Skip to content

[None][fix] Initialize KV cache fixtures and unwaive nine cases on all platforms #19382

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 2 commits into from
Sep 18, 2026
+2 −9
Merged

[None][fix] Initialize KV cache fixtures and unwaive nine cases on all platforms #19382

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
e5a8fc2
[None][fix] Initialize KV cache test fixture state
yizhang-nv Sep 18, 2026
743b06b
[None][test] Unwaive nine KV cache fixture cases on all platforms
yizhang-nv Sep 18, 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
9 changes: 0 additions & 9 deletions tests/integration/test_lists/waives.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -256,15 +256,6 @@ unittest/_torch/attention/sparse/dsa/test_req_idx_per_token.py::test_on_update_k
unittest/_torch/attention/test_attention_backends.py::test_attention_backend[exaone_moe_gqa_swa128-ctx-bf16-HND-p32-v1] SKIP (https://nvbugs/6668773)
unittest/_torch/attention/test_attention_backends.py::test_attention_backend[qwen2_0_5b_gqa_hd64-ctx-bf16-HND-p32-v1] SKIP (https://nvbugs/6641268)
unittest/_torch/executor/kv_cache/test_kv_cache_budget_split.py::TestSplitGpuBudgetForDraft::test_target_cost_uses_derived_layer_type_windows SKIP (https://nvbugs/6800223)
unittest/_torch/executor/kv_cache/test_kv_cache_manager_v2.py::test_prepare_context_drops_cached_token_attribution_while_estimating SKIP (https://github.com/NVIDIA/TensorRT-LLM/pull/19382)
unittest/_torch/executor/kv_cache/test_kv_cache_manager_v2.py::test_prepare_context_keeps_cached_token_attribution_staged_by_the_core SKIP (https://github.com/NVIDIA/TensorRT-LLM/pull/19382)
unittest/_torch/executor/test_kv_connector_v2_prefix.py::TestAskTiming::test_the_scheduling_pass_does_not_ask SKIP (https://github.com/NVIDIA/TensorRT-LLM/pull/19382)
unittest/_torch/executor/test_kv_connector_v2_prefix.py::TestReEntryAfterAServe::test_a_dead_allocation_drops_the_floor SKIP (https://github.com/NVIDIA/TensorRT-LLM/pull/19382)
unittest/_torch/executor/test_kv_connector_v2_prefix.py::TestReEntryAfterAServe::test_the_chunk_still_spans_to_the_end_of_the_prompt SKIP (https://github.com/NVIDIA/TensorRT-LLM/pull/19382)
unittest/_torch/executor/test_kv_connector_v2_prefix.py::TestReEntryAfterAServe::test_the_same_holds_for_a_synchronous_serve SKIP (https://github.com/NVIDIA/TensorRT-LLM/pull/19382)
unittest/_torch/executor/test_kv_connector_v2_prefix.py::TestReEntryAfterAServe::test_the_served_position_survives_re_entry SKIP (https://github.com/NVIDIA/TensorRT-LLM/pull/19382)
unittest/_torch/executor/test_kv_connector_v2_prefix.py::TestSwaScratchReuse::test_scratch_reuse_is_disabled_before_the_scheduler_can_take_slots SKIP (https://github.com/NVIDIA/TensorRT-LLM/pull/19382)
unittest/_torch/executor/test_kv_connector_v2_prefix.py::TestSwaScratchReuse::test_scratch_reuse_survives_without_a_connector SKIP (https://github.com/NVIDIA/TensorRT-LLM/pull/19382)
unittest/_torch/modeling/test_gemma4_e2e_dummy.py::test_e2e_text_31b_dummy SKIP (https://nvbugs/6607482)
unittest/_torch/modeling/test_modeling_nemotron_h.py::test_nemotron_h_breakable_prefill_cuda_graph -k "tp1" SKIP (https://nvbugs/6797747)
unittest/_torch/modeling/test_modeling_nemotron_h.py::test_nemotron_h_breakable_prefill_cuda_graph -k "tp2" SKIP (https://nvbugs/6797215)
Expand Down
1 change: 1 addition & 0 deletions tests/unittest/_torch/executor/kv_cache/test_kv_cache_manager_v2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1831,6 +1831,7 @@ def _make_admission_manager(
"""
manager = object.__new__(KVCacheManagerV2)
manager.conversation_manager = None
manager.kv_connector_manager = None
Comment thread
yizhang-nv marked this conversation as resolved.
manager.enable_block_reuse = True
manager.tokens_per_block = TOKENS_PER_BLOCK
manager.is_draft = False
Expand Down
1 change: 1 addition & 0 deletions tests/unittest/_torch/executor/test_kv_connector_v2_prefix.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -201,6 +201,7 @@ def make_manager(connector, num_extra_kv_tokens=0, is_draft=False):
manager = object.__new__(KVCacheManagerV2)
manager.kv_connector_manager = connector
manager.is_draft = is_draft
manager.is_estimating_kv_cache = False
manager.tokens_per_block = TOKENS_PER_BLOCK
manager.num_extra_kv_tokens = num_extra_kv_tokens
manager.kv_cache_map = {}
Expand Down
Loading