reduce for kqmax_new_j is unnecessary#1032

Merged

JohannesGaessler merged 2 commits intoggml-org:masterfrom mahorozte:feature/fa-opt-vec-f16

Dec 3, 2024

-2

Commits on Nov 29, 2024

kqmax_new_j in every thread within warp is same after operate at line 199,this reduce can be omit
ZhaoXiaoYu
committed

Commits on Dec 2, 2024

same problem in vec32
ZhaoXiaoYu
committed