Skip to content

[https://nvbugs/6564714][fix] Fix the FP8BlockScaleMoERunner get_fallback_tactic #17328

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
leslie-fang25 merged 2 commits into from
Aug 6, 2026
+4 −24
Merged

[https://nvbugs/6564714][fix] Fix the FP8BlockScaleMoERunner get_fallback_tactic #17328

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
34b9784
[https://nvbugs/6564714][fix] Fix the FP8BlockScaleMoERunner get_fall…
leslie-fang25 Aug 5, 2026
622eda1
address comments
leslie-fang25 Aug 6, 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
8 changes: 4 additions & 4 deletions tensorrt_llm/_torch/custom_ops/trtllm_gen_custom_ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -903,8 +903,8 @@ def __init__(
# that influence tactic validity here. e.g. we are tuning FC1 and FC2 so the routing
# type does not matter
def unique_id(self):
return (self.top_k, self.intermediate_size, self.local_num_experts,
self.act_type)
return (self.top_k, self.num_fused_shared_experts,
self.intermediate_size, self.local_num_experts, self.act_type)

def get_runner(self):
instance_key = ()
Expand All @@ -922,8 +922,8 @@ def get_fallback_tactic(self, hidden_size: int,
the profiling cache. The result depends only on the key below, so
memoize it the same way ``runner_dict`` memoizes the runner.
"""
key = (self.top_k, hidden_size, self.intermediate_size,
self.local_num_experts, num_tokens)
key = (self.top_k, self.num_fused_shared_experts, hidden_size,
Comment thread
leslie-fang25 marked this conversation as resolved.
self.intermediate_size, self.local_num_experts, num_tokens)
Comment thread
coderabbitai[bot] marked this conversation as resolved.
tactic = FP8BlockScaleMoERunner.fallback_tactic_dict.get(key)
if tactic is None:
tactic = tuple(
Expand Down
20 changes: 0 additions & 20 deletions tests/integration/test_lists/waives.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -361,7 +361,6 @@ perf/test_perf_sanity.py::test_e2e[aggr_upload-deepseek_r1_fp4_v2_grace_blackwel
perf/test_perf_sanity.py::test_e2e[aggr_upload-deepseek_r1_fp8_blackwell-r1_fp8_tp8_mtp3_8k1k] SKIP (https://nvbugs/6432948)
perf/test_perf_sanity.py::test_e2e[aggr_upload-dynamo_gpt_oss_120b_fp4_blackwell-gpt_oss_fp4_tep4_adp_cutlass_8k1k] SKIP (https://nvbugs/6374910)
perf/test_perf_sanity.py::test_e2e[aggr_upload-glm5_fp4_blackwell-glm5_fp4_tep8_mtp3_8k1k] SKIP (https://nvbugs/6329155)
perf/test_perf_sanity.py::test_e2e[aggr_upload-host_perf_deepseek_v3_lite-v3lite_fp8_bs8_128_256] SKIP (https://nvbugs/6564714)
perf/test_perf_sanity.py::test_e2e[aggr_upload-k25_thinking_fp4_2_nodes_grace_blackwell-k25_thinking_fp4_tep8_32k8k] SKIP (https://nvbugs/6422339)
perf/test_perf_sanity.py::test_e2e[aggr_upload-qwen3_5_397b_fp4_blackwell-qwen3_5_397b_fp4_dep8_8k1k] SKIP (https://nvbugs/6535767)
perf/test_perf_sanity.py::test_e2e[aggr_upload-qwen3_5_397b_fp4_blackwell-qwen3_5_397b_fp4_dep8_mtp3_8k1k] SKIP (https://nvbugs/6535767)
Expand Down Expand Up @@ -441,27 +440,8 @@ unittest/auto_deploy/multigpu/smoke/test_ad_allreduce_strategies.py::test_allred
unittest/auto_deploy/multigpu/smoke/test_ad_allreduce_strategies.py::test_allreduce_strategies[ONESHOT] SKIP (https://nvbugs/6517839)
unittest/auto_deploy/multigpu/smoke/test_ad_allreduce_strategies.py::test_allreduce_strategies[SYMM_MEM] SKIP (https://nvbugs/6517839)
unittest/auto_deploy/multigpu/smoke/test_ad_allreduce_strategies.py::test_allreduce_strategies[TWOSHOT] SKIP (https://nvbugs/6517839)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[256-2-4-256-1] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[256-2-4-256-4] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[256-2-4-512-1] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[256-2-4-512-4] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[256-2-8-256-1] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[256-2-8-256-4] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[256-2-8-512-1] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[256-2-8-512-4] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[512-2-4-256-1] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[512-2-4-256-4] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[512-2-4-512-1] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[512-2-4-512-4] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[512-2-8-256-1] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[512-2-8-256-4] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[512-2-8-512-1] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/custom_ops/moe/test_trtllm_moe.py::test_trtllm_finegrained_fp8_moe_blackwell[512-2-8-512-4] SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/models/test_glm4_moe_modeling.py::test_glm4_moe_decoder_layer_equivalence[1-dtype0-1-8] SKIP (https://nvbugs/6566730)
unittest/auto_deploy/singlegpu/models/test_glm4_moe_modeling.py::test_glm4_moe_moe_equivalence[dtype0-1-8] SKIP (https://nvbugs/6566730)
unittest/auto_deploy/singlegpu/transformations/library/test_moe_fusion.py SKIP (https://nvbugs/6564714)
unittest/auto_deploy/singlegpu/transformations/library/test_moe_fusion.py::test_trtllm_quant_finegrained_fp8_moe_fused_correctness SKIP (https://nvbugs/6564714)
unittest/bindings/test_transfer_agent_bindings.py::TestMooncakeFunctionalTransfer::test_mooncake_wait_in_progress_on_zero_timeout SKIP (https://nvbugs/6517836)
unittest/bindings/test_transfer_agent_bindings.py::TestMooncakeFunctionalTransfer::test_mooncake_write_transfer_gpu_tensor SKIP (https://nvbugs/6517836)
unittest/bindings/test_transfer_agent_bindings.py::TestMooncakeFunctionalTransfer::test_mooncake_write_transfer_multiple_chunks SKIP (https://nvbugs/6517836)
Expand Down
Loading