Skip to content

Optimize block unrolling operations using AVX-512 #83798

@BruceForstall

Description

@BruceForstall

CodeGen::genCodeForInitBlkUnroll and CodeGen::genCodeForCpBlkUnroll have specific cases for optimizing using YMM registers. Enhance these to use ZMM registers if available. Consider the performance implications of the change; measure the improvements.

Related: update Compiler::getUnrollThreshold as appropriate.

Metadata

Metadata

Assignees

Labels

area-CodeGen-coreclrCLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMIavx512Related to the AVX-512 architecture

Type

No type

Projects

No projects

Milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions