Skip to content

[https://nvbugs/6412108][fix] Restore original order — all_reduce the routed partial first, then add the… #15922

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
trtllm-agent wants to merge 1 commit into
base: main
Choose a base branch
from
+7 −7
Open

[https://nvbugs/6412108][fix] Restore original order — all_reduce the routed partial first, then add the… #15922

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
13 changes: 7 additions & 6 deletions tensorrt_llm/_torch/auto_deploy/models/custom/modeling_qwen3_5_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -774,13 +774,14 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
layer_type="moe",
)

# Single merge-point all_reduce for routed + shared partial sums.
# Both branches produce per-rank partial outputs under TP/EP sharding
# (routed: MoEShardableNode; shared: rowwise down_proj inside the MLP).
# One reduction on the sum lifts both to full; reducing before the add
# would mix a full routed contribution with an unreduced shared one.
expert_output = expert_output + shared_expert_output
# The shared expert is replicated (Qwen3_5MoeMLP intentionally omits
# ``layer_type`` and the yaml ``shard_layers`` whitelist excludes it),
# so its output is already the full value on every rank. All-reduce
# the sharded routed-expert partial first, then add the replicated
# shared output; adding before would scale the shared output by the
# TP world size.
expert_output = torch.ops.auto_deploy.all_reduce(expert_output, layer_type="moe")
expert_output = expert_output + shared_expert_output

expert_output = expert_output.reshape(batch_size, sequence_length, hidden_dim)
return expert_output
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 @@ -39,7 +39,6 @@ accuracy/test_llm_api_autodeploy.py::TestNemotronNanoV3::test_accuracy[bf16-4-at
accuracy/test_llm_api_autodeploy.py::TestNemotronNanoV3::test_accuracy[fp8-4-attn_dp_off-trtllm] SKIP (https://nvbugs/6367792)
accuracy/test_llm_api_autodeploy.py::TestNemotronNanoV3::test_accuracy[nvfp4-1-attn_dp_off-trtllm] SKIP (temporary ToT main waive; B200 AutoDeploy NVFP4 GSM8K accuracy below threshold)
accuracy/test_llm_api_autodeploy.py::TestNemotronNanoV3::test_accuracy[nvfp4-4-attn_dp_off-trtllm] SKIP (https://nvbugs/6367792)
accuracy/test_llm_api_autodeploy.py::TestQwen3_5_397B_MoE::test_nvfp4[8] SKIP (https://nvbugs/6412108)
accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_fp8_blockscale[throughput] SKIP (https://nvbugs/6379333)
accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_fp8_blockscale[throughput_mtp] SKIP (https://nvbugs/6281818)
accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_fp8_blockscale[throughput_mtp_trtllm] SKIP (https://nvbugs/6281818)
Expand Down
Loading