brent A Schwartz

Wei-Yu Chen

Jorge F Garcia Pabon

Shubh Shah

Konrad Trifunovic

Buqi Cheng

Weiyu Chen

Gang Y Chen

Ashwin J Shivani

Thomas A Piazza

Pratik J Ashar

Described herein is a graphics processor including a plurality of processing clusters coupled with a host interface, each processing cluster comprising a plurality of multiprocessors, the plurality of multiprocessors interconnected via a data interconnect, and each multiprocessor comprising sparse matrix multiply acceleration hardware including a systolic processing array with feedback inputs.

Scalable sparse matrix multiply acceleration using systolic arrays with feedback inputs

Graphics processors and graphics processing units having dot product accumulate instructions for a hybrid floating point format are disclosed. In one embodiment, a graphics multiprocessor comprises an instruction unit to dispatch instructions and a processing resource coupled to the instruction unit. The processing resource is configured to receive a dot product accumulate instruction from the instruction unit and to process the dot product accumulate instruction using a bfloat16 number (BF16) format.

Graphics processors and graphics processing units having dot product accumulate instruction for hybrid floating point format

Described herein is a graphics processing unit (GPU) configured to receive an instruction having multiple operands, where the instruction is a single instruction multiple data (SIMD) instruction configured to use a bfloat16 (BF16) number format and the BF16 number format is a sixteen-bit floating point format having an eight-bit exponent. The GPU can process the instruction using the multiple operands, where to process the instruction includes to perform a multiply operation, perform an addition to a result of the multiply operation, and apply a rectified linear unit function to a result of the addition.

Examples described herein relate to a software and hardware optimization that manages scenarios where a write operation to a register is less than an entirety of the register. A compiler detects instructions that make partial writes to the same register, groups such instructions, and provides hints to hardware of the partial write. The execution unit combines the output data for grouped instructions and updates the destination register as single write instead of multiple separate partial writes.

Compiler assisted register file write reduction

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.

Described herein is an accelerator device including a host interface, a fabric interconnect coupled with the host interface, and one or more hardware tiles coupled with the fabric interconnect, the one or more hardware tiles including sparse matrix multiply acceleration hardware including a systolic array with feedback inputs.

Described herein is a graphics processing unit (GPU) comprising a single instruction, multiple thread (SIMT) multiprocessor comprising an instruction cache, a shared memory coupled with the instruction cache, and circuitry coupled with the shared memory and the instruction cache, the circuitry including multiple texture units, a first core including hardware to accelerate matrix operations, and a second core configured to receive an instruction having multiple operands in a bfloat16 (BF16) number format, wherein the multiple operands include a first source operand, a second source operand, and a third source operand, and the BF16 number format is a sixteen-bit floating point format having an eight-bit exponent and process the instruction, wherein to process the instruction includes to multiply the second source operand by the third source operand and add a first source operand to a result of the multiply.

Embodiments provide support for divergent control flow in heterogeneous compute operations on a fused execution unit. On embodiment provides for a processing apparatus comprising a fused execution unit including multiple graphics execution units having a common instruction pointer; logic to serialize divergent function calls by the fused execution unit, the logic configured to compare a call target of execution channels within the fused execution unit and create multiple groups of channels, each group of channels associated with a single call target; and wherein the fused execution unit is to execute a first group of channels via a first execution unit and a second group of channels via a second execution unit.

Divergent control flow for fused EUs

Embodiments described herein provide a graphics processor in which dependency tracking hardware is simplified via the use of compiler provided software scoreboard information. In one embodiment the shader compiler for shader programs is configured to encode software scoreboard information into each instruction. Dependencies can be evaluated by the shader compiler and provided as scoreboard information with each instruction. The hardware can then use the provided information when scheduling instructions. In one embodiment, a software scoreboard synchronization instruction is provided to facilitate software dependency handling within a shader program. Using software to facilitate software dependency handling and synchronization can simplify hardware design, reducing the area consumed by the hardware. In one embodiment, dependencies can be evaluated by the shader compiler instead of the GPU hardware. The compiler can then insert a software scoreboard sync immediate instruction into compiled program code to manage instruction dependencies and prevent data hazards from occurring.

Software scoreboard information and synchronization

An apparatus and method for a SIMD structured branching. For example, one embodiment of a processor comprises: an execution unit having a plurality of channels to execute instructions; and a branch unit to process control flow instructions and to maintain a per channel count for each channel and a control instruction count for the control flow instructions, the branch unit to enable and disable the channels based at least on the per channel count.

Method and apparatus for SIMD structured branching

An apparatus and method for a SIMD unstructured branching. For example, one embodiment of a processor comprises: an execution unit having a plurality of channels to execute instructions; and a branch unit to process unstructured control flow instructions and to maintain a per channel count value for each channel, the branch unit to store instruction pointer tags for the unstructured control flow instructions in a memory and identify the instruction pointer tags using tag addresses, the branch unit to further enable and disable the channels based at least on the per channel count value.

Growing community of inventors

Roseville, CA, United States of America

Darin Starkey