hybrid_w4a16: Tune qwen3-vl-4b TTFT shapes for gfx1150 with 2 448x448 image and 256 text token input #1012
Merged
Merged
Conversation
mgehre-amd
reviewed
Jun 22, 2026
qingxuamd
force-pushed
the
qingxu/qwen3-vl-optimize2
branch
from
1828d6d to
504cc57
Compare
mgehre-amd
reviewed
Jun 24, 2026
qingxuamd
force-pushed
the
qingxu/qwen3-vl-optimize2
branch
from
504cc57 to
5f2e7ac
Compare
mgehre-amd
approved these changes
Jun 26, 2026
mgehre-amd
changed the title
This optimization is for gfx1150. The model of Qwen3-VL-4B-Instruct-AWQ-4bit-lm_head_int8, for triton_w4a16_skinny_fmt_kernel, it cost more then 60% latency in prefill. The required input = 2 448x448 img + 256 token, then, tok num ~=660 tok. M: ~660 (2x448x448 + 256 token prefill） N,K（hot shape）： 660 x 19456 x 2560 660 x 2560 x 9728 660 x 6144 x 2560 660 x 2560 x 4096 warps, currently 8 is best num so: BLOCK_M=64 BLOCK_N=256 BLOCK_K=64 num_warps=8 GEMM：M=660，N/K as above kernel tile：64,256,64,8 Signed-off-by: Xu Qing <qing.xu2@amd.com>
qingxuamd
force-pushed
the
qingxu/qwen3-vl-optimize2
branch
from
5f2e7ac to
169f9c0
Compare
Add two missing blank lines between module-level functions as required by ruff-format (E302). Signed-off-by: Matthias Gehre <matthias.gehre@amd.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Specifically optimize the kernel tile according to qwen3-vl-4b and input shape, without this PR, TTFT=1231 ms, with this PR, TTFT = 980 ms.