llama.cpp
metal: per-op source split + parallel compile (replay)
#26561
Open

metal: per-op source split + parallel compile (replay) #26561

forforever73 wants to merge 8 commits into master from stack/metal-split
forforever73
forforever73 forforever73 requested a review 8 days ago
github-actions github-actions added ggml
github-actions github-actions added Apple Metal
forforever73 forforever73 force pushed from 864c08ed to 2934b5b3 7 days ago
forforever73 metal : per-op source split + parallel compile (#24021)
ff8169ce
ggerganov metal: add col2im_1d op (f32/f16/bf16) (#25176)
d2112aa3
ggerganov metal : add set_rows with src0 f16 (#25434)
f348d513
ggerganov metal : add CONV_2D_DW (depthwise convolution) support (#21565)
7649856e
ggerganov metal : add Q2_0 support (#25419)
0b6cbb57
ggerganov metal: fuse snake activation (mul, sin, sqr, mul, add) (#25459)
b2d79a7e
ggerganov ggml-metal: FWHT kernel for metal backend (#25924)
db0a8ce8
forforever73 metal : port new kernels into the split sources
55224983
forforever73 forforever73 force pushed from 2934b5b3 to 55224983 7 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
No reviews
Assignees
No one assigned
Labels
Milestone