Skip to content

[ExecuTorch][WebGPU] test: q4gsw shmem-GEMM prefill golden configs #20606

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
JulianCloudNTH wants to merge 2 commits into
base: gh/JulianCloudNTH/80/base
Choose a base branch
from
+7 −0
Open

[ExecuTorch][WebGPU] test: q4gsw shmem-GEMM prefill golden configs #20606

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
adf6c95
Update
JulianCloudNTH Jun 29, 2026
1164e70
Update
JulianCloudNTH Jun 30, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions backends/webgpu/test/ops/test_quantized_linear.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,6 +53,10 @@ class Q4gswConfig:
Q4gswConfig("kv_proj_4k", 4096, 2048, 512),
Q4gswConfig("q_proj_8k", 8192, 2048, 2048, heavy=True), # 67MB golden
Q4gswConfig("kv_proj_8k", 8192, 2048, 512, heavy=True),
# Prefill shapes routing to the shmem GEMM (K>=4096 or N>=4096); M=128.
Q4gswConfig("gate_proj_pf", 128, 2048, 8192), # gate/up prefill (shmem via N)
Q4gswConfig("down_proj_pf", 128, 8192, 2048), # down prefill (shmem via K)
Q4gswConfig("shmem_edge", 130, 4096, 2056), # partial 32-tile bounds
]


Expand Down
3 changes: 3 additions & 0 deletions backends/webgpu/test/test_webgpu_native.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -288,6 +288,9 @@ static const Q4gswConfig kQ4gswConfigs[] = {
{"kv_proj_4k", 4096, 2048, 512, 1e-4f, 1e-3f, true, false},
{"q_proj_8k", 8192, 2048, 2048, 1e-4f, 1e-3f, false, true},
{"kv_proj_8k", 8192, 2048, 512, 1e-4f, 1e-3f, false, true},
{"gate_proj_pf", 128, 2048, 8192, 1e-4f, 1e-3f, true, false}, // shmem via N
{"down_proj_pf", 128, 8192, 2048, 1e-3f, 1e-2f, true, false}, // shmem via K
{"shmem_edge", 130, 4096, 2056, 1e-4f, 1e-3f, true, false}, // partial tiles
};

// /16 ramp over the flat index; mirrors test_quantized_linear.py _ramp_input.
Expand Down
Loading