]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
overview / pkg / ggml / sources / ggml / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 5d6c37b)
vulkan: optimizations for deepseek prompt processing (llama/14555)
authorJeff Bolz <redacted>
Sat, 12 Jul 2025 09:51:58 +0000 (04:51 -0500)
committerGeorgi Gerganov <redacted>
Sat, 12 Jul 2025 13:05:00 +0000 (16:05 +0300)
commit5eb294f4ff6e53272306d10b461487188bb6487e
tree636448271781589fe36740af16632ccc0a0b72d1tree
parent5d6c37b4b23117bb54e1969963d97be4859d3f85commit | diff
vulkan: optimizations for deepseek prompt processing (llama/14555)

* vulkan: allow unclamped loads in coopmat2 mul_mat_id shader

* vulkan: increase coopmat2 mul_mat_id tile size

* vulkan: optimize mat_mul_id row_ids search to batch loads, and port to coopmat1 path

* vulkan: use smaller FA row size when head size is large. applies to both scalar and CM2 paths (CM1 isn't used due to shared memory limits)
src/ggml-vulkan/ggml-vulkan.cpp diff | blob | history
src/ggml-vulkan/vulkan-shaders/mul_mm.comp diff | blob | history
src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp diff | blob | history
Packaging of ggml-org/ggml