Skip to content

docs: clarify VNNI dispatch tiers — F32x16 is the floor, no scalar on x86 avx512vnni (64 MACs) and avxvnniint8 (32 MACs) are mutually exclusive by hardware generation. The scalar i32 path in matvec_dispatch only exists for non-x86 correctness. On x86, the thinking engine dispatches to F32x16 FMA (16 MACs) when no VNNI is detected — never reaches the scalar path. https://claude.ai/code/session_01ChLvBfpJS8dQhHxRD4pYNp#83

Merged
AdaWorldAPI merged 1 commit into
masterfrom
claude/setup-embedding-pipeline-Fa65C
Apr 4, 2026

Conversation

@AdaWorldAPI
Copy link
Copy Markdown
Owner

No description provided.

… x86

avx512vnni (64 MACs) and avxvnniint8 (32 MACs) are mutually exclusive
by hardware generation. The scalar i32 path in matvec_dispatch only
exists for non-x86 correctness. On x86, the thinking engine dispatches
to F32x16 FMA (16 MACs) when no VNNI is detected — never reaches
the scalar path.

https://claude.ai/code/session_01ChLvBfpJS8dQhHxRD4pYNp
@chatgpt-codex-connector
Copy link
Copy Markdown

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.

@AdaWorldAPI AdaWorldAPI merged commit bab9f2e into master Apr 4, 2026
4 of 10 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants