Skip to content

[https://nvbugs/6095851][fix] Refresh FlashMLA metadata for draft KV cache #15251

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
yizhang-nv merged 1 commit into from
Jul 3, 2026
+8 −2
Merged

[https://nvbugs/6095851][fix] Refresh FlashMLA metadata for draft KV cache #15251

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
8 changes: 8 additions & 0 deletions tensorrt_llm/_torch/speculative/interface.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -104,6 +104,8 @@ def prepare_attn_metadata_for_draft_replay(attn_metadata,
attn_metadata.kv_cache_block_offsets = attn_metadata.draft_kv_cache_block_offsets
attn_metadata.host_kv_cache_block_offsets = (
draft_kv_cache_manager.host_kv_cache_block_offsets)
if attn_metadata.enable_flash_mla:
attn_metadata.prepare_flash_mla()

from ..attention_backend.sparse.dsa import (DSAtrtllmAttentionMetadata,
Indexer)
Expand Down Expand Up @@ -157,6 +159,8 @@ def restore_attn_metadata_after_draft_replay(attn_metadata, saved_state):
saved_state['target_kv_cache_block_offsets'])
attn_metadata.host_kv_cache_block_offsets = (
saved_state['target_host_kv_cache_block_offsets'])
if attn_metadata.enable_flash_mla:
attn_metadata.prepare_flash_mla()
saved_dsa = saved_state.get('saved_dsa_state')
if saved_dsa is not None:
m = attn_metadata
Expand Down Expand Up @@ -1599,6 +1603,8 @@ def draft_kv_cache_context(self, attn_metadata, draft_kv_cache_manager):
attn_metadata.kv_cache_manager = draft_kv_cache_manager
attn_metadata.kv_cache_block_offsets = attn_metadata.draft_kv_cache_block_offsets
attn_metadata.host_kv_cache_block_offsets = draft_kv_cache_manager.host_kv_cache_block_offsets
if attn_metadata.enable_flash_mla:
attn_metadata.prepare_flash_mla()

try:
yield
Expand All @@ -1607,6 +1613,8 @@ def draft_kv_cache_context(self, attn_metadata, draft_kv_cache_manager):
attn_metadata.kv_cache_manager = target_kv_cache_manager
attn_metadata.kv_cache_block_offsets = target_kv_cache_block_offsets
attn_metadata.host_kv_cache_block_offsets = target_host_kv_cache_block_offsets
if attn_metadata.enable_flash_mla:
attn_metadata.prepare_flash_mla()

def _sample_tokens_for_batch(
self,
Expand Down
2 changes: 0 additions & 2 deletions tests/integration/test_lists/waives.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,8 +25,6 @@ accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=0-
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=0-attention_dp=False-cuda_graph=True-overlap_scheduler=True-torch_compile=True-enable_chunked_prefill=False-v2_kv_cache=True] SKIP (https://nvbugs/6305365)
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=0-attention_dp=True-cuda_graph=False-overlap_scheduler=False-torch_compile=True-enable_chunked_prefill=False-v2_kv_cache=True] SKIP (https://nvbugs/6305404)
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=2-attention_dp=False-cuda_graph=False-overlap_scheduler=False-torch_compile=True-enable_chunked_prefill=False-v2_kv_cache=True] SKIP (https://nvbugs/6388129)
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=2-attention_dp=False-cuda_graph=False-overlap_scheduler=True-torch_compile=False-enable_chunked_prefill=False-v2_kv_cache=True] SKIP (https://nvbugs/6084720)
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16[mtp_nextn=2-attention_dp=False-cuda_graph=True-overlap_scheduler=False-torch_compile=False-enable_chunked_prefill=False-v2_kv_cache=True] SKIP (https://nvbugs/6095851)
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16_4gpus[ep4-mtp_nextn=2-attention_dp=True-cuda_graph=True-overlap_scheduler=True-torch_compile=True] SKIP (https://nvbugs/6402058)
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16_4gpus[pp4-mtp_nextn=0-attention_dp=False-cuda_graph=False-overlap_scheduler=False-torch_compile=False] SKIP (https://nvbugs/6278337)
accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16_4gpus[pp4-mtp_nextn=0-attention_dp=False-cuda_graph=True-overlap_scheduler=False-torch_compile=False] SKIP (https://nvbugs/6278337)
Expand Down
Loading