SIMD processing unit with local data share and access to a global data share of a GPU

摘要

A graphics processing unit is disclosed, the graphics processing unit having a processor having one or more SIMD processing units, and a local data share corresponding to one of the one or more SIMD processing units, the local data share comprising one or more low latency accessible memory regions for each group of threads assigned to one or more execution wavefronts, and a global data share comprising one or more low latency memory regions for each group of threads.

主权项

1. A graphics processing unit comprising:
a plurality of SIMD processing units, each SIMD processing unit comprising:
a plurality of thread processors,a plurality of general purpose register (GPR) files, each GPR file m communication with one of the plurality of thread processors, anda local data share (LDS) in bi-directional communication with the plurality of thread processors, wherein the plurality of thread processors are assigned a plurality of threads for execution, wherein the LDS has a private write space assigned to one of the plurality of thread processors and a public write space accessible by more than one of the plurality of thread processors, wherein only the one of the plurality of thread processors has both read and write access to the private write space, and wherein the plurality of thread processors have direct read access to the private write space; and a global data share (GDS) in communication with each of the plurality of thread processors in each of the plurality of SIMD processing units.