llama.cpp
graph : remove redundant GDN state transposes
#20443
Merged

graph : remove redundant GDN state transposes #20443

ggerganov merged 6 commits into master from gg/gdn-fix-state-transpose
ggerganov
ggerganov ggerganov requested a review from CISC CISC 13 days ago
github-actions github-actions added model
github-actions github-actions added Nvidia GPU
github-actions github-actions added ggml
github-actions github-actions added Apple Metal
CISC
CISC approved these changes on 2026-03-12
0cc4m
ggerganov
ORippler
ORippler
arkavo-com ggml : transpose fused GDN state access for coalesced memory reads (#…
b0dbb39e
arkavo-com ggml : use SIMD dot products in CPU GDN kernel, couple AR/chunked fus…
fb32cd48
ggerganov llama : rever fgdn argument changes
d9a7ab36
ggerganov graph : remove GDN state transposes
fe3ef4a0
ggerganov ggerganov force pushed from 7ea6ee4f to fe3ef4a0 12 days ago
ggerganov vulkan : adapt
2882a4b8
ggerganov ggerganov requested a review from 0cc4m 0cc4m 12 days ago
github-actions github-actions added Vulkan
ORippler
ORippler cuda : remove obsolete smem code
d466d89b
ggerganov ggerganov merged e30f1fdf into master 12 days ago
ggerganov ggerganov deleted the gg/gdn-fix-state-transpose branch 12 days ago
jeffbolznv
0cc4m
jeffbolznv
pedapudi

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone