Skip to content

[https://nvbugs/6159132][fix] Differentiate the two paths via extra_acc_spec="tp_attn" when attention_dp=False #13922

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

Merged
2ez4bz merged 2 commits into from
Aug 10, 2026
+6 −2
Merged

[https://nvbugs/6159132][fix] Differentiate the two paths via extra_acc_spec="tp_attn" when attention_dp=False #13922

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
7e11d1d
[nvbugs/6159132][fix] MiniMaxM2 GSM8K reference for TP-sharded attent…
tensorrt-cicd May 8, 2026
be462ad
[nvbugs/6159132][chore] Remove stale waiver after fix
tensorrt-cicd May 15, 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
3 changes: 3 additions & 0 deletions tests/integration/defs/accuracy/references/gsm8k.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -491,6 +491,9 @@ MiniMaxAI/MiniMax-M2:
- accuracy: 89.045
- quant_algo: FP8_BLOCK_SCALES
accuracy: 93.75
- quant_algo: FP8_BLOCK_SCALES
extra_acc_spec: tp_attn
accuracy: 92.0
MiniMaxAI/MiniMax-M2.5:
- quant_algo: FP8_BLOCK_SCALES
accuracy: 93.75
Expand Down
4 changes: 3 additions & 1 deletion tests/integration/defs/accuracy/test_llm_api_pytorch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7740,7 +7740,9 @@ def test_4gpus(self, tp_size, ep_size, attention_dp, cuda_graph,
enable_attention_dp=attention_dp) as llm:
assert llm.args.quant_config.quant_algo == QuantAlgo.FP8_BLOCK_SCALES
task = GSM8K(self.MODEL_NAME)
task.evaluate(llm)
# TP-sharded path uses fused minimax_allreduce_rms_qk with slightly different numerics.
task.evaluate(llm,
extra_acc_spec=None if attention_dp else "tp_attn")


@skip_pre_hopper
Expand Down
1 change: 0 additions & 1 deletion tests/integration/test_lists/waives.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -78,7 +78,6 @@ accuracy/test_llm_api_pytorch.py::TestLlama3_1_8BInstruct::test_guided_decoding_
accuracy/test_llm_api_pytorch.py::TestLlama3_3_70BInstruct::test_fp4_tp2pp2[torch_compile=False-enable_gemm_allreduce_fusion=False] SKIP (https://nvbugs/6428089)
accuracy/test_llm_api_pytorch.py::TestLlama3_3_70BInstruct::test_fp4_tp2pp2[torch_compile=True-enable_gemm_allreduce_fusion=False] SKIP (https://nvbugs/6427411)
accuracy/test_llm_api_pytorch.py::TestLlama3_3_70BInstruct::test_fp4_tp2pp2[torch_compile=True-enable_gemm_allreduce_fusion=True] SKIP (https://nvbugs/6211441)
accuracy/test_llm_api_pytorch.py::TestMiniMaxM2::test_4gpus[attention_dp=False-cuda_graph=True-overlap_scheduler=True-tp_size=4-ep_size=4] SKIP (https://nvbugs/6159132)
accuracy/test_llm_api_pytorch.py::TestNemotronV3Super::test_nvfp4_8gpus_mtp SKIP (https://nvbugs/6581065)
accuracy/test_llm_api_pytorch.py::TestNemotronV3Super::test_nvfp4_parallelism[TP4_PP2] SKIP (https://nvbugs/6427411)
accuracy/test_llm_api_pytorch.py::TestNemotronV3Ultra::test_nvfp4_4gpus_static_eplb[moe_backend=CUTLASS] SKIP (https://nvbugs/6535767)
Expand Down
Loading