Skip to content

Actions: pytorch/executorch

Actions

Test Vulkan Backend (specialized runners)

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
vulkan.yml
600 workflow runs
600 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
Use dynamic CPU count for cmake --build -j in docs and test scripts (… Test Vulkan Backend (specialized runners) #581: Commit 7f86746 pushed by JakeStevens
44s main
main
44s
Cortex-M backend: lower quantized GELU via the activation LUT (#21025) Test Vulkan Backend (specialized runners) #580: Commit 7628aa6 pushed by rascani
41m 46s main
main
41m 46s
Arm backend: Document RIFE accuracy handover (#21059) Test Vulkan Backend (specialized runners) #579: Commit 53180df pushed by bdemirb
41s main
main
41s
Arm backend: Support SNORM RGBA runtime format (#20964) Test Vulkan Backend (specialized runners) #578: Commit 4f3f8ed pushed by robell
39s main
main
39s
[ET-VK][sdpa] Vendor-adaptive head_dim output-tiling (TILE_N4=2) in G… Test Vulkan Backend (specialized runners) #577: Commit 266e0dc pushed by SS-JIA
37m 52s main
main
37m 52s
[ET-VK][sdpa] Add SDPA operator perf benchmark binary (test_sdpa) Test Vulkan Backend (specialized runners) #576: Pull request #21116 synchronize by SS-JIA
37m 27s gh/SS-JIA/575/orig
gh/SS-JIA/575/orig
37m 27s
[ET-VK][sdpa] Reuse shared V cache across GQA query heads in AV coop-GEMV Test Vulkan Backend (specialized runners) #575: Pull request #21117 synchronize by SS-JIA
37m 56s gh/SS-JIA/576/orig
gh/SS-JIA/576/orig
37m 56s
[ET-VK][sdpa] Vendor-adaptive head_dim output-tiling (TILE_N4=2) in GQA AV coop-GEMV Test Vulkan Backend (specialized runners) #574: Pull request #21118 opened by pytorchbot
43m 56s gh/SS-JIA/577/orig
gh/SS-JIA/577/orig
43m 56s
[ET-VK][sdpa] Reuse shared V cache across GQA query heads in AV coop-GEMV Test Vulkan Backend (specialized runners) #573: Pull request #21117 opened by pytorchbot
43m 47s gh/SS-JIA/576/orig
gh/SS-JIA/576/orig
43m 47s
[ET-VK][sdpa] Add SDPA operator perf benchmark binary (test_sdpa) Test Vulkan Backend (specialized runners) #572: Pull request #21116 opened by pytorchbot
43m 24s gh/SS-JIA/575/orig
gh/SS-JIA/575/orig
43m 24s
Fuse consecutive clamps into a single clamp (#21013) Test Vulkan Backend (specialized runners) #571: Commit b22a902 pushed by meta-codesync Bot
45s main
main
45s
Fix viable strict updating job (#21111) Test Vulkan Backend (specialized runners) #570: Commit 181ded4 pushed by metascroy
36s main
main
36s
Flakey CI fix (#21110) Test Vulkan Backend (specialized runners) #569: Commit f6eaba2 pushed by metascroy
38m 20s main
main
38m 20s
[ET-VK][sdpa] Vendor-adaptive head_dim output-tiling (TILE_N4=2) in GQA AV coop-GEMV Test Vulkan Backend (specialized runners) #568: Pull request #21064 synchronize by SS-JIA
37m 45s gh/SS-JIA/577/head
gh/SS-JIA/577/head
37m 45s
[ET-VK][sdpa] Reuse shared V cache across GQA query heads in AV coop-GEMV Test Vulkan Backend (specialized runners) #567: Pull request #21063 synchronize by SS-JIA
37m 53s gh/SS-JIA/576/head
gh/SS-JIA/576/head
37m 53s
[ET-VK][sdpa] Add SDPA operator perf benchmark binary (test_sdpa) Test Vulkan Backend (specialized runners) #566: Pull request #21062 synchronize by SS-JIA
37m 50s gh/SS-JIA/575/head
gh/SS-JIA/575/head
37m 50s
Qualcomm AI Engine Direct - improve llama3.2 3B TPS (#20903) Test Vulkan Backend (specialized runners) #565: Commit fcc1020 pushed by psiddh
40s main
main
40s
[ExecuTorch][WebGPU] Fuse attention QKV projections in prefill (#21107) Test Vulkan Backend (specialized runners) #564: Commit 5828778 pushed by JCNTH
46s main
main
46s
Preserve BatchNorm running state in composable quantizer Test Vulkan Backend (specialized runners) #563: Commit 6bd0355 pushed by denizkilinc
1m 10s main
main
1m 10s
Revert "Arm backend: Rerun duplicate-user fusion after TOSA lowering"… Test Vulkan Backend (specialized runners) #562: Commit cd380e7 pushed by metascroy
44m 31s main
main
44m 31s
Make extension/cuda:caller_stream fbcode-only Test Vulkan Backend (specialized runners) #561: Commit 37400d9 pushed by meta-codesync Bot
45s main
main
45s
gate fbcode-only TARGETS dirs with is_fbcode() and rename to BUCK Test Vulkan Backend (specialized runners) #560: Commit b3ac2df pushed by meta-codesync Bot
38m 3s main
main
38m 3s
Add cooperative matrix dispatch for quantized linear Test Vulkan Backend (specialized runners) #559: Commit 4a95dff pushed by meta-codesync Bot
42m 54s main
main
42m 54s
[ExecuTorch][WebGPU] Add fp32 matmul ops (mm, bmm, linear) to the Web… Test Vulkan Backend (specialized runners) #558: Commit 666ba0b pushed by JCNTH
25m 20s main
main
25m 20s
Proper implementation for VulkanBackend::is_available() Test Vulkan Backend (specialized runners) #557: Commit 8b14fc2 pushed by meta-codesync Bot
37m 46s main
main
37m 46s