 Jeff Bolz
|
ecc93d0558
vulkan: compile a test shader in cmake to check for coopmat2 support (#10713)
|
1 year ago |
 0cc4m
|
3df784b305
Vulkan: VK_KHR_cooperative_matrix support to speed up prompt processing (#10597)
|
1 year ago |
|
Jeff Bolz
|
c9c6e01dae
vulkan: Add VK_NV_cooperative_matrix2 support for mul_mat and flash attention (#10206)
|
1 year ago |
|
Jeff Bolz
|
2759916d86
vulkan: Implement "fast divide" (mul+shift) for unary ops like copy (#10642)
|
1 year ago |
|
Jeff Bolz
|
cc98896db8
vulkan: optimize and reenable split_k (#10637)
|
1 year ago |
 Eve
|
0533e7fb38
vulkan: Dynamic subgroup size support for Q6_K mat_vec (#10536)
|
1 year ago |
|
Jeff Bolz
|
f095a649ec
vulkan: get the first command buffer submitted sooner (#10499)
|
1 year ago |
|
Jeff Bolz
|
5b3466bedf
vulkan: Handle GPUs with less shared memory (#10468)
|
1 year ago |
|
Jeff Bolz
|
904109ed0d
vulkan: fix group_norm (#10496)
|
1 year ago |
 Diego Devesa
|
5931c1f233
ggml : add support for dynamic loading of backends (#10469)
|
1 year ago |
|
Jeff Bolz
|
1bacb9f625
vulkan: further optimize mul_mat_vec using larger loads (#10387)
|
1 year ago |
|
Jeff Bolz
|
b3e585988f
vulkan: Optimize soft_max (#10301)
|
1 year ago |
|
0cc4m
|
9b75f03cd2
Vulkan: Fix device info output format specifiers (#10366)
|
1 year ago |
|
Jeff Bolz
|
772703c8ff
vulkan: Optimize some mat-vec mul quant shaders (#10296)
|
1 year ago |
 thewh1teagle
|
3225008973
ggml : vulkan logs (whisper/2547)
|
1 year ago |
|
Diego Devesa
|
ae8de6d50a
ggml : build backends as libraries (#10256)
|
1 year ago |
