Made fixes to AVX3 detection on macOS #2083
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jan-wassenberg
approved these changes
Apr 12, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Updated DetectTargets() to fix a bug that can occur on macOS 12.1 or earlier on x86_64 CPU's that have AVX3 support.
There is a bug on macOS 12.1 or earlier on x86_64 CPU's with AVX3 support where ZMM16-ZMM31, the upper 256 bits of ZMM0-ZMM15, and K0-K7 are not always properly preserved across context switches, and this bug is described at https://community.intel.com/t5/Software-Tuning-Performance/MacOS-Darwin-kernel-bug-clobbers-AVX-512-opmask-register-state/m-p/1327259, golang/go#49233, and simdutf/simdutf#236.
On macOS, it is also possible for bits 5, 6, and 7 of XCR0 to be zero on x86_64 CPU's that support AVX3 before the first AVX512 instruction is executed by the current thread as macOS only preserves ZMM16-ZMM31, the upper 256 bits of ZMM0-ZMM15, and K0-K7 across context switches on threads that have executed an AVX512 instruction.
DetectTargets() has been updated to check for AVX3 support on macOS by doing
IsMacOs12_2OrLater() && HasCpuFeature("hw.optional.avx512f")instead of checking that bits 5, 6, and 7 of XCR0 are set to avoid false negative results and inconsistent behavior on x86_64 CPU's that support AVX3 on macOS and to disable AVX3 targets on macOS versions earlier than 12.2.HasCpuFeature is a wrapper around the macOS sysctlbyname API function that returns true if a particular CPU feature is supported and returns false if a CPU feature is not supported.