]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 11f0af5)
ggml: Correct SVE implementation in ggml_vec_dot_f16_unroll (#16518)
authorsirus20x6 <redacted>
Sun, 12 Oct 2025 05:15:00 +0000 (00:15 -0500)
committerGitHub <redacted>
Sun, 12 Oct 2025 05:15:00 +0000 (08:15 +0300)
commit20cc625edc2264aae2779e71bef1593e6a4e8c43
tree3306864a50de1a009218a49b4978ede2f7a2e376tree
parent11f0af5504252e453d57406a935480c909e3ff37commit | diff
ggml: Correct SVE implementation in ggml_vec_dot_f16_unroll (#16518)

The previous SVE implementation for `ggml_vec_dot_f16_unroll` contained a bug due to a copy-paste error. The wrong variable was used in an FMA instruction, leading to incorrect results. This commit corrects the variable usage and improves the clarity of the code by renaming variables to avoid confusion.

Co-authored-by: Aaron <redacted>
ggml/src/ggml-cpu/vec.h diff | blob | history
Packaging of ggml-org/llama.cpp