ggml-cpu: implement MXFP4 SIMD for s390x #16193
ggml-cpu: impl mxfp4 s390x
618ef467
ggml-cpu: missing s = sumf
6549e3bc
ggml-cpu: fix incorrect kval_mxfp4 type
377d0fca
ggml-cpu: rework mxfp4
35389304
ggml-cpu: missing delta calc
cf927d8d
ggml-cpu: fix typo
ae718c7d
ggml-cpu: fix typo for vec_splats
f7e7539c
ggml-cpu: expand to 2 blocks per loop
5fb1bb99
ggml-cpu: add unroll to boost perf
4f85c33d
ggml-cpu: back to 1 block per loop to test perf
1fe55724
Revert "ggml-cpu: back to 1 block per loop to test perf"
1f99e511
slaren
approved these changes
on 2025-09-24
ggml-cpu: rm unroll from single block
96cba331
ggerganov
approved these changes
on 2025-09-26
ggerganov
merged
9b265118
into master 355 days ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub