Skip to content

refactor(triton): reorganize conv modules and unify gated FP8 quant path #3048

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
hellozhuo-amd wants to merge 8 commits into
base: main
Choose a base branch
from
Open

refactor(triton): reorganize conv modules and unify gated FP8 quant path #3048

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
e7fbac1
refactor(triton): reorganize conv modules and unify gated FP8 quant path
hellozhuo-amd May 6, 2026
f7616c4
Updated op_tests/triton_tests/quant/test_fused_rms_gated_fp8_group_qu…
hellozhuo-amd May 6, 2026
658a217
fix: add group boundary masking in gated FP8 quant kernel loop
Copilot May 6, 2026
faf4ee2
fix: pass constant placeholders for gated-only constexprs in non-gate…
Copilot May 6, 2026
076ab91
Merge branch 'main' into zhuo/triton-pr2423-reorg
hellozhuo-amd May 6, 2026
8b756ac
fix(triton): backward-compat conv imports and align conv tests
hellozhuo-amd May 6, 2026
e1687f3
Merge branch 'main' into zhuo/triton-pr2423-reorg
hellozhuo-amd May 6, 2026
2e74ecd
Merge branch 'main' into zhuo/triton-pr2423-reorg
hellozhuo-amd May 7, 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
2 changes: 1 addition & 1 deletion .github/scripts/split_tests.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,7 +134,7 @@ elif [[ "$TEST_TYPE" == "triton" ]]; then
FILE_TIMES[op_tests/triton_tests/attention/test_mha.py]=1452
FILE_TIMES[op_tests/triton_tests/test_pa_decode_gluon.py]=718
FILE_TIMES[op_tests/triton_tests/attention/test_pa_decode.py]=635
FILE_TIMES[op_tests/triton_tests/test_causal_conv1d.py]=634
FILE_TIMES[op_tests/triton_tests/conv/test_causal_conv1d.py]=634
FILE_TIMES[op_tests/triton_tests/gemm/batched/test_batched_gemm_a8w8_a_per_token_group_prequant_w_per_batched_tensor_quant.py]=402
FILE_TIMES[op_tests/triton_tests/attention/test_flash_attn_kvcache.py]=357
FILE_TIMES[op_tests/triton_tests/attention/test_chunked_pa_prefill.py]=336
Expand Down
6 changes: 5 additions & 1 deletion .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,4 +79,8 @@ scripts/results/*.csv

# Jupyter notebook checkpoints and outputs
.ipynb_checkpoints/
**/.ipynb_checkpoints/
**/.ipynb_checkpoints/

# vim swaps
*.swo
*.swp
3 changes: 3 additions & 0 deletions aiter/ops/triton/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -141,6 +141,9 @@
# Quant modules (quant/)
"fused_fp8_quant": "quant.fused_fp8_quant",
"fused_mxfp4_quant": "quant.fused_mxfp4_quant",
# Conv modules (conv/)
"causal_conv1d": "conv.causal_conv1d",
"causal_conv1d_update_single_token": "conv.causal_conv1d_update_single_token",
}


Expand Down
6 changes: 6 additions & 0 deletions aiter/ops/triton/_triton_kernels/conv/__init__.py
View file
Open in desktop
Copy link
Copy Markdown
Contributor

@Boss2002n Boss2002n May 8, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can we keep this file empty

Original file line number Diff line number Diff line change
@@ -0,0 +1,6 @@
# SPDX-License-Identifier: MIT
# Copyright (C) 2024-2026, Advanced Micro Devices, Inc. All rights reserved.

from .causal_conv1d import PAD_SLOT_ID

__all__ = ["PAD_SLOT_ID"]
0 ...s/triton/_triton_kernels/causal_conv1d.py → ...ton/_triton_kernels/conv/causal_conv1d.py
View file
Open in desktop
File renamed without changes.
0 ...nels/causal_conv1d_update_single_token.py → ...conv/causal_conv1d_update_single_token.py
View file
Open in desktop
File renamed without changes.
Loading
Loading