Intel Corporation

Buqi Cheng

Subramaniam M Maiyuran

Guei-Yuan Lueh

Chandra S Gurram

Ashutosh Garg

Jorge E Parra

Wei-Yu Chen

Supratim Pal

Jorge F Garcia Pabon

Darin Starkey

Shubh Shah

Kaiyu Chen

Konrad Trifunovic

Weiyu Chen

Brent A Schwartz

Ashwin J Shivani

Li Liu

Gansha Wu

Tin-Fook Ngai

Zhaohui Du

Peinan Zhang

Mechanisms for reducing register bank conflicts based on software hint and hardware thread switch are disclosed. In some embodiments, an apparatus for thread switching includes a graphics processing unit (GPU) that includes a plurality of register banks to store operands that are assigned at least partially to avoid register bank conflicts. Decoding circuitry checks a thread switching field of a first instruction to be executed by a first thread. The GPU performs a thread switch mechanism to cause a second instruction to be executed by a second thread when the thread switching field of the first instruction is set.

Systems and methods for reducing register bank conflicts based on a software hint bit causing a hardware thread switch

Embodiments are generally directed to register bank conflict reduction for multi-threaded processor execution units. An embodiment of an apparatus includes a processor including one or more execution units (EUs), at least a first execution unit (EU) to process a plurality of threads, the first EU including a register file including multiple register banks with each register bank including multiple registers, and one or more read multiplexers to read registers from the register file, wherein attempting to read more than one register from a single register bank of the register file in a same clock cycle generates a register bank conflict. Registers for each thread for the first EU are distributed across the registers banks within the register file such that a first register for a first thread of the plurality of threads and a following second register for the first thread are located in different register banks within the register file.

Register bank conflict reduction for multi-threaded processor

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

A method to provide effective control and data flow information in an Intermediate Representation (IR) form. A Path Sensitive single Assignment (PSA) IR form with effective and explicit control and data path information supports control flow sensitive optimizations such as path sensitive symbolic substitution, array privatization and speculative multi threading. In the definition of PSA form, besides defining new versioned variables, the gamma functions keep control path information. The gamma function in PSA form keeps the basic attribute of SSA IR form and only one definition exists for each use. Therefore, all existing Single Static Assignment (SSA) IR form based analysis can be applied in PSA form. The gamma function in PSA form keeps all essential control flow information and eliminates unnecessary predicates at the same time.

Method and system for intermediate representation of source code

In one embodiment, the present invention includes a method for constructing a data dependency graph (DDG) for a loop to be transformed, performing statement shifting to transform the loop into a first transformed loop according to at least one of first and second algorithms, performing unimodular and echelon transformations of a selected one of the first or second transformed loops, partitioning the selected transformed loop to obtain maximum outer level parallelism (MOLP), and partitioning the selected transformed loop into multiple sub-loops. Other embodiments are described and claimed.

Method and apparatus to achieve maximum outer level parallelism of a loop

Growing community of inventors

San Jose, CA, United States of America

Buqi Cheng