llama.cpp
vulkan: chunked GATED_DELTA_NET implementation
#30207
Open
Overview
Commits
16
Changes
View On GitHub

vulkan: chunked GATED_DELTA_NET implementation #30207

0cc4m wants to merge 16 commits into master from 0cc4m/vulkan-gdn-cm2
0cc4m
0cc4m vulkan: add cm2 chunked GATED_DELTA_NET shader
f200094f
0cc4m fix shmem bank conflict bottleneck
bde3ccf6
0cc4m optimize further
df078fb5
0cc4m add coopmat maintenance1 variant
a4c938c9
0cc4m optimize for RDNA4
aa5f6ca9
0cc4m add gmem version RDNA4
94eb47f0
0cc4m optimize for Ampere
094d5670
0cc4m fix nvidia shmem overuse
da8d4883
0cc4m fix cm2 validation errors
d9cedfea
0cc4m improve cm2 perf
7ff8fe48
0cc4m improve cm1 perf
75823219
0cc4m clean up
b25101f9
0cc4m enable by default
76535993
0cc4m show cm maintenance1 marker in device info
51443839
0cc4m properly check for cm_m1 features
1de478bf
0cc4m add tests
bfc04053
github-actions github-actions added testing
github-actions github-actions added Vulkan
github-actions github-actions added ggml
jeffbolznv

Login to write a write a comment.

Login via GitHub

Reviewers
No reviews
Assignees
No one assigned
Labels
testing Vulkan ggml
Milestone
No milestone
Terms Of Use Privacy Policy Refund Policy Impressum