19.1

Adjusted Signal

19.1

Raw Pattern Score

100%

Time Factor

2026-10-31

Last Push

6.8K

Stars

Cuda

Language

4.3M

Lines of Code

6.3K

Files

49.9K

Pattern Hits

2026-10-31

Scan Date

0.83

HC Hit Rate

Risk Review · Payload v1

Use each vector to guide manual verification of the underlying findings.

Payload schema: 1.0.0 · Engine: 2.0.0 · Scoring: heuristic-density-v1 · Scan date: 2026-10-31.

Cognitive Ownership Risk 24.4/100 · LOW

Can maintainers explain intent, assumptions, and trade-offs? 26688 finding(s).

Hallucinated Dependency & Supply Chain Risk 9.8/100 · LOW

Are imports declared and package/API names verified? 8719 finding(s).

Agent Over-Engineering Ratio 53.8/100 · HIGH

Are wrappers, guards, and repeated structures necessary? 34158 finding(s).

PR Review Velocity Impact 65.6/100 · HIGH

Which changes need focused human review? 49875 finding(s).

What These Metrics Mean

Adjusted Signal: Legacy pattern score normalized per 1,000 scanned lines and multiplied by the repository-level time factor. It is not an estimate of AI authorship.
Raw Pattern Score: Sum of severity- and context-weighted pattern findings before applying the time factor.
Time Factor: A coarse multiplier based on repository last-push metadata. It does not identify when individual code was written.
Pattern Hits: Count of configured pattern findings. Review each location and surrounding code before acting.
HC Hit Rate: High- and critical-severity findings per scanned file. Severity is a configured detector weight, not proof of a vulnerability or authorship.
Lines / Files: Scan coverage denominators used by the legacy normalized score.

Pattern-Signal History

This chart tracks the legacy adjusted pattern score across scan runs. Changes can reflect source changes, scan coverage, or detector/scoring versions; they do not measure AI adoption or prove that a team refactored code. Compare runs only when their scoring versions are compatible. The dashed line tracks the raw pattern-hit count.

Severity Breakdown

Groups findings by configured severity weight. Severity is not a confidence estimate, security rating, or proof that a package is missing. Verify imports against the project's manifests and inspect each finding in its local design context.

CRITICAL 84HIGH 5173MEDIUM 7548LOW 37070

Directory Score Breakdown

This horizontal bar chart shows where configured pattern scores accumulate across directories. Treat the bars as locations to inspect, not as evidence of authorship or code quality. Use the file-level findings to decide whether a directory needs a focused ownership, dependency, complexity, or review check.

Pattern Findings

The scanner identified 49875 distinct pattern matches across 25 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a detector match to inspect in context. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.

Reading the findings table: The Severity column indicates the configured finding weight (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — legacy comment-context matches receive a ×1.5 score weight. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier under the legacy scoring model.

Cross-File Repetition4891 hits · 24455 pts

Severity	File	Line	Snippet	Context
HIGH	build_backend.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/test_minimax_h3_fp8_varlen_attention.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_dsa_indexer.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_kimi_k3_fp8_projection.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_w8a8_megamoe_ep16.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_kda_det_train.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_mla_varq_dcp_decode.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	…ts/experimental/test_cake_dense_projection_det_gemm.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_balanced_gqa_decode.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_sm110_gqa_decode.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_kimi_k3_latent_moe.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_dsa_train.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_mla_nvfp4_paged_decode.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_sm110_gqa_decode_prepared.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_lm_head_loss.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_vsa_sm90.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	…rimental/test_cake_moe_grouped_det_gemm_determinism.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_experimental_api.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	…erimental/test_cake_dense_projection_det_gemm_names.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_msa_nvfp4_decode.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_vsa_sm90_api.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_qsa_nvfp4.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_kimi_k3_fused_router.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_kimi_k3_vision_tower.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_mhc_pre_big_fuse_with_norm.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	…/experimental/test_cake_minimax_h3_varlen_attention.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_mxfp8_megamoe_ep16.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/experimental/test_cake_moe_grouped_det_gemm.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_nvfp4_attention_sm120.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_prefill_uniform_q_len_plan.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_mla_auto_backend.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_lse_base.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_bidirectional_ranges_attention.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_workspace_size.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_auto_backend_upgrade.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_cake_vsa.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_eagle_verify_fp8kv_sm90.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/attention/test_cake_vsa_planning.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/moe_ep/test_sm90_push_fp8_gemm.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/cpu/trace/test_b12x_moe_trace.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/cpu/utils/test_collect_env.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/cpu/jit/test_bgmv_moe_jit.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/cpu/jit/test_aot_sm103_modules.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/cpu/jit/test_cutlass_fused_moe_aot.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/cpu/jit/test_cake_bgmv_moe_jit.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/cpu/jit/test_blackwell_msa_jit.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/mamba/test_cake_ssd_combined.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/mamba/test_cake_ssd_combined_capture.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/comm/test_mixed_comm_checkpoint.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/comm/test_pcie_ipc_all_reduce.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/comm/test_ulysses_head_chunk.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/utils/test_gen_module_symlink_race_condition.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/utils/test_sampling_seed_offset.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/utils/test_cake_sampling_upstream.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/utils/test_gated_act_mxfp8.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/gemm/test_cake_mm_fp4.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/gemm/test_group_gemm_fp4.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/gemm/test_cudnn_knob_tactics.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/quantization/test_cake_nvfp4_quantize.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
HIGH	tests/gdn/test_replayssm_spec_fold.py	0	copyright (c) 2025 by flashinfer team. licensed under the apache license, version 2.0 (the "license"); you may not use t	STRING
4831 more matches not shown…

Decorative Section Separators7214 hits · 21807 pts

Severity	File	Line	Snippet	Context
MEDIUM	csrc/fused_moe/monomoe/src/ptx_utils.h	14	// ── cp.async scalar helpers (sm_80+) ──────────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/ptx_utils.h	35	// ── Hopper WGMMA (sm_90a) helpers ─────────────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/ptx_utils.h	148	// ── Hopper mbarrier helpers (sm_90a) ──────────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/ptx_utils.h	309	// ── Hopper TMA bulk-tensor copy (sm_90a) ──────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_tma.h	214	// ─── Device-side TMA load helpers ────────────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_tma.h	437	// ─── Down-projection (Phase 4) TMA load helpers ──────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_interface.h	40	// ── Single fixed-shape Dims (E=256, N=512, K=2048 block-wise FP8) ─────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	24	// ── KernelConfig opt-in detection ───────────────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	257	// ── Sentinel-based Phase 3 → 4 handoff state (tail fields) ────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	285	// ── Phase 4 → 5 readiness flags (replaces the site-#3 barrier) ────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	285	// ── Phase 4 → 5 readiness flags (replaces the site-#3 barrier) ────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	306	// ── Sentinel handoff helpers (Phase 3 → 4) ────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	409	// ── WGMMA tile geometry ───────────────────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	427	// ── Down-proj outer K-step (K_STEP_DOWN tunable) ──────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	437	// ── Down-proj TMA pipeline depth (DOWN_PIPE_DEPTH tunable) ────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	451	// ── Up-proj outer K-step (K_STEP_UP tunable) ──────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	462	// ── Up-proj weight-TMA lookahead (UP_W_SLOTS tunable) ─────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	483	// ── Up-proj col-halves (M-axis 128-row atoms per block) ───────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	501	// ── Down-projection grid layout ───────────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	665	// ── mbarriers (16-B alignment required by SM90 mbarrier PTX) ─────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	678	// ── Phase 3 → Phase 4 (expert, token) reorganization tables ──────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	730	// ── Common fields ────────────────────────────────────────────────────────	COMMENT
MEDIUM	csrc/fused_moe/monomoe/src/moe_internal.h	781	// ── Warp identity helpers ───────────────────────────────────────────────────	COMMENT
MEDIUM	tests/test_minimax_h3_mxfp8.py	324	# --------------------------------------------------------------------------	COMMENT
MEDIUM	tests/test_minimax_h3_mxfp8.py	326	# --------------------------------------------------------------------------	COMMENT
MEDIUM	tests/test_minimax_h3_nvfp4.py	509	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	tests/test_minimax_h3_nvfp4.py	511	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	tests/test_minimax_h3_nvfp4.py	1015	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	tests/test_minimax_h3_nvfp4.py	1017	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/test_minimax_h3_qkv_quantize_pack.py	65	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/test_minimax_h3_qkv_quantize_pack.py	67	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/test_minimax_h3_qkv_quantize_pack.py	191	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/test_minimax_h3_qkv_quantize_pack.py	193	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	tests/test_cake_causal_conv1d_det_train_kcp.py	588	# ----------------------------------------------------------------------------------------------------------------	COMMENT
MEDIUM	tests/kda/test_kda_prefill_trained_gate_distribution.py	395	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	tests/kda/test_kda_prefill_trained_gate_distribution.py	408	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_kda_output_only.py	493	# =============================================================================	COMMENT
MEDIUM⚡	tests/kda/test_kda_output_only.py	495	# =============================================================================	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	752	# ==============================================================================	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	754	# ==============================================================================	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	1468	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	1470	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	1589	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	1591	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	1655	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	1657	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	1687	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	1689	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2109	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2111	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2155	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2157	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2227	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2229	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2287	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2289	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2457	# ==============================================================================	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2459	# ==============================================================================	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2468	# ------------------------------------------------------------------------------	COMMENT
MEDIUM⚡	tests/kda/test_recurrent_kda.py	2470	# ------------------------------------------------------------------------------	COMMENT
7154 more matches not shown…

Hyper-Verbose Identifiers18556 hits · 16300 pts

Severity	File	Line	Snippet	Context
LOW⚡	build_backend.py	1153	def get_requires_for_build_wheel(config_settings=None):	CODE
LOW⚡	build_backend.py	1158	def get_requires_for_build_sdist(config_settings=None):	CODE
LOW⚡	build_backend.py	1163	def get_requires_for_build_editable(config_settings=None):	CODE
LOW⚡	build_backend.py	1168	def prepare_metadata_for_build_wheel(metadata_directory, config_settings=None):	CODE
LOW⚡	build_backend.py	1173	def prepare_metadata_for_build_editable(metadata_directory, config_settings=None):	CODE
LOW	build_backend.py	359	def _synthesize_nccl_builddir(build: Path) -> None:	CODE
LOW	build_backend.py	417	def _check_nccl_version_drift(wheel: Path) -> None:	CODE
LOW	build_backend.py	698	def _install_nvep_runtime_wheels(built_nixl: bool) -> None:	CODE
LOW	build_backend.py	792	def _system_cuda_tile_compiler_available() -> bool:	CODE
LOW	build_backend.py	801	def _install_cuda_tile_compile_deps() -> None:	CODE
LOW	build_utils.py	43	def get_dependency_requirements(	CODE
LOW⚡	build_utils.py	69	def get_build_dependency_requirements(	CODE
LOW⚡	build_utils.py	76	def get_cuda_extra_dependency_requirements(	CODE
LOW⚡	build_utils.py	83	def get_ci_image_dependency_requirements(	CODE
LOW⚡	build_utils.py	90	def get_cuda_tile_compile_dependency_requirements() -> list[str]:	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_ffe79cdc3573a82e5cc0_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_dd0b4fbada3615175a95_kernel.py	112	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_5c2a4f52a093fd54e5fc_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_6cb3a150103691dde0de_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_7f291b75fc80b9a071f0_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_6cfe5b1f49445aeb8696_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_c6d09364a66d1c1190ad_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_31c6bccc1180f0dd2350_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_3842336f9e9e393d3dda_kernel.py	109	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_7dccce76ace7a05df2b3_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_6fd2cab425181a6f291d_kernel.py	110	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_1d943b27569adbab332c_kernel.py	112	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_72ece5210802b11d0db8_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_f859d3c2a0ab081471d7_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_ddf905709278cfc39a9b_kernel.py	109	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_b7a69392ef67f34fcc0d_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_529b7eb9bd7bef286ab5_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_420577304e5b3471f743_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_335a69c359d728d6af44_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_3627d2e2661cafcee2bd_kernel.py	112	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_c954193ac8e9caf92181_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_066e94e570b6a8a9b087_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_b3ce0d3341158795cdc2_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_af57b7cab33f085e2f65_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_4e5ee0eae63df027f2b0_kernel.py	112	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_bd3d1ae26b97c5fddbb5_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_9efcd6af56b94dc0151d_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_e0ed95f2da24a7c364d3_kernel.py	112	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_9679b7dd9bf720adfbcc_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_172cacdf988f22944dc2_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_cc2104779e227b135ea3_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_4a452481524e0de7d29b_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_cdad4980b1d2baab7a70_kernel.py	109	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_660e0aba2121514f9ee6_kernel.py	116	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_6d1700289b7122de7347_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_185c95e41de2f577bde8_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_a251cfee19c1d1e8e8f1_kernel.py	115	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_394ae8bc3951d3e66cbe_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_b0a44ae685b334a65639_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_85464c39902415732dc9_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_222a7f18ed572ce91300_kernel.py	115	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_4c4ded85112c2ae18068_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_f147c22769f47c2f9ec4_kernel.py	109	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
LOW	…03a/cake_mxfp4_situ_moe_da3b1d07a34de335619a_kernel.py	102	def _cake_launch_cluster_spread(kernel_launcher, **kwargs):	CODE
18496 more matches not shown…

Unused Imports8527 hits · 7602 pts

Severity	File	Line	Context
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	18	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	23	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	26	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	27	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	30	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	31	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	32	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	33	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	42	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	43	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	18	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	23	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	26	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	27	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	30	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	31	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	32	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	33	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	42	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	43	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	18	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	23	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	26	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	27	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	30	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	31	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	32	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	33	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	42	CODE
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	43	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	18	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	23	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	26	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	27	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	30	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	31	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	32	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	33	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	42	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	43	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	18	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	26	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	30	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	31	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	32	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	33	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	34	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	34	CODE
8467 more matches not shown…

Deep Nesting3178 hits · 2626 pts

Severity	File	Line	Context
LOW	build_backend.py	556	CODE
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_426dd8da968b2c0eebc9_kernel.py	103	CODE
LOW	…03a/cake_mxfp4_situ_moe_680c12cbf8b8834e8265_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_ffe79cdc3573a82e5cc0_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_dd0b4fbada3615175a95_kernel.py	127	CODE
LOW	…03a/cake_mxfp4_situ_moe_5c2a4f52a093fd54e5fc_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_6cb3a150103691dde0de_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_7f291b75fc80b9a071f0_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_6cfe5b1f49445aeb8696_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_cd07eab0b1a2997ea73c_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_c6d09364a66d1c1190ad_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_31c6bccc1180f0dd2350_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_ea96c4c8eb07264a9fd4_kernel.py	109	CODE
LOW	…03a/cake_mxfp4_situ_moe_3842336f9e9e393d3dda_kernel.py	124	CODE
LOW	…03a/cake_mxfp4_situ_moe_7dccce76ace7a05df2b3_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_1d4a75945e58e438939d_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_27fbec2708fb3b41a41a_kernel.py	103	CODE
LOW	…03a/cake_mxfp4_situ_moe_fbf2bf37da9a9c3613da_kernel.py	109	CODE
LOW	…03a/cake_mxfp4_situ_moe_6fd2cab425181a6f291d_kernel.py	125	CODE
LOW	…03a/cake_mxfp4_situ_moe_1d943b27569adbab332c_kernel.py	127	CODE
LOW	…03a/cake_mxfp4_situ_moe_72ece5210802b11d0db8_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_258d09e3bdc0b94a6873_kernel.py	109	CODE
LOW	…03a/cake_mxfp4_situ_moe_f859d3c2a0ab081471d7_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_ddf905709278cfc39a9b_kernel.py	124	CODE
LOW	…03a/cake_mxfp4_situ_moe_b7a69392ef67f34fcc0d_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_f8ffa5365621d970babf_kernel.py	109	CODE
LOW	…03a/cake_mxfp4_situ_moe_03552e950e5afe3aa67a_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_d4eee9dcefe1606b0a09_kernel.py	103	CODE
LOW	…03a/cake_mxfp4_situ_moe_529b7eb9bd7bef286ab5_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_54f0c94b1ac5812dc6ed_kernel.py	103	CODE
LOW	…03a/cake_mxfp4_situ_moe_420577304e5b3471f743_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_335a69c359d728d6af44_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_3627d2e2661cafcee2bd_kernel.py	127	CODE
LOW	…03a/cake_mxfp4_situ_moe_c954193ac8e9caf92181_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_066e94e570b6a8a9b087_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_b3ce0d3341158795cdc2_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_af57b7cab33f085e2f65_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_4e5ee0eae63df027f2b0_kernel.py	127	CODE
LOW	…03a/cake_mxfp4_situ_moe_6a021d55b7aaa069c148_kernel.py	111	CODE
LOW	…03a/cake_mxfp4_situ_moe_9d64a6ec84389ae76702_kernel.py	112	CODE
LOW	…03a/cake_mxfp4_situ_moe_bd3d1ae26b97c5fddbb5_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_04d4dac8da67c2e6463b_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_9efcd6af56b94dc0151d_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_e0ed95f2da24a7c364d3_kernel.py	127	CODE
LOW	…03a/cake_mxfp4_situ_moe_6d44b0fe7249401b957e_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_9679b7dd9bf720adfbcc_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_172cacdf988f22944dc2_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_cc2104779e227b135ea3_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_4a452481524e0de7d29b_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_cdad4980b1d2baab7a70_kernel.py	124	CODE
LOW	…03a/cake_mxfp4_situ_moe_660e0aba2121514f9ee6_kernel.py	131	CODE
LOW	…03a/cake_mxfp4_situ_moe_6d1700289b7122de7347_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_185c95e41de2f577bde8_kernel.py	117	CODE
LOW	…03a/cake_mxfp4_situ_moe_28f8827fbe1200a49d77_kernel.py	111	CODE
LOW	…03a/cake_mxfp4_situ_moe_de3bc9b31b36fa668a4a_kernel.py	110	CODE
LOW	…03a/cake_mxfp4_situ_moe_a251cfee19c1d1e8e8f1_kernel.py	130	CODE
3118 more matches not shown…

Over-Commented Block3050 hits · 2595 pts

Severity	File	Line	Snippet	Context
LOW	pytest.ini	1	[pytest]	COMMENT
LOW	requirements.txt	1	apache-tvm-ffi>=0.1.11,<0.2	COMMENT
LOW	requirements.txt	21	# pin nvidia-nccl-cuXX==2.30.7 exactly, which trips the torch-eviction problem	COMMENT
LOW	.pre-commit-config.yaml	1	# To use:	COMMENT
LOW	.pre-commit-config.yaml	21	# upstream kernel drops that are re-synced periodically	COMMENT
LOW	pyproject.toml	1	# Copyright (c) 2024 by FlashInfer team.	COMMENT
LOW	pyproject.toml	281	# The vendored cuDNN Frontend tree of the Qwen 3.5 gated attention block (see the ruff exclude note).	COMMENT
LOW	build_backend.py	501	shutil.copy(sopath, dst / soname)	COMMENT
LOW	csrc/topk_transform_checks.h	21	// (radix_topk_page_table_transform and cub_topk_page_table_transform), so the backends	COMMENT
LOW	csrc/fmha_v2/fused_multihead_attention.h	201	int64_t softmax_stats_stride_in_bytes;	COMMENT
LOW	…/fmha_v2/fused_multihead_attention_kernel_4x1_hopper.h	461	// Create buffer_full barriers with 1 arrive count	COMMENT
LOW	csrc/fmha_v2/fused_multihead_attention_utils.h	801	int const* cu_mask_rows) {	COMMENT
LOW	…/fused_multihead_flash_attention_kernel_noloop_tiled.h	241	float global_sum[Softmax::ROWS_PER_THREAD];	COMMENT
LOW	csrc/fmha_v2/fused_multihead_attention_kernel.h	21	#include <fmha/smem_tile_o.h>	COMMENT
LOW	csrc/fmha_v2/fused_multihead_attention_kernel.h	41	// - P = norm * K^T * Q , where norm is the normalization term (a scalar),	COMMENT
LOW	…c/fmha_v2/fused_multihead_attention_demo_bert_params.h	101	// The stride between matrices of packed mask.	COMMENT
LOW	…c/fmha_v2/fused_multihead_attention_demo_bert_params.h	181	#ifdef SKIP_SOFTMAX_STAT	COMMENT
LOW	csrc/fmha_v2/fmha/softmax.h	1041	}	COMMENT
LOW	csrc/fmha_v2/fmha/mask.h	341	}	COMMENT
LOW	csrc/fmha_v2/fmha/utils.h	21	#include <__clang_cuda_builtin_vars.h>	COMMENT
LOW	csrc/fmha_v2/fmha/utils.h	61	using Unpack_remaining = typename Unpack<Ns...>::Type;	COMMENT
LOW	csrc/fmha_v2/fmha/utils.h	81	// (84): here	COMMENT
LOW	csrc/fmha_v2/fmha/utils.h	101	struct Cond_inspect_ns {	COMMENT
LOW	csrc/fmha_v2/fmha/utils.h	801	// asm volatile("fma.rn.f16x2.relu %0, %1, %2, %3;" : "=r"(c) : "r"(a), "r"(one),	COMMENT
LOW	csrc/fmha_v2/fmha/utils.h	821	// return dst;	COMMENT
LOW	csrc/fmha_v2/fmha/paged_kv_cache.h	21	struct Kv_block_array {	COMMENT
LOW	csrc/fmha_v2/fmha/gmem_tile_o_packed.h	1281	// The scale.	COMMENT
LOW	csrc/fmha_v2/fmha/kernel_traits.h	21	#include <fmha/gmem_tile_qkv.h>	COMMENT
LOW	csrc/fmha_v2/fmha/fragment.h	141	f.reg(4 * ii + 0) = tmp.x;	COMMENT
LOW	csrc/fmha_v2/fmha/smem_tile_v.h	481	t[3] = y;	COMMENT
LOW	csrc/fmha_v2/fmha/smem_tile_v.h	601	// Jump by 8*16 rows per K but account for packing.	COMMENT
LOW	csrc/fmha_v2/fmha/smem_tile_v.h	921	uint2 tmp_0;	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/dma.h	341	typename Shared::Head_info info{	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/dma.h	721	// now-flipped barrier -> deadlock at bar.sync 1, 128 in transpose_v_tile. Same hazard as	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/dma.h	741	// Logical:	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/compute.h	581	// results once. For example, STEP_KV=256, qgmma.K=32, then 256/32=8 MMAs are needs,	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/kernel_traits.h	41	int Q_BUFFERS_,	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/kernel_traits.h	201	// Apply the exp2f optimization (fuse bmm1_scale and -max into FMAs).	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/kernel_traits.h	421	// The step size in key/value sequence dimension (N of BMM1 and K of BMM2).	COMMENT
LOW	csrc/fmha_v2/fmha/hopper/arrive_wait.h	21		COMMENT
LOW	csrc/fmha_v2/fmha/hopper/arrive_wait.h	41	#endif	COMMENT
LOW	csrc/fmha_v2/fmha/hopper/smem_tile.h	1801	tmp[nii][ki].w); // PRMT 2+3	COMMENT
LOW	csrc/fmha_v2/fmha/hopper/smem_tile.h	1821	// [g0, g1] [o0, o1] => y: [g0, h0, o0, p0], i.e. col 0	COMMENT
LOW	csrc/fmha_v2/fmha/hopper/smem_tile.h	2001	// [00, 01]	COMMENT
LOW	csrc/fmha_v2/fmha/hopper/smem_tile.h	2121	BYTES_PER_STS, 1, 0, ROWS_PER_XOR_PATTERN, 1>;	COMMENT
LOW	csrc/fmha_v2/fmha/hopper/smem_tile.h	2321	// w: +24 (o)	COMMENT
LOW	csrc/fmha_v2/fmha/hopper/gmem_tile_o_packed.h	601	int64_t row_offset =	COMMENT
LOW	csrc/fmha_v2/fmha/hopper/fragment.h	281	Gmma_single_desc_b::TRANS_MODE == fmha::Gmma_descriptor_transpose::TRANS ? true : false,	COMMENT
LOW	csrc/fused_moe/cake_mxfp4_situ_moe/ruff.toml	1	# The CuTe DSL units under cute/ are generated (rendered from the pinned Cake	COMMENT
LOW	…03a/cake_mxfp4_situ_moe_4dcb477d20a7fa3877b7_kernel.py	1	#	COMMENT
LOW	…03a/cake_mxfp4_situ_moe_fe2484ee16bded7e091e_kernel.py	1	#	COMMENT
LOW	…03a/cake_mxfp4_situ_moe_f1b98a7a271393c79232_kernel.py	1	#	COMMENT
LOW	…03a/cake_mxfp4_situ_moe_cfbc94891bd6ce2c40c8_kernel.py	1	#	COMMENT
LOW	…3a/cake_mxfp4_situ_moe_a01f9a3440d0afbad475_binding.py	1	#	COMMENT
LOW	…03a/cake_mxfp4_situ_moe_01f9756f0519b43b177a_kernel.py	1	#	COMMENT
LOW	…03a/cake_mxfp4_situ_moe_426dd8da968b2c0eebc9_kernel.py	1	#	COMMENT
LOW	…03a/cake_mxfp4_situ_moe_680c12cbf8b8834e8265_kernel.py	1	#	COMMENT
LOW	…3a/cake_mxfp4_situ_moe_bd3d1ae26b97c5fddbb5_binding.py	1	#	COMMENT
LOW	…3a/cake_mxfp4_situ_moe_61a5a1672f9167742db3_binding.py	1	#	COMMENT
LOW	…3a/cake_mxfp4_situ_moe_9efcd6af56b94dc0151d_binding.py	1	#	COMMENT
2990 more matches not shown…

Modern Structural Boilerplate1177 hits · 1130 pts

Severity	File	Line	Snippet	Context
LOW	flashinfer-cubin/flashinfer_cubin/__init__.py	80	__all__ = ["get_cubin_dir", "list_cubins", "get_cubin_path", "CUBIN_DIR"]	CODE
LOW	…che-provider/flashinfer_jit_cache_provider/__init__.py	27	__all__ = ["get_provider"]	CODE
LOW	tests/prims_ts/test_moe_partition_execution.py	547	def _set_activation_params(workload: _Workload, case: str) -> None:	CODE
LOW	tests/prims_ts/test_moe_partition_execution.py	1112	def set_routes(*expert_ids: int) -> None:	CODE
LOW	tests/experimental/test_cake_mhc_train.py	234	def _set_ieee_fp32() -> None:	CODE
LOW	tests/experimental/test_cake_mhc_train_iters.py	159	def _set_ieee_fp32() -> None:	CODE
LOW	…/experimental/b12x/attention/test_attention_mla_api.py	77	def set_split_chunk_config(self, *, kv_chunk_size: int, num_chunks: int) -> None:	CODE
LOW	…/experimental/b12x/attention/test_attention_mla_api.py	89	def set_decode_chunk_config(self, *, kv_chunk_size: int, num_chunks: int) -> None:	CODE
LOW	tests/experimental/deepgemm_common/__init__.py	19	__all__ = [	CODE
LOW	tests/attention/conftest.py	28	logger = logging.getLogger(__name__)	CODE
LOW	tests/moe_ep/_sm90_mxfp4_humming_reference.py	158	__all__ = [	CODE
LOW	tests/comm/test_nvshmem_allreduce.py	12	logger = logging.getLogger(__name__)	CODE
LOW	tests/comm/test_trtllm_mnnvl_allreduce.py	43	def _set_device_from_launcher_local_rank() -> None:	CODE
LOW	tests/comm/cake_a2a_driver.py	22	__all__ = ["CakeA2ADriver", "metainfo_index", "workspace_bytes_per_rank"]	CODE
LOW	tests/comm/test_vllm_custom_allreduce.py	18	logger = logging.getLogger(__name__)	CODE
LOW	tests/utils/test_topk.py	1091	def set_selected_scores(use_prefix: bool) -> None:	CODE
LOW	tests/gemm/test_cute_dsl_gemm_allreduce_two_shot.py	36	logger = logging.getLogger(__name__)	CODE
LOW	tests/gdn/reference_gdn_materialize.py	51	__all__ = ["materialize_ref", "active_request_indices_from"]	CODE
LOW	tests/mhc/test_cake_mhc_det_train.py	212	def _set_ieee_fp32() -> None:	CODE
LOW	tests/mhc/test_cake_mhc_det_train_iters.py	92	def _set_ieee_fp32() -> None:	CODE
LOW	tests/test_helpers/cake_balanced_gqa_plan.py	468	__all__ = [	CODE
LOW	tests/test_helpers/kcp_collective_shims.py	124	__all__ = ["CollectiveCounter", "RecomputePhase"]	CODE
LOW	tests/test_helpers/cake_fmha_capability.py	559	__all__ = [	CODE
LOW	tests/test_helpers/cake_dsa_indexer_reference.py	1609	__all__ = [	CODE
LOW	tests/test_helpers/qwen_3_8_block_cases.py	29	__all__ = [	CODE
LOW	tests/test_helpers/cake_qsa_nvfp4_reference.py	110	__all__ = [	CODE
LOW	tests/test_helpers/cake_kcp_fake_backend.py	698	__all__ = [	CODE
LOW	flashinfer-jit-cache/flashinfer_jit_cache/__init__.py	25	logger = logging.getLogger(__name__)	CODE
LOW	flashinfer-jit-cache/flashinfer_jit_cache/__init__.py	134	__all__ = [	CODE
LOW	examples/pytorch/wan/transformer_wan_flashinfer.py	63	logger = logging.getLogger(__name__)	CODE
LOW	examples/pytorch/wan/transformer_wan_flashinfer.py	303	def set_ulysses_communicator(comm: Optional["UlyssesCommunicator"]) -> None:	CODE
LOW	benchmarks/experimental/b12x/benchmark_iq2_xs_moe.py	77	def set_routes(inputs: Inputs, pattern: str, *, experts: int, offset: int = 0) -> None:	CODE
LOW	scripts/aot_memory_monitor.py	265	def update_stats(stats: MemoryStats, sample: dict[str, int]) -> None:	CODE
LOW	scripts/test_sharding/__init__.py	5	__all__ = ["ALGORITHM_VERSION", "SCHEMA_VERSION"]	CODE
LOW	flashinfer/gated_act_mxfp8.py	158	__all__ = [	STRING
LOW	flashinfer/cake_dcp.py	1401	__all__ = [	CODE
LOW	flashinfer/fp8_quantization.py	20	__all__ = [	CODE
LOW	flashinfer/cake_rmsnorm_silu_gate_det_train.py	152	__all__ = [	CODE
LOW	flashinfer/mega_moe_v3.py	12	__all__ = [	CODE
LOW	flashinfer/compilation_context.py	25	logger = logging.getLogger(__name__)	CODE
LOW	flashinfer/cake_fmha_request_ordered.py	377	__all__ = [	CODE
LOW	flashinfer/gdn_train.py	84	__all__ = [	CODE
LOW	flashinfer/cake_rmsnorm_det_train.py	173	__all__ = [	CODE
LOW	flashinfer/kda_train.py	113	__all__ = [	CODE
LOW	flashinfer/kimi_k3_latent_moe.py	245	__all__ = [	STRING
LOW	flashinfer/cake_minimax_h3.py	594	__all__ = [	CODE
LOW	flashinfer/cake_fused_qk_rope_append.py	171	__all__ = ["cake_fused_qk_rmsnorm_rope_append_paged_kv_cache", "launch_plan"]	STRING
LOW	flashinfer/cake_kda_det_train_kcp.py	454	__all__ = [	CODE
LOW	flashinfer/sm110_xqa.py	13	__all__ = ["prepare", "attention"]	CODE
LOW	flashinfer/kda_prefill_cute.py	592	__all__ = [	CODE
LOW	flashinfer/cake_mhc_det_train.py	598	__all__ = [	STRING
LOW	flashinfer/kimi_k3_vision.py	132	__all__ = ["kimi_k3_vision_tower", "prepare_kimi_k3_vision_tower"]	STRING
LOW	flashinfer/cake_rmsnorm_silu_gate_train.py	181	__all__ = [	CODE
LOW	flashinfer/cake_vsa_sm90.py	100	__all__ = [	CODE
LOW	flashinfer/utils.py	40	logger = logging.getLogger(__name__)	CODE
LOW	flashinfer/utils.py	906	def set_log_level(lvl_str: str) -> None:	CODE
LOW	flashinfer/qsa.py	15	__all__ = ["qsa_forward", "qsa_indexed_forward"]	CODE
LOW	flashinfer/cake_mhc_train.py	47	__all__ = [	CODE
LOW	flashinfer/cake_gdn_det_train_kcp.py	424	__all__ = [	CODE
LOW	flashinfer/fp4_quantization.py	45	__all__ = [	CODE
1117 more matches not shown…

Excessive Try-Catch Wrapping1029 hits · 1015 pts

Severity	File	Line	Snippet	Context
LOW	build_backend.py	142	except Exception:	CODE
LOW	build_backend.py	221	except Exception:	CODE
LOW	build_backend.py	236	except Exception:	CODE
LOW	build_backend.py	355	except Exception:	CODE
LOW	build_backend.py	432	except Exception:	CODE
LOW	build_backend.py	588	except Exception:	CODE
LOW	build_backend.py	1009	except Exception as e:	CODE
LOW	build_utils.py	116	except Exception:	CODE
LOW	docker/test_ci_image.py	105	except Exception as error:	CODE
LOW	tests/kda/test_recurrent_kda_decode_export.py	348	except Exception:	STRING
LOW	tests/kda/test_recurrent_kda_prefill_sm120.py	1061	except Exception as exc: # noqa: BLE001 -- the type and text are the assertion	CODE
LOW	tests/kda/test_recurrent_kda_prefill_sm120.py	1447	except Exception as exc: # noqa: BLE001 -- reported below	CODE
MEDIUM	tests/kda/test_recurrent_kda_prefill_sm120.py	1058	def outcome(backend):	CODE
LOW⚡	tests/prims_ts/test_moe_bf16_support.py	93	except Exception:	CODE
LOW	tests/gdp/test_cake_gdp_prefill.py	967	except Exception as ex: # noqa: BLE001 - profiler unavailable on this host	CODE
MEDIUM	tests/experimental/test_cake_qsa_nvfp4.py	920	def worker(i):	CODE
LOW	tests/experimental/test_cake_qsa_nvfp4.py	926	except Exception as error: # reported by the test thread	CODE
MEDIUM	tests/experimental/test_cake_msa_icp_indexer.py	514	def _get_open_port() -> int:	CODE
LOW	tests/experimental/test_cake_msa_icp_indexer.py	605	except Exception as exc: # pragma: no cover - reported to the parent	CODE
LOW	tests/experimental/test_allreduce_mhc_post_prenorm.py	605	except Exception as exc: # no launcher / single process	CODE
LOW	tests/experimental/test_allreduce_mhc_post_prenorm.py	643	except Exception as exc: # noqa: BLE001	CODE
LOW	tests/experimental/test_cake_dsa_train_api.py	294	except Exception as exc: # noqa: BLE001 -- the shipped kernel predates the mode	CODE
LOW	tests/experimental/test_cake_dsa_train_api.py	330	except Exception as exc: # noqa: BLE001	CODE
MEDIUM	tests/experimental/test_cake_kimi_k3_latent_moe.py	3516	def first_call():	CODE
LOW	tests/experimental/test_cake_kimi_k3_latent_moe.py	3520	except Exception as exc: # noqa: BLE001	CODE
LOW	tests/experimental/test_cake_dsa_train.py	5996	except Exception as exc: # noqa: BLE001 -- the shipped kernel predates the interface	CODE
LOW	tests/experimental/test_cake_dsa_train.py	6028	except Exception as exc: # noqa: BLE001 -- the shipped kernel predates the mode	CODE
LOW	tests/experimental/test_cake_gdn_det_train_kcp_dist.py	352	except Exception: # noqa: BLE001 - a registry without library_name: the manifest's names decide	CODE
LOW	tests/experimental/test_cake_gdn_det_train_kcp_dist.py	483	except Exception: # noqa: BLE001 - foreign objects may refuse attribute access	CODE
LOW	tests/experimental/test_cake_gdn_det_train_kcp_dist.py	512	except Exception: # noqa: BLE001	CODE
LOW	tests/experimental/test_cake_gdn_det_train_kcp_dist.py	1254	except Exception as exc: # noqa: BLE001 - recorded for the decision table	CODE
LOW	tests/experimental/test_cake_gdn_det_train_kcp_dist.py	1356	except Exception as exc: # noqa: BLE001 - recorded; the gate decides below	CODE
LOW	…s/experimental/qwen_3_8_block/test_import_isolation.py	202	except Exception as exc: # noqa: BLE001	CODE
LOW	…s/experimental/b12x/attention/test_qsa_program_keys.py	112	except Exception as error:	CODE
LOW	…l/b12x/attention/test_attention_mla_sglang_boundary.py	32	except Exception as exc: # pragma: no cover - environment-dependent import path	CODE
LOW	…ttention/test_attention_dsa_indexer_sglang_boundary.py	32	except Exception as exc: # pragma: no cover - environment-dependent import path	CODE
LOW	…ttention/test_attention_nsa_backend_sglang_boundary.py	22	except Exception as exc: # pragma: no cover - environment-dependent import path	CODE
LOW	…s/experimental/qwen_3_5_block/test_import_isolation.py	135	except Exception as exc: # noqa: BLE001	CODE
LOW	tests/attention/conftest.py	449	except Exception as e:	CODE
LOW	tests/attention/conftest.py	494	except Exception as error:	CODE
LOW	tests/attention/conftest.py	541	except Exception: # pragma: no cover - diagnostics must never fail the run	CODE
LOW	tests/attention/test_cudnn_decode.py	1216	except Exception as exc: # noqa: BLE001 -- the frontend's error types vary	CODE
LOW	tests/attention/test_deepseek_mla.py	110	except Exception as e:	CODE
MEDIUM	tests/attention/test_deepseek_mla.py	50	def warmup_jit():	CODE
LOW	…sts/trace/test_tgv_gemm_sm100_reference_correctness.py	37	except Exception as exc:	CODE
LOW	tests/trace/test_mm_bf16_fp4_reference_correctness.py	52	except Exception as exc:	CODE
LOW	tests/trace/test_single_decode_reference_correctness.py	31	except Exception as exc:	CODE
LOW	…ize_fp8_append_paged_kv_cache_reference_correctness.py	69	except Exception as exc:	CODE
LOW	…ize_fp8_append_paged_kv_cache_reference_correctness.py	291	except Exception as exc:	CODE
LOW	tests/trace/test_bmm_fp8_reference_correctness.py	38	except Exception as exc:	CODE
LOW	…/trace/test_cutlass_fused_moe_reference_correctness.py	63	except Exception as exc:	CODE
LOW	…s/trace/test_segment_gemm_run_reference_correctness.py	34	except Exception as exc:	CODE
LOW	…s/trace/test_block_sparse_run_reference_correctness.py	58	except Exception as exc:	CODE
LOW	…s/trace/test_xqa_batch_decode_reference_correctness.py	74	except Exception as exc:	CODE
LOW	…ace/test_var_block_sparse_run_reference_correctness.py	67	except Exception as exc:	CODE
LOW	tests/trace/test_bmm_bf16_reference_correctness.py	29	except Exception as exc:	CODE
LOW	tests/trace/test_mm_bf16_reference_correctness.py	29	except Exception as exc:	CODE
LOW	tests/trace/test_fi_trace_template_consistency.py	473	except Exception as exc: # noqa: BLE001	CODE
LOW	…ace/test_xqa_batch_decode_mla_reference_correctness.py	72	except Exception as exc:	CODE
LOW	…trace/test_cudnn_batch_decode_reference_correctness.py	80	except Exception as exc:	CODE
969 more matches not shown…

AI Structural Patterns1134 hits · 877 pts

Severity	File	Line	Context
LOW	tests/kda/test_recurrent_kda_decode_export.py	1366	CODE
LOW	tests/kda/test_recurrent_kda_cudnn_backend.py	83	CODE
LOW	tests/prims_ts/test_batched_gemm_mx.py	189	CODE
LOW	tests/prims_ts/test_batched_gemm_fp4_fc2.py	77	CODE
LOW	tests/prims_ts/test_batched_gemm_fp4_fc2.py	602	CODE
LOW	tests/prims_ts/test_batched_gemm_bf16_fc2.py	65	CODE
LOW	tests/prims_ts/test_batched_gemm_fp4_fc1.py	86	CODE
LOW	tests/prims_ts/test_batched_gemm_bf16_fc1.py	107	CODE
LOW	tests/prims_ts/test_batched_gemm_bf16_fc1.py	271	CODE
LOW	tests/prims_ts/test_batched_gemm_bf16_fc1.py	469	CODE
LOW	tests/gdp/test_prefill_gdp_cudnn_backend.py	62	CODE
LOW	tests/experimental/test_cake_rubin_mla_nvfp4_decode.py	172	CODE
LOW	tests/experimental/test_mok_bf16.py	58	CODE
LOW	tests/experimental/test_cake_lm_head_loss.py	167	CODE
LOW	tests/experimental/test_cake_kda_det_train_kcp.py	3639	CODE
LOW	…al/b12x/attention/test_dsa_indexer_mxfp4_boundaries.py	98	CODE
LOW	…mental/b12x/attention/test_attention_mla_compressed.py	161	CODE
LOW	…s/experimental/b12x/gemm/test_dense_gemm_expected_m.py	1263	CODE
LOW	tests/experimental/b12x/sequence/test_kda_prefill.py	61	CODE
LOW	tests/experimental/b12x/sequence/test_gdn_decode_kda.py	46	CODE
LOW	tests/experimental/b12x/sequence/test_gdn_decode.py	78	CODE
LOW	tests/experimental/b12x/preparation/test_session.py	186	CODE
LOW	…sts/experimental/b12x/moe/test_dynamic_w4a8_trellis.py	51	CODE
LOW	tests/experimental/b12x/moe/test_route_pack_cache.py	7	CODE
LOW	tests/experimental/b12x/moe/test_w4a8_dynamic_kernel.py	100	CODE
LOW	tests/attention/test_modular_fmha_prefill_paged.py	127	CODE
LOW	tests/attention/test_attention_ts_decode.py	1717	CODE
LOW	tests/attention/test_attention_ts_decode.py	6174	CODE
LOW	tests/attention/test_dcp_spec_fp8.py	624	CODE
LOW	tests/attention/test_cake_fmha.py	3984	CODE
LOW	tests/attention/test_attention_ts_block_sparse.py	950	CODE
LOW	tests/attention/test_sparse_mla_blackwell_dsv4_nvfp4.py	187	CODE
LOW	tests/attention/test_fmha_v2_prefill.py	68	CODE
LOW	tests/attention/test_mla_auto_backend.py	471	CODE
LOW	tests/attention/test_xqa.py	124	CODE
LOW	tests/attention/test_attention_ts_sparse_mla.py	34	CODE
LOW	tests/attention/test_bidirectional_ranges_attention.py	139	CODE
LOW	…/test_attention_ts_q_token_kv_block_sparse_metadata.py	283	CODE
LOW	tests/attention/test_cute_dsl_mla_decode.py	327	CODE
LOW	tests/attention/test_trtllm_ragged_kv_stride.py	40	CODE
LOW	tests/attention/test_trtllm_gen_attention_decode.py	631	CODE
LOW	tests/attention/test_trtllm_gen_attention_prefill.py	300	CODE
LOW	tests/attention/test_batch_attention.py	99	CODE
LOW	tests/attention/test_cudnn_decode.py	1371	CODE
LOW	tests/msa_ops/test_msa_nvfp4_decode_sm100.py	578	CODE
LOW	tests/msa_ops/test_blackwell_msa_sm100.py	31	CODE
LOW	…ts/moe_ep/test_moe_ep_sm90_pull_bf16_mega_multirank.py	346	CODE
LOW	…ts/moe_ep/test_moe_ep_sm90_pull_bf16_mega_multirank.py	384	CODE
LOW	tests/moe_ep/test_nvfp4_cutedsl_kernel_vs_reference.py	556	CODE
LOW	tests/moe_ep/test_nvfp4_cutedsl_kernel_vs_reference.py	732	CODE
LOW	…sts/moe_ep/test_moe_ep_sm90_pull_fp8_mega_multirank.py	381	CODE
LOW	…sts/moe_ep/test_moe_ep_sm90_pull_fp8_mega_multirank.py	431	CODE
LOW	tests/moe_ep/test_sm90_push_fp8_kernel.py	110	CODE
LOW	tests/moe_ep/test_sm90_push_fp8_gemm_contract.py	444	CODE
LOW	tests/moe_ep/test_sm90_bf16_push_cake_backend.py	140	CODE
LOW	…sts/moe_ep/test_moe_ep_nvfp4_cutedsl_mega_multirank.py	446	CODE
LOW	…sts/moe_ep/test_moe_ep_nvfp4_cutedsl_mega_multirank.py	1436	CODE
LOW	tests/diffusion_ops/test_minimax_h3_mlp.py	1220	CODE
LOW	tests/mamba/test_chunk_scan_combined.py	27	CODE
LOW	tests/mamba/test_cake_ssd_combined.py	328	CODE
1074 more matches not shown…

Empty Error Handlers155 hits · 775 pts

Severity	File	Line	Snippet	Context
HIGH	build_backend.py	142	except Exception:	CODE
HIGH	build_backend.py	588	except Exception:	CODE
HIGH	flashinfer-cubin/flashinfer_cubin/__init__.py	55	except ImportError:	CODE
HIGH	flashinfer-cubin/flashinfer_cubin/__init__.py	72	except (ImportError, AttributeError):	CODE
HIGH	tests/conftest.py	232	except ImportError:	CODE
HIGH	tests/test_cake_causal_conv1d_det_train_kcp.py	193	except ImportError:	CODE
HIGH	tests/experimental/test_cake_gdn_det_train_kcp_dist.py	1627	except FileNotFoundError:	CODE
HIGH	tests/experimental/b12x/moe/test_nvfp4_split_backend.py	481	except ValueError:	CODE
HIGH	tests/attention/conftest.py	541	except Exception: # pragma: no cover - diagnostics must never fail the run	CODE
HIGH	tests/attention/conftest.py	290	except ImportError: # pragma: no cover - fixture still builds these	CODE
HIGH	tests/trace/example.py	772	except Exception:	CODE
HIGH	tests/trace/example.py	794	except Exception:	CODE
HIGH	tests/trace/example.py	805	except Exception:	CODE
HIGH	tests/trace/example.py	823	except Exception:	CODE
HIGH	tests/trace/example.py	833	except Exception:	CODE
HIGH	tests/trace/example.py	851	except Exception:	CODE
HIGH	tests/trace/example.py	865	except Exception:	CODE
HIGH	tests/trace/example.py	877	except Exception:	CODE
HIGH	tests/msa_ops/test_msa_nvfp4_decode_sm100.py	527	except _LeaveBlock:	CODE
HIGH	tests/moe_ep/test_moe_ep_compute_correctness.py	129	except ImportError:	CODE
HIGH	tests/comm/test_ulysses_communicator.py	2671	except Boom:	CODE
HIGH	tests/comm/test_ulysses_communicator.py	2861	except queue_mod.Empty:	CODE
HIGH	tests/comm/test_ulysses_topology.py	905	except queue_mod.Empty:	CODE
HIGH	tests/gemm/test_ragged_block_scaled_bmm_cutile.py	111	except Exception:	CODE
HIGH	tests/test_helpers/prims_ts_attention_prebuild.py	79	except pytest.skip.Exception:	CODE
HIGH	benchmarks/bench_cake_fused_kda_decode.py	603	except ValueError:	CODE
HIGH	benchmarks/bench_cake_fused_kda_decode.py	1276	except ValueError:	CODE
HIGH	benchmarks/bench_cudnn_wrapper_host.py	58	except (OSError, subprocess.CalledProcessError):	CODE
HIGH	benchmarks/bench_cake_lm_head_loss.py	147	except (TypeError, RuntimeError):	CODE
HIGH	benchmarks/bench_checkpointing_ssu.py	1253	except Exception:	CODE
HIGH	benchmarks/experimental/b12x/benchmark_roce_oneshot.py	74	except OSError:	CODE
HIGH	benchmarks/experimental/b12x/benchmark_roce_oneshot.py	141	except (OSError, subprocess.SubprocessError):	CODE
HIGH	benchmarks/experimental/b12x/benchmark_mla.py	841	except Exception:	CODE
HIGH	benchmarks/routines/attention.py	18	except ImportError:	CODE
HIGH	scripts/run_with_output_watchdog.py	247	except (OSError, subprocess.TimeoutExpired):	CODE
HIGH	scripts/unit_test_runner.py	204	except Exception:	CODE
HIGH	scripts/pr_checks/check_api_docs.py	47	except Exception:	CODE
HIGH	scripts/pr_checks/experimental_test_scope.py	295	except ValueError:	CODE
HIGH	scripts/pr_checks/experimental_test_scope.py	308	except ValueError:	CODE
HIGH	scripts/pr_checks/experimental_test_scope.py	333	except ValueError:	CODE
HIGH	scripts/pr_checks/experimental_test_scope.py	258	except ValueError:	CODE
HIGH	scripts/test_sharding/junit.py	152	except (OSError, AttributeError, json.JSONDecodeError):	CODE
HIGH	scripts/test_sharding/workers.py	189	except (OSError, ValueError):	CODE
HIGH	scripts/test_sharding/state.py	43	except FileExistsError:	CODE
HIGH	flashinfer/api_logging.py	2515	except Exception:	CODE
HIGH	flashinfer/collect_env.py	461	except Exception:	CODE
HIGH	flashinfer/__init__.py	485	except (ImportError, AttributeError):	CODE
HIGH	flashinfer/__init__.py	634	except Exception:	CODE
HIGH	flashinfer/cake_kda_tf32_runtime.py	1365	except OSError:	CODE
HIGH	flashinfer/cake_kda_tf32_runtime.py	1234	except Exception:	CODE
HIGH	flashinfer/cake_gdn_det_train_kcp.py	395	except ImportError:	CODE
HIGH	flashinfer/fi_trace.py	139	except AttributeError:	CODE
HIGH	flashinfer/fi_trace.py	164	except Exception:	CODE
HIGH	flashinfer/prims_ts/moe/tools/sweep/measure.py	58	except (ImportError, PackageNotFoundError, RuntimeError, ValueError):	CODE
HIGH	flashinfer/prims_ts/moe/tools/sweep/engine.py	1195	except (ImportError, RuntimeError):	CODE
HIGH	flashinfer/prims_ts/moe/tools/sweep/engine.py	1308	except Exception:	CODE
HIGH	flashinfer/prims_ts/moe/tools/sweep/engine.py	1147	except (ImportError, RuntimeError):	CODE
HIGH	flashinfer/prims_ts/moe/tools/sweep/workers.py	475	except (AttributeError, ImportError, RuntimeError):	CODE
HIGH	flashinfer/prims_ts/moe/tools/sweep/report.py	193	except (AttributeError, ImportError, RuntimeError):	CODE
HIGH	flashinfer/prims_ts/moe/tools/sweep/report.py	216	except Exception:	CODE
95 more matches not shown…

Hallucination Indicators84 hits · 720 pts

Severity	File	Line	Snippet	Context
CRITICAL	flashinfer/fused_moe/runners.py	4675	activation_name = self.config.activation.type.name.lower()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	783	txl.ptx.st.release.gpu.global_.s32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	791	txl.ptx.fence.proxy.async_.shared__cta()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	798	txl.ptx.tcgen05.wait__ld.sync.aligned()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1018	txl.ptx.tcgen05.wait__ld.sync.aligned()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1022	txl.ptx.tcgen05.wait__ld.sync.aligned()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1026	txl.ptx.tcgen05.wait__st.sync.aligned()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1075	txl.ptx.ldmatrix.sync.aligned.m8n8.x4.trans.shared.b16(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1080	txl.ptx.stmatrix.sync.aligned.m8n8.x4.shared.b16(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1086	txl.ptx.movmatrix.sync.aligned.m8n8.trans.b16(dst[z], src[z])	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1094	txl.ptx.mma.sync.aligned.m16n8k16.row.col.f32.bf16.bf16.f32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1233	txl.ptx.st.shared.v4.u32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1236	txl.ptx.st.shared.v4.u32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1239	txl.ptx.fence.proxy.async_.shared__cta()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1441	txl.ptx.fence.proxy.async_.shared__cta()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1449	txl.ptx.fence.proxy.async_.shared__cta()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1451	txl.ptx.stmatrix.sync.aligned.m8n8.x4.shared.b16(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1458	txl.ptx.fence.proxy.async_.shared__cta()	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1585	txl.ptx.ld.shared.v4.b32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1590	txl.ptx.st.shared.v4.b32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1596	txl.ptx.rcp.approx.ftz.f32(r, x)	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1653	txl.ptx.ld.global_.v4.f32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1660	txl.ptx.ld.global_.v4.f32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1667	txl.ptx.st.shared.v4.f32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1674	txl.ptx.st.shared.v4.f32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1684	txl.ptx.ld.shared.v4.f32(	CODE
CRITICAL⚡	flashinfer/kda_kernels/tirx/fused.py	1687	txl.ptx.ld.shared.v4.f32(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	241	txl.ptx.fence.proxy.async_.shared__cta()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	243	txl.ptx.fence.mbarrier_init.release.cluster()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	350	txl.ptx.mbarrier.try_wait.parity.shared.b64(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	435	txl.ptx.cvt.rn.bf16x2.f32(r, hi, lo)	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	455	txl.ptx.ex2.approx.ftz.f32(r, x)	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	507	txl.ptx.mbarrier.arrive.expect_tx.shared.b64(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	517	txl.ptx.fence.proxy.async_.shared__cta()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	534	txl.ptx.mbarrier.arrive.expect_tx.shared.b64(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	695	txl.ptx.ld.global_.v4.f32(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	707	txl.ptx.tcgen05.wait__st.sync.aligned()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	727	txl.ptx.ld.acquire.gpu.global_.s32(fl, flags.ptr_to([cta]))	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	746	txl.ptx.tcgen05.wait__ld.sync.aligned()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	755	txl.ptx.st.global_.v4.f32(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	824	txl.ptx.stmatrix.sync.aligned.m8n8.x4.trans.shared.b16(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	831	txl.ptx.fence.proxy.async_.shared__cta()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	851	txl.ptx.cp.async_.bulk.commit_group()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	852	txl.ptx.cp.async_.bulk.wait_group.read(0)	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	869	txl.ptx.ld.shared.v4.b32(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	872	txl.ptx.st.global_.v4.b32(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	894	txl.ptx.tcgen05.wait__ld.sync.aligned()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	897	txl.ptx.tcgen05.wait__st.sync.aligned()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	909	txl.ptx.tcgen05.wait__ld.sync.aligned()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	970	txl.ptx.tcgen05.wait__ld.sync.aligned()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	973	txl.ptx.tcgen05.wait__st.sync.aligned()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	990	txl.ptx.ld.shared.v4.f32(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	1000	txl.ptx.mul.rn.ftz.f32x2(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	1253	txl.ptx.ld.shared.v2.f32(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	1266	txl.ptx.tcgen05.wait__ld.sync.aligned()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	1269	txl.ptx.ld.shared.v2.f32(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	1307	txl.ptx.tcgen05.wait__ld.sync.aligned()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	1340	txl.ptx.stmatrix.sync.aligned.m8n8.x4.shared.b16(	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	1393	txl.ptx.fence.proxy.async_.shared__cta()	CODE
CRITICAL	flashinfer/kda_kernels/tirx/fused.py	1486	txl.ptx.rsqrt.approx.ftz.f32(norm, n0 + n1 + txl.float32(EPS))	CODE
24 more matches not shown…

Cross-Language Confusion104 hits · 514 pts

Severity	File	Line	Snippet	Context
HIGH	…ts/experimental/test_cake_dense_projection_det_gemm.py	7001	# the measured rules never key one weight-gradient row in both orientations (the decision would be undefined)	COMMENT
HIGH	tests/experimental/test_cake_lm_head_loss.py	1964	# and in the fp32 fused cast; the bf16 cast keeps the 1-D raster; a pinned ``group_n`` (0 or null) keeps the default	COMMENT
HIGH⚡	…ts/experimental/b12x/preparation/test_delta_prefill.py	46	tensors["recurrent_state"][null].fill_(float("nan"))	CODE
HIGH⚡	…ts/experimental/b12x/preparation/test_delta_prefill.py	47	tensors["initial_state_indices"].fill_(null)	CODE
HIGH⚡	…ts/experimental/b12x/preparation/test_delta_prefill.py	53	expected, state = oracle(case, tensors, null_state_index=null)	CODE
HIGH	tests/attention/test_cake_fused_qk_rope_append_pdl.py	102	TVM_FFI_ICHECK(out.device().device_type == kDLCUDA && out.IsContiguous() &&	CODE
HIGH	tests/attention/test_cake_fused_qk_rope_append_pdl.py	106	TVM_FFI_ICHECK(bytes == in.numel() * in.dtype().bits / 8 && bytes % 16 == 0)	CODE
HIGH	tests/attention/test_cake_fmha.py	731	" CAKE_FMHA_HAS_SINK == 1 && CAKE_FMHA_HAS_WINDOW == 0 && "	CODE
HIGH	tests/attention/test_cake_fmha.py	834	" CAKE_FMHA_HAS_SINK == 0 && CAKE_FMHA_HAS_WINDOW == 0 && "	CODE
HIGH	…s/moe_ep/test_moe_ep_sm90_pull_mxfp4_mega_multirank.py	202	"""Canonical ``gate \|\| up`` rows to the kernel's ``gate8, up8`` order."""	STRING
HIGH⚡	tests/moe_ep/test_nvlink_one_sided_lamport.py	127	" if (rank >= 0 && token % 7 == 3) rank = (rank + 1) % p.ep_size;\n",	CODE
HIGH⚡	tests/moe_ep/test_nvlink_one_sided_lamport.py	133	" if (token == 0 && target_mask) {\n"	CODE
HIGH⚡	tests/moe_ep/test_nvlink_one_sided_lamport.py	143	" if (group == 0 && num_sources > 2) {\n"	CODE
HIGH⚡	tests/cpu/trace/test_paged_mqa_logits_checker.py	80	With rnorm == 0 the relative error is undefined; substituting 0.0 made	STRING
HIGH	tests/cpu/jit/test_cake_bgmv_moe_jit.py	277	"major == CAKE_BGMV_MOE_CC_MAJOR && minor == CAKE_BGMV_MOE_CC_MINOR" in binding	CODE
HIGH	tests/cpu/jit/test_cake_bgmv_moe_jit.py	765	"major == CAKE_BGMV_MOE_CC_MAJOR && minor == CAKE_BGMV_MOE_CC_MINOR" in binding	CODE
HIGH	tests/cpu/jit/test_cake_bgmv_moe_jit.py	783	"shrink_form == kShrinkFormPrefill \|\| shrink_form == kShrinkFormPrefillS3"	CODE
HIGH	tests/cpu/jit/test_cake_bgmv_moe_jit.py	809	assert "pdl_mode == 1 \|\| pdl_mode == 2" in binding	CODE
HIGH	tests/mamba/test_cake_ssd_combined.py	2084	state undefined) -- both documented in D12, both flagged. A clean call	STRING
HIGH	tests/comm/test_trtllm_allreduce_reduction.py	43	TORCH_CHECK(num_threads > 0 && num_threads <= 1024, "invalid block size");	CODE
HIGH	tests/comm/test_cake_moe_alltoall_fused_adapter.py	229	"(params.top_k == 1 \|\| params.top_k == 2 \|\| params.top_k == 4) &&"	CODE
HIGH	tests/comm/test_cake_moe_alltoall_fused_adapter.py	233	"params.dtype == nvinfer1::DataType::kBF16 && params.elements_per_token % 8 == 0;"	CODE
HIGH	tests/comm/test_cake_moe_alltoall_fused_adapter.py	346	assert "if (bid == 0 && blockIdx.y == 0)" in topk6_source	CODE
HIGH	tests/comm/test_cake_moe_alltoall_fused_adapter.py	426	"params.local_num_tokens > 0 && params.elements_per_token % 32 == 0 && "	CODE
HIGH	tests/comm/test_trtllm_mnnvl_allreduce.py	318	ordinary data; a `val == 0.F && signbit(val)` check would FTZ-flush	STRING
HIGH	tests/comm/test_cake_ulysses_a2a.py	348	TORCH_CHECK(dest.scalar_type() == at::kInt && status.scalar_type() == at::kInt &&	CODE
HIGH	tests/comm/test_cake_ulysses_a2a.py	351	TORCH_CHECK(dest.numel() >= 1 && dest.numel() <= 160, "entry_poll: 1 .. 160 CTAs");	CODE
HIGH	tests/utils/test_jit_example.py	193	if (qo_idx < qo_len && kv_idx < kv_len) {	CODE
HIGH	tests/utils/test_norm.py	301	TORCH_CHECK(flag.scalar_type() == at::kInt && observed.scalar_type() == at::kInt,	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_gemm.py	3612	"x.size(1) == 5120 \|\| x.size(1) == 7168 \|\| x.size(1) == 14336" not in combined	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_gemm.py	3662	assert "x.size(1) == 5120 \|\| x.size(1) == 5376 \|\|" not in combined	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_gemm.py	3677	assert "if constexpr (K == 5120 \|\| K == 5376 \|\| K == 7168)" in producer	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_host.py	2897	if (tactic < 0 \|\| tactic >= kNvfp4SvdquantGemmSm120NumTactics)	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_host.py	3431	assert(argc == 4 \|\| argc == 5);	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_host.py	3448	if (entry == 3 && TEST_RANK != 32) {	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_host.py	3479	assert(backend_calls == (rejected \|\| empty ? 0 : (entry == 3 ? 2 : 1)));	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_host.py	3645	assert(state.initialized && state.device_id == current_device);	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_host.py	3666	assert(failed && live_resources == before);	CODE
HIGH	tests/gemm/test_nvfp4_svdquant_host.py	3671	assert(recovered.initialized && live_resources == before + 6);	CODE
HIGH	tests/gdn/test_decode_ucache.py	57	source env.sh && pytest tests/gdn/test_decode_ucache.py -v	STRING
HIGH	tests/jit/test_flash_kda_packed_t1_jit.py	207	assert "minor == 0 \|\| minor == 3 \|\| minor == 7" in binding	CODE
HIGH	tests/jit/test_flash_kda_decode_jit.py	144	assert "else if (has_token && k_lane == 0)" in generated_body	CODE
HIGH⚡	tests/jit/test_flash_kda_decode_jit.py	187	assert "major == 10 && (minor == 0 \|\| minor == 3)" in common	CODE
HIGH⚡	tests/jit/test_flash_kda_decode_jit.py	188	assert "major == 10 && minor == expected_minor" in common	CODE
HIGH⚡	tests/jit/test_flash_kda_decode_jit.py	192	assert "GateKind == 0 \|\| GateKind == 1 \|\| GateKind == 2" in common	CODE
HIGH	tests/moe/test_cake_warp_decode_pdl_gpu.py	111	TVM_FFI_ICHECK(tensor.device().device_type == kDLCUDA && tensor.IsContiguous())	CODE
HIGH	benchmarks/bench_monomoe.py	86	Uses mono_moe's `[gate \|\| up]` fc1 half-ordering and the given (already	STRING
HIGH	benchmarks/bench_cake_dsa_train.py	1668	"skipped: hole rows (cuDNN DSA takes dense top-k rows; -1 slots before valid keys are undefined)"	CODE
HIGH	benchmarks/routines/attention.py	4091	# Leave empty (null) when not explicitly overridden so legacy	COMMENT
HIGH	flashinfer/collect_env.py	156	_run("grep -sq -e docker -e containerd -e kubepods /proc/1/cgroup && echo 1")	CODE
HIGH	flashinfer/kda_decode.py	1037	# slots are padding (the vLLM drop-in reserves slot 0 as null).	STRING
HIGH	flashinfer/prims_ts/moe/runner.py	282	"""Pick finalize weights. Match C++ ``ndim()==2 && size(0)>0``.	STRING
HIGH	…experimental/cake_moe_grouped_det_gemm/cake_backend.py	755	if (threadIdx.x != 0 \|\| smid >= 1024) return;	CODE
HIGH	…/experimental/b12x/attention/_shared/mla/prefill_mg.py	181	XV path enables it (USE_WFP8_ROW_XOR = DUAL_CACHE && pbs_extra==2,	STRING
HIGH	…fer/experimental/b12x/testing/reference/prefill_ref.py	38	if attn_sink != null:	STRING
HIGH	…imental/b12x/sequence/gdn_prefill/_parallel_kernels.py	600	result = slot == Int64(self.null)	CODE
HIGH	…hinfer/experimental/b12x/sequence/gdn_prefill/_impl.py	83	object.__setattr__(self, "null_state_index", null)	CODE
HIGH	…hinfer/experimental/b12x/sequence/kda_prefill/_impl.py	66	object.__setattr__(self, "null_state_index", null)	CODE
HIGH	…er/experimental/b12x/sequence/kda_prefill/reference.py	117	initial state from ``initial_state_indices[i]`` (zero when null), writes its	STRING
HIGH	…experimental/b12x/moe/_shared/kernels/w4a16/prepare.py	2242	``payload`` uses the reference container order ``record0 \|\| record1`` and	STRING
44 more matches not shown…

AI Slop Vocabulary142 hits · 422 pts

Severity	File	Line	Snippet	Context
MEDIUM	tests/experimental/test_cake_gdn_det_train_kcp.py	66	import gdn_reference # the FP64 reference of the validation harness	CODE
MEDIUM⚡	tests/experimental/test_cake_stepfun_moe_api.py	245	# Case construction (trtllm-gen harness data, unified-API inputs)	COMMENT
MEDIUM	tests/experimental/test_cake_stepfun_moe_api.py	129	# One entry per public precision: the quantization, the trtllm-gen test harness that produces	COMMENT
MEDIUM	tests/experimental/test_cake_stepfun_moe_api.py	172	# The per-tensor FP8 harness keeps the native plural key spelling for the scale scalars.	COMMENT
MEDIUM	tests/experimental/test_cake_stepfun_moe_api.py	221	# The harness reference is torch fp32 on the device; pin the matmul precision so the error	COMMENT
MEDIUM	tests/experimental/test_cake_dsa_train_api.py	79	# for the iid top-k-2048 configuration and are enforced by the project harness, not per test shape.	COMMENT
MEDIUM	tests/experimental/test_cake_dsa_train_api.py	83	# harness values); rope parts at twice the latent gate, peaked attention at one common gate.	COMMENT
MEDIUM	tests/experimental/test_cake_dsa_train.py	78	# for the iid top-k-2048 configuration and are enforced by the project harness, not per test shape.	COMMENT
MEDIUM	tests/experimental/test_cake_dsa_train.py	82	# harness values); rope parts at twice the latent gate, peaked attention at one common gate.	COMMENT
MEDIUM	tests/experimental/test_cake_kda_det_train_kcp.py	334	# byte totals over the ranks match the harness	COMMENT
MEDIUM	tests/experimental/test_cake_kda_det_train_kcp.py	368	# --------------------------------------------------------------------------- ownership, plan rule, limits (no harness)	COMMENT
MEDIUM	tests/experimental/test_cake_kda_det_train_kcp.py	4444	) # natural-log decay in (-5, 0), the harness's default regime	CODE
MEDIUM	tests/experimental/b12x/attention/test_fused_indexer.py	347	# value-multiset parity (tie-robust); fp8 e4m3 dot in f32 matches the MMA.	COMMENT
MEDIUM	tests/experimental/b12x/attention/test_fused_indexer.py	717	# returned index (rather than asserting an index order) keeps this robust	COMMENT
MEDIUM	…ental/b12x/attention/test_attention_dsa_indexer_api.py	1470	# tie-robust check is that the selected logit-VALUE multiset matches the	COMMENT
MEDIUM	tests/experimental/b12x/moe/test_nvfp4_phase_kernels.py	556	# direct torch chain must match the oracle too (sanity of the harness)	COMMENT
MEDIUM⚡	tests/msa_ops/test_msa_nvfp4_decode_sm100.py	407	# same numbers a consumer's capability probe and the kernel-level A/B harness	COMMENT
MEDIUM⚡	tests/msa_ops/test_msa_nvfp4_decode_sm100.py	1856	# A warm launch is not a caller's call, and the A/B harness differences	COMMENT
MEDIUM	tests/moe_ep/test_moe_ep_sm120_w4a8_cake_mega.py	504	# Element-wise bound scaled to the output range (the Cake harness' real-weight rule: max \|error\| <= 2 % of	COMMENT
MEDIUM	tests/moe_ep/test_sm120_w4a8_cake_fixture.py	188	# -- canonical -> kernel layout reproduces the generator harness's operands --	COMMENT
MEDIUM	tests/moe_ep/test_sm120_w4a8_cake_fixture.py	279	# -- bit-exact parity with the generator harness (optional) ---------------	COMMENT
MEDIUM	tests/cpu/autotuner/test_autotuner_distributed.py	228	# independent of how the harness imported the tests (a ``spawn`` worker that	COMMENT
MEDIUM	…sion_ops/test_minimax_h3_sm120_nvfp4_fp8pv_programs.py	268	# default scale is spelled the same way, so the two arms of the kitchen's parity harness pass one float.	COMMENT
MEDIUM	tests/mamba/test_philox_rounding.py	794	"""Software e4m3 SR matches hardware PTX path bitwise across a comprehensive	STRING
LOW	tests/mamba/triton_reference/causal_conv1d_triton.py	171	# first chunk and does not have prior-token, so just set to 0	COMMENT
MEDIUM⚡	tests/comm/test_cake_fp4_allreduce.py	1250	# process harness	COMMENT
MEDIUM	tests/gemm/test_nvfp4_svdquant_host.py	2859	# Optional helpers let the same behavioral harness run against the pre-fix source.	COMMENT
MEDIUM⚡	tests/gemm/test_unified_gemm_fuzz.py	595	# + tight tolerance. Same model as the MoE harness: with exact-grid inputs the oracle measures the	COMMENT
MEDIUM	tests/gemm/test_unified_gemm_fuzz.py	549	# layout_128x4 / mxfp4 default this harness prepares). Toggling it on the cutlass/cudnn/auto	COMMENT
MEDIUM⚡	tests/gdn/test_decode_delta_rule.py	1482	# Test MTP kernel with FP32 state, cache ON, state update ON (comprehensive)	COMMENT
MEDIUM	tests/test_helpers/cake_stepfun_moe_reference.py	106	# Per-tensor FP8 calibration of the FC1 output (the previous test's harness: StepFun is bounded	COMMENT
LOW	tests/test_helpers/alibi_reference.py	78	# Since it's causal mask we can just use $[0, 1, \dots, N]$ too.	STRING
MEDIUM	tests/test_helpers/cake_profiler.py	3	"""CUDA launch records of one call under ``torch.profiler``, robust to a session that loses its device records.	STRING
MEDIUM	tests/test_helpers/cake_kcp_fake_backend.py	188	# the harness seeds its jitter by (row name, W, document index of the pack it cuts)	COMMENT
MEDIUM	tests/moe/test_trtllm_moe_valid_dims.py	187	# harness / CUDA-only deps on machines that only run the unit-level checks above.	COMMENT
MEDIUM	tests/moe/test_trtllm_moe_valid_dims.py	523	# GEMM2 output width equal to hidden_size, which the harness FP4	COMMENT
MEDIUM	tests/moe/test_trtllm_moe_valid_dims.py	705	# The harness FP4 prepare_static_weights_for_kernel reshapes GEMM2 by a single	COMMENT
MEDIUM⚡	tests/moe/test_unified_moe_fuzz.py	3424	# different shape would produce a wrong answer here. (Shares the harness's snap/reference/prep.)	COMMENT
MEDIUM	tests/moe/test_unified_moe_fuzz.py	1565	# Covers the in-kernel-routing bug cluster the pre-routed harness structurally can't reach:	COMMENT
MEDIUM	tests/moe/test_unified_moe_fuzz.py	1721	# separate validation, and EP collectives are out of scope for this single-GPU harness.	COMMENT
MEDIUM	tests/moe/test_unified_moe_fuzz.py	1871	# In-kernel routing (FromLogits) headline cases the pre-routed harness can't reach:	COMMENT
MEDIUM	tests/moe/test_unified_moe_fuzz.py	1568	RoutingMethodType.RenormalizeNaive, # == the harness's original host routing	CODE
MEDIUM	benchmarks/bench_gdn_prefix_materialize.py	177	# --timing graph the harness captures on a side stream; a stream bound	COMMENT
MEDIUM⚡	benchmarks/bench_mamba_ssd_combined.py	257	# Layer 2: bench_mode — batched or varlen harness	COMMENT
MEDIUM	benchmarks/diag_cake_dsa_indexer_sampling.py	823	# The benchmark harness bounds a row at max(seg_k); time the arms with the call's own bound so the	COMMENT
MEDIUM	benchmarks/bench_silu_and_mul_nvfp4_quantize.py	132	# more comprehensive correctness coverage.	COMMENT
MEDIUM⚡	benchmarks/bench_cake_dense_projection_det_gemm.py	115	# Inputs (the Cake eval harness distributions)	COMMENT
MEDIUM	benchmarks/bench_recurrent_kda_prefill.py	660	# harness while keeping route logging out of every timed call.	COMMENT
MEDIUM	…chmarks/experimental/b12x/benchmark_paged_attention.py	2069	# copies introduced solely by the benchmark harness.	COMMENT
MEDIUM	benchmarks/routines/kda.py	329	# difference between the two APIs, not an artifact of this harness, and it	COMMENT
MEDIUM	benchmarks/routines/kda.py	560	# published combination. They need a protocol this harness cannot	COMMENT
MEDIUM	benchmarks/routines/kda.py	567	# what this harness is actually able to set up.	COMMENT
MEDIUM	scripts/task_test_moe_ep_smoke.sh	4	# Run the moe_ep smoke + multi-rank harness via torchrun.	COMMENT
LOW	flashinfer/api_logging.py	1569	# The safest way is to just call the class constructor.	COMMENT
MEDIUM	flashinfer/aot.py	368	# The holistic (persistent) batch-attention kernel	COMMENT
MEDIUM	flashinfer/utils.py	1756	# Hopper and newer supports a more nuanced query of confidential	COMMENT
MEDIUM	flashinfer/cake_sampling.py	1033	# in the interleaved CUPTI harness, +0.8 / +2.4 % in two idle-node matrices at k = 10 / 50), while the 16 %-full	COMMENT
MEDIUM	flashinfer/autotune_cache.py	168	# execution_mode. Used by the accuracy harness to A/B	COMMENT
MEDIUM⚡	flashinfer/prims_ts/batched_gemm/batched_gemm_run.py	4377	# Perf-harness interface (``@pytest.mark.perf`` profiler sweep).	COMMENT
MEDIUM	flashinfer/prims_ts/moe/tools/sweep/calibration.py	477	"""Derive a robust symmetric comparison band from repeated timings.	STRING
82 more matches not shown…

Structural Annotation Overuse211 hits · 314 pts

Severity	File	Line	Snippet	Context
LOW	…_llm/kernels/moe/communication/moeAlltoAllCftManager.h	124	// Step 1: Load LE APIs from driver via cuGetProcAddress.	COMMENT
LOW	…_llm/kernels/moe/communication/moeAlltoAllCftManager.h	158	// Step 2: Create this rank's unicast LE and bind it to fabric memory.	COMMENT
LOW	…_llm/kernels/moe/communication/moeAlltoAllCftManager.h	406	// Step 3: Exchange LE handles across all ranks.	COMMENT
LOW	tests/attention/test_rope.py	915	# Step 1: Pre-populate cache with existing tokens	COMMENT
LOW	tests/attention/test_rope.py	1126	# Step 2: Append new tokens to the pre-populated cache	COMMENT
LOW	…_quantize_fp8_append_paged_kv_cache_h8_kv2_rope64.json	194	"reference": "from __future__ import annotations\nimport math\nimport torch\nimport torch.nn.functional as F\n\n@torch	CODE
LOW	…tize_fp8_append_paged_kv_cache_h8_kv2_rope64_ps16.json	195	"reference": "from __future__ import annotations\nimport math\nimport torch\nimport torch.nn.functional as F\n\n@torch	CODE
LOW⚡	tests/utils/test_fp4_quantize.py	404	# Step 1: Quantize with fp4_quantize	COMMENT
LOW⚡	tests/utils/test_fp4_quantize.py	414	# Step 2: Dequantize with e2m1_and_ufp8sf_scale_to_float	COMMENT
LOW	tests/utils/test_fp4_quantize.py	429	# Step 3: Compare results	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1081	# Step 1: Compute g = -exp(A_log) * softplus(a + dt_bias)	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1136	# Step 1: Apply gating to hidden state: h *= exp(g)	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1139	# Step 2: Delta rule: v -= k^T @ h ([B,H,K] x [B,H,K,V] -> [B,H,V])	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1142	# Step 3: Apply beta gating: v *= beta	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1145	# Step 4: Update hidden state with outer product: h += k ⊗ v_new	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1148	# Step 5: Compute output: o = q^T @ h	COMMENT
LOW	tests/gdn/reference_delta_rule.py	1098	# Step 2: Compute beta = sigmoid(b)	COMMENT
LOW	.claude/skills/self-review/SKILL.md	16	## Step 1: Get the diff	COMMENT
LOW	.claude/skills/self-review/SKILL.md	29	## Step 2: Apply the review guidance	COMMENT
LOW	.claude/skills/self-review/SKILL.md	52	## Step 3: Apply the PR rules	COMMENT
LOW	.claude/skills/self-review/SKILL.md	71	## Step 4: Report, then fix	COMMENT
LOW	.claude/skills/benchmark-kernel/SKILL.md	63	### Step 1: Choose Your Test Routine	COMMENT
LOW	.claude/skills/benchmark-kernel/SKILL.md	70	### Step 2: Run a Single Benchmark	COMMENT
LOW	.claude/skills/benchmark-kernel/SKILL.md	118	### Step 3: Understand the Output	COMMENT
LOW	.claude/skills/benchmark-kernel/SKILL.md	134	### Step 4: Run Batch Benchmarks	COMMENT
LOW	.claude/skills/benchmark-kernel/SKILL.md	156	### Step 5: Common Flags	COMMENT
LOW	.claude/skills/benchmark-kernel/SKILL.md	174	### Step 1: Write Your Benchmark Script	COMMENT
LOW	.claude/skills/benchmark-kernel/SKILL.md	210	### Step 2: Run Your Benchmark	COMMENT
LOW	.claude/skills/benchmark-kernel/SKILL.md	229	### Step 3: Advanced Options	COMMENT
LOW	.claude/skills/debug-cuda-crash/SKILL.md	24	## Step 1: Enable API Logging	COMMENT
LOW	.claude/skills/debug-cuda-crash/SKILL.md	109	## Step 2: Reproduce the Crash	COMMENT
LOW	.claude/skills/debug-cuda-crash/SKILL.md	146	## Step 3: Common CUDA Errors and How to Debug	COMMENT
LOW	.claude/skills/debug-cuda-crash/SKILL.md	249	## Step 4: Multi-Process Debugging	COMMENT
LOW	.claude/skills/debug-cuda-crash/SKILL.md	268	## Step 5: Advanced Debugging with compute-sanitizer	COMMENT
LOW	.claude/skills/debug-cuda-crash/SKILL.md	309	## Step 6: Kernel-Level Debugging with printf()	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	18	## Step 1: Define CUDA Kernel in `include/`	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	76	## Step 2: Create Launcher in `csrc/`	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	163	## Step 3: Create TVM-FFI Binding in `csrc/`	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	183	## Step 4: Create JIT Generator (No Jinja for Simple Case)	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	387	## Step 5: Create Python API in `flashinfer/`	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	628	## Step 6: Add a Trace Template	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	776	## Step 7: Write Tests in `tests/`	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	846	## Step 7: Register in AOT	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	872	## Step 8: Export API	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	883	## Step 9: Run and Test	COMMENT
LOW	.claude/skills/add-cuda-kernel/SKILL.md	893	## Step 10: Add Benchmark	COMMENT
LOW	docs/tutorials/jax_tvm_ffi/flashinfer_jax_tvm_ffi.py	176	# Step 1 - Compile and load the TVM FFI module	COMMENT
LOW	docs/tutorials/jax_tvm_ffi/flashinfer_jax_tvm_ffi.py	197	# Step 2 - Register as a JAX FFI target	COMMENT
LOW	docs/tutorials/jax_tvm_ffi/flashinfer_jax_tvm_ffi.py	240	# Step 3 - Call as a regular JAX expression	COMMENT
LOW⚡	benchmarks/bench_cute_dsl_add_rmsnorm_fp4quant.py	149	# Step 1: Add	COMMENT
LOW⚡	benchmarks/bench_cute_dsl_add_rmsnorm_fp4quant.py	151	# Step 2: RMSNorm	COMMENT
LOW⚡	benchmarks/bench_cute_dsl_add_rmsnorm_fp4quant.py	153	# Step 3: FP4 quantize (with global_scale for NVFP4)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	179	# Step 1: Apply decay to hidden state: h *= exp(g)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	182	# Step 2: Delta rule: v -= sum(h * k, dim=0) = k @ h	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	187	# Step 3: Apply beta gating: v *= beta	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	190	# Step 4: Update hidden state: h += outer(k, v) = k[:, None] * v[None, :]	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	193	# Step 5: Compute output: o = q @ h = sum(q[:, None] * h, dim=0)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	361	# Step 1: Apply decay: h *= exp(g)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	364	# Step 2: Delta rule: v -= h @ k (h is [BV, BK], k is [BK])	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	369	# Step 3: Apply beta gating	COMMENT
151 more matches not shown…

Verbosity Indicators184 hits · 277 pts

Severity	File	Line	Snippet	Context
LOW	…mha_v2/fused_multihead_flash_attention_kernel_noloop.h	286	// Do we need to check if there are negative inf for softmax row_max ?	COMMENT
LOW	…/fused_multihead_flash_attention_kernel_noloop_tiled.h	92	// Do we need to check if there are negative inf for softmax row_max ?	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/epilogue.h	73	// Whether we need to check if local_max could be -inf or not.	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/epilogue.h	543	// Whether we need to check if local_max could be -inf or not.	COMMENT
LOW	csrc/fmha_v2/fmha/warpspec/epilogue.h	734	// Whether we need to check if local_max could be -inf or not.	COMMENT
LOW	…_llm/kernels/moe/communication/moeAlltoAllCftManager.h	124	// Step 1: Load LE APIs from driver via cuGetProcAddress.	COMMENT
LOW	…_llm/kernels/moe/communication/moeAlltoAllCftManager.h	158	// Step 2: Create this rank's unicast LE and bind it to fabric memory.	COMMENT
LOW	…_llm/kernels/moe/communication/moeAlltoAllCftManager.h	406	// Step 3: Exchange LE handles across all ranks.	COMMENT
LOW	tests/attention/test_rope.py	915	# Step 1: Pre-populate cache with existing tokens	COMMENT
LOW	tests/attention/test_rope.py	1126	# Step 2: Append new tokens to the pre-populated cache	COMMENT
LOW	…_quantize_fp8_append_paged_kv_cache_h8_kv2_rope64.json	194	"reference": "from __future__ import annotations\nimport math\nimport torch\nimport torch.nn.functional as F\n\n@torch	CODE
LOW	…tize_fp8_append_paged_kv_cache_h8_kv2_rope64_ps16.json	195	"reference": "from __future__ import annotations\nimport math\nimport torch\nimport torch.nn.functional as F\n\n@torch	CODE
LOW⚡	tests/utils/test_fp4_quantize.py	404	# Step 1: Quantize with fp4_quantize	COMMENT
LOW⚡	tests/utils/test_fp4_quantize.py	414	# Step 2: Dequantize with e2m1_and_ufp8sf_scale_to_float	COMMENT
LOW	tests/utils/test_fp4_quantize.py	429	# Step 3: Compare results	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1081	# Step 1: Compute g = -exp(A_log) * softplus(a + dt_bias)	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1136	# Step 1: Apply gating to hidden state: h *= exp(g)	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1139	# Step 2: Delta rule: v -= k^T @ h ([B,H,K] x [B,H,K,V] -> [B,H,V])	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1142	# Step 3: Apply beta gating: v *= beta	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1145	# Step 4: Update hidden state with outer product: h += k ⊗ v_new	COMMENT
LOW⚡	tests/gdn/reference_delta_rule.py	1148	# Step 5: Compute output: o = q^T @ h	COMMENT
LOW	tests/gdn/reference_delta_rule.py	1098	# Step 2: Compute beta = sigmoid(b)	COMMENT
LOW⚡	benchmarks/bench_cute_dsl_add_rmsnorm_fp4quant.py	149	# Step 1: Add	COMMENT
LOW⚡	benchmarks/bench_cute_dsl_add_rmsnorm_fp4quant.py	151	# Step 2: RMSNorm	COMMENT
LOW⚡	benchmarks/bench_cute_dsl_add_rmsnorm_fp4quant.py	153	# Step 3: FP4 quantize (with global_scale for NVFP4)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	179	# Step 1: Apply decay to hidden state: h *= exp(g)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	182	# Step 2: Delta rule: v -= sum(h * k, dim=0) = k @ h	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	187	# Step 3: Apply beta gating: v *= beta	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	190	# Step 4: Update hidden state: h += outer(k, v) = k[:, None] * v[None, :]	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	193	# Step 5: Compute output: o = q @ h = sum(q[:, None] * h, dim=0)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	361	# Step 1: Apply decay: h *= exp(g)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	364	# Step 2: Delta rule: v -= h @ k (h is [BV, BK], k is [BK])	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	369	# Step 3: Apply beta gating	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	372	# Step 4: Update state: h += outer(v, k) = v[:, None] * k[None, :]	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	376	# Step 5: Compute output: o = h @ q = sum(h * q[None, :], axis=1) -> [BV]	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	633	# Step 1: Apply decay to hidden state: h *= exp(g)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	636	# Step 2: Delta rule: v -= h @ k = sum(h * k[None, :], axis=1)	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	640	# Step 3: Apply beta gating: v *= beta	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	643	# Step 4: Update hidden state: h += outer(v, k) = v[:, None] * k[None, :]	COMMENT
LOW⚡	benchmarks/gdn_triton_reference.py	646	# Step 5: Compute output: o = h @ q = sum(h * q[None, :], axis=1)	COMMENT
LOW⚡	benchmarks/bench_cute_dsl_rmsnorm_fp4quant.py	157	# Step 1: RMSNorm	COMMENT
LOW⚡	benchmarks/bench_cute_dsl_rmsnorm_fp4quant.py	159	# Step 2: FP4 quantize (with global_scale for NVFP4)	COMMENT
LOW	benchmarks/routines/norm.py	1356	# Step 2: RMSNorm on residual	COMMENT
LOW⚡	scripts/pr_checks/check_api_docs.py	63	# Step 1: collect all @flashinfer_api-decorated function names via AST	COMMENT
LOW⚡	scripts/pr_checks/check_api_docs.py	91	# Step 2: collect documented symbols from docs/api/*.rst	COMMENT
LOW⚡	scripts/pr_checks/check_api_docs.py	162	# Step 3: compare and report	COMMENT
LOW⚡	…/deepseek_sparse_attention/indexer_top_k/block_scan.py	77	# Step 1: Warp-level prefix sum using shuffle	COMMENT
LOW⚡	…/deepseek_sparse_attention/indexer_top_k/block_scan.py	80	# Step 2: Store warp prefix sums	COMMENT
LOW⚡	…/deepseek_sparse_attention/indexer_top_k/block_scan.py	85	# Step 3: Prefix sum across warps	COMMENT
LOW⚡	…/deepseek_sparse_attention/indexer_top_k/block_scan.py	94	# Step 4: Add warp-level prefix	COMMENT
LOW⚡	…/deepseek_sparse_attention/indexer_top_k/block_scan.py	98	# Step 5: Get total sum if need_total_sum is True	COMMENT
LOW	flashinfer/trace/templates/rope.py	1059	# Step 3: append into paged cache (GQA/MHA) — materialize the quantized	COMMENT
LOW⚡	…e_cutedsl_megakernel/src/moe_hopper_fp8/mega_runner.py	196	# Step 1: deterministic FP8 input + weight generation	COMMENT
LOW⚡	…e_cutedsl_megakernel/src/moe_hopper_fp8/mega_runner.py	625	# Step 2: FP8 reference	COMMENT
LOW⚡	…e_cutedsl_megakernel/src/moe_hopper_fp8/mega_runner.py	696	# Step 5: FP8 validation	COMMENT
LOW⚡	…e_cutedsl_megakernel/src/moe_hopper_fp8/mega_runner.py	794	# Step 4: FP8 kernel launch	COMMENT
LOW	…le_cutedsl_megakernel/src/moe_nvfp4_swapab/epilogue.py	871	# Step 1: absmax over the weighted regs.	COMMENT
LOW	…l_megakernel/src/moe_nvfp4_swapab/epilogue_refactor.py	1407	# Step 0: load tmem	COMMENT
LOW	…l_megakernel/src/moe_nvfp4_swapab/epilogue_refactor.py	1454	# Step 1: perform swiglu on the first part, interleave with the second's 32x32 tmem transpose.	COMMENT
LOW	…l_megakernel/src/moe_nvfp4_swapab/epilogue_refactor.py	1482	# Step 2: Quant	COMMENT
124 more matches not shown…

Self-Referential Comments78 hits · 204 pts

Severity	File	Line	Snippet	Context
MEDIUM	tests/conftest.py	351	# Create a dict with module info for reporting	COMMENT
MEDIUM	tests/kda/test_recurrent_kda_decode_export.py	38	# This file is the deprecated facade's regression suite, so its own deprecation	COMMENT
MEDIUM	tests/attention/test_attention_sink.py	303	# Create a fragmented page allocation pattern	COMMENT
MEDIUM	tests/attention/test_trtllm_gen_attention_decode.py	2020	# Create a larger tensor with 2*D in the last dimension	COMMENT
MEDIUM⚡	tests/trace/test_fi_trace.py	1225	# Create a dummy instance — we don't call run(), only fi_trace()	COMMENT
MEDIUM	tests/moe_ep/nccl_ep/test_handle_mock.py	803	# Create the handle at 222 rows on a fleet sized for 128, so the	COMMENT
MEDIUM	tests/cpu/autotuner/test_autotuner_configs.py	802	# Create a cache file with a known config	COMMENT
MEDIUM	tests/cpu/autotuner/test_autotuner_configs.py	1000	# Create a cache file	COMMENT
MEDIUM	tests/mamba/utils.py	266	# Create a simple linear chain structure by default	COMMENT
MEDIUM	tests/comm/test_cake_moe_alltoall.py	334	# Create a random input tensor	COMMENT
MEDIUM	tests/comm/test_cake_moe_alltoall.py	726	# Create a random input tensor	COMMENT
MEDIUM	tests/comm/test_cake_moe_alltoall.py	1398	# Create a random input tensor	COMMENT
MEDIUM	tests/comm/test_nvlink_two_sided_alltoall.py	100	# Create a random input tensor	COMMENT
MEDIUM	tests/comm/test_nvlink_two_sided_alltoall.py	163	# Create a random input tensor	COMMENT
MEDIUM	tests/comm/test_mnnvl_memory.py	138	# Create a random input tensor	COMMENT
MEDIUM	tests/comm/test_cake_moe_alltoall_multi_gpu.py	347	# Create the all-to-all driver	COMMENT
MEDIUM	tests/comm/test_quantized_allreduce.py	213	# This file is single-node spawn-managed; use a file store to avoid	COMMENT
MEDIUM	tests/utils/test_fp4_quantize.py	342	# Create a test scale factors tensor with uint8 dtype	COMMENT
MEDIUM	tests/utils/test_fp4_quantize.py	392	# Create a reasonable test tensor	COMMENT
MEDIUM	tests/utils/test_logging.py	520	# Create a tensor with specific properties	COMMENT
MEDIUM	tests/utils/test_logging_replay.py	1131	# Create a mock dump with safetensors files	COMMENT
MEDIUM	tests/utils/test_norm.py	1457	# Create a JIT spec for norm module without ENABLE_FP8 flag	COMMENT
MEDIUM	tests/utils/test_load_cubin_compile_race_condition.py	40	# Define the target file - same for all processes	COMMENT
MEDIUM	tests/utils/test_load_cubin_compile_race_condition.py	75	# Create a temporary directory for FLASHINFER_CUBIN_DIR	COMMENT
MEDIUM	tests/cli/test_cli_cmds.py	542	# Create a temporary JIT directory with some dummy cache files	STRING
MEDIUM	tests/cli/test_cli_cmds.py	585	# Create a temporary cubin directory with some dummy cubin files	STRING
MEDIUM	tests/test_helpers/rope_reference.py	113	# The following code is from the vLLM's implementation of RoPE.	COMMENT
MEDIUM	tests/moe/test_dpsk_fused_moe_fp8.py	596	# Create a mock MoE implementation for skip_checks	COMMENT
MEDIUM	docs/tutorials/jax_tvm_ffi/flashinfer_jax_tvm_ffi.py	1	# This file is the canonical Sphinx-Gallery source for the tutorial.	COMMENT
MEDIUM	docs/tutorials/jax_tvm_ffi/gemma3_flashinfer_jax.py	1	# This file is the canonical Sphinx-Gallery source for the tutorial.	COMMENT
MEDIUM	examples/pytorch/wan/transformer_wan_flashinfer.py	4	# This file is adapted from the diffusers WanTransformer3DModel implementation:	COMMENT
MEDIUM	scripts/task_test_multi_gpu_comm_kernels.sh	19	# Define the specific test files for multi-GPU comm tests (single-node)	COMMENT
MEDIUM	scripts/task_test_multi_node_comm_kernels.sh	23	# Define the specific test files for multi-node comm tests	COMMENT
MEDIUM	scripts/test_utils.sh	3	# This file is meant to be sourced by test runner scripts	COMMENT
MEDIUM	scripts/test_utils.sh	1022	# Create a bash array with the node IDs	COMMENT
MEDIUM	scripts/test_utils.sh	1192	# Create a temporary directory for parallel job state	COMMENT
MEDIUM	scripts/test_utils.sh	1207	# Create a results file for each test	COMMENT
MEDIUM	scripts/pr_checks/check_skip_all.py	585	# Create a JUnit XML with skip reasons for the GDN file.	COMMENT
MEDIUM	flashinfer/experimental/b12x/_lib/utils.py	368	# Create a numpy array	STRING
MEDIUM	flashinfer/experimental/b12x/_lib/dense_gemm.py	29	# This file is ported from the CUTLASS dense block-scaled GEMM example	COMMENT
MEDIUM	flashinfer/experimental/deepseek_v41/hca_v41.py	2253	# This function is reached only by the dequant warp interval, so	COMMENT
MEDIUM	…e_cutedsl_megakernel/src/moe_nvfp4_swapab/moe_utils.py	221	# Create a new cute.Pointer with generic address space, preserving alignment	COMMENT
MEDIUM	…e_cutedsl_megakernel/src/moe_nvfp4_swapab/moe_utils.py	249	# Create a new cute.Pointer with gmem address space, preserving alignment	COMMENT
MEDIUM	…b_cutedsl_megakernel/src/moe_nvfp4_swapab/moe_utils.py	220	# Create a new cute.Pointer with generic address space, preserving alignment	COMMENT
MEDIUM	…b_cutedsl_megakernel/src/moe_nvfp4_swapab/moe_utils.py	248	# Create a new cute.Pointer with gmem address space, preserving alignment	COMMENT
MEDIUM	…dsl_megakernel/src/moe_sm120_mxfp8_swapab/moe_utils.py	260	# Create a new cute.Pointer with generic address space, preserving alignment	COMMENT
MEDIUM	…dsl_megakernel/src/moe_sm120_mxfp8_swapab/moe_utils.py	288	# Create a new cute.Pointer with gmem address space, preserving alignment	COMMENT
MEDIUM	…m100/cutedsl_megamoe/src/moe_nvfp4_swapab/moe_utils.py	219	# Create a new cute.Pointer with generic address space, preserving alignment	COMMENT
MEDIUM	…m100/cutedsl_megamoe/src/moe_nvfp4_swapab/moe_utils.py	247	# Create a new cute.Pointer with gmem address space, preserving alignment	COMMENT
MEDIUM	flashinfer/comm/dlpack_utils.py	224	# Create a new capsule each time	STRING
MEDIUM	flashinfer/fused_moe/core.py	575	# Create the output	COMMENT
MEDIUM	flashinfer/fused_moe/cute_dsl/rubin/custom_pipeline.py	44	# This file is copied and modified from https://github.com/NVIDIA/cutlass/tree/main/python/CuTeDSL/cutlass/pipeline	COMMENT
MEDIUM	flashinfer/fused_moe/cute_dsl/rubin/utils.py	146	# Create a numpy array	STRING
MEDIUM	…hinfer/fused_moe/cute_dsl/blackwell/custom_pipeline.py	44	# This file is copied and modified from https://github.com/NVIDIA/cutlass/tree/main/python/CuTeDSL/cutlass/pipeline	COMMENT
MEDIUM	flashinfer/fused_moe/cute_dsl/blackwell/utils.py	44	# This file is copied and modified from cutlass https://github.com/NVIDIA/cutlass/blob/main/python/CuTeDSL/cutlass/cute/	COMMENT
MEDIUM	…blockscaled_contiguous_grouped_gemm_finalize_fusion.py	317	"""This class implements batched matrix multiplication (C = A x SFA x B x SFB) with support for various data types	STRING
MEDIUM	…ockscaled_contiguous_gather_grouped_gemm_act_fusion.py	332	"""This class implements contiguous grouped matrix multiplication with gather operation and SwiGLU fusion	STRING
MEDIUM	flashinfer/fused_moe/cute_dsl/common/kernel_utils.py	44	# This file is copied and modified from cutlass https://github.com/NVIDIA/cutlass/blob/main/python/CuTeDSL/cutlass/cute/	COMMENT
MEDIUM	flashinfer/gemm/kernels/dense_blockscaled_gemm_sm103.py	29	# This file is ported from CUTLASS's sm103_dense_blockscaled_gemm_persistent.py	COMMENT
MEDIUM	flashinfer/gemm/kernels/dense_blockscaled_gemm_sm103.py	54	"""This class implements batched matrix multiplication (C = A x SFA x B x SFB) with support for FP4 data types	STRING
18 more matches not shown…

Redundant / Tautological Comments70 hits · 104 pts

Severity	File	Line	Snippet	Context
LOW	build_backend.py	1059	# Check if we're in a git repository	COMMENT
LOW	flashinfer-cubin/build_backend.py	75	# Check if we're in a git repository	COMMENT
LOW	tests/attention/test_attention_sink.py	934	# Check if any request has qo_len > kv_len for causal case	COMMENT
LOW	…s/comm/test_trtllm_allreduce_fusion_group_fp8_quant.py	218	# Check if current workspace is large enough; recreate if not	COMMENT
LOW	tests/utils/test_fp8_quantize.py	1833	# Check if the bits of paddings are zero.	COMMENT
LOW	tests/utils/test_load_cubin_compile_race_condition.py	44	# Read the file from FLASHINFER_CUBIN_DIR	COMMENT
LOW	tests/model_optimizations/test_dsv3_fused_routing.py	276	# Check if kernel's selection matches expected experts (exact or tied)	COMMENT
LOW	tests/test_helpers/sink_attention_reference.py	95	# Check if variable length mode is indicated by presence of indptr	COMMENT
LOW	flashinfer-jit-cache/build_backend.py	89	# Check if we're in a git repository	COMMENT
LOW	benchmarks/bench_tgv_gemm.py	184	# Check if results are close enough	COMMENT
LOW	benchmarks/bench_logging_overhead.py	224	# Check if CUDA is available	COMMENT
LOW	benchmarks/bench_dcp_alltoall.py	297	# Print results (rank 0 only)	COMMENT
LOW	scripts/xfails_tracker.py	52	# Check if this is a test function (starts with "test_" or has @pytest marks)	COMMENT
LOW	scripts/xfails_tracker.py	70	# Check if this is a pytest.param call	COMMENT
LOW	scripts/xfails_tracker.py	77	# Check if this is a runtime pytest.xfail() call	COMMENT
LOW	scripts/test_utils.sh	1671	# Check if parallel execution is enabled	COMMENT
LOW	scripts/codeowner_analyzer.py	89	# Check if gh CLI is available	COMMENT
LOW	scripts/codeowner_analyzer.py	114	# Check if it's already in our cache	COMMENT
LOW	scripts/codeowner_analyzer.py	220	# Check if username is in allowed list (case-insensitive)	COMMENT
LOW	scripts/codeowner_analyzer.py	290	# Check if it's a relevant code file	COMMENT
LOW	scripts/codeowner_analyzer.py	420	# Check if contributor is in allowed users list	COMMENT
LOW	scripts/codeowner_analyzer.py	629	# Write file-level overrides LAST (CODEOWNERS uses last-match-wins)	COMMENT
LOW	.github/workflows/update-xfails-report.yml	43	# Check if xfails report file is new (unstaged) or has changes	COMMENT
LOW	.github/workflows/update-codeowners.yml	50	# Check if CODEOWNERS file is new (unstaged) or has changes	COMMENT
LOW	.github/workflows/issue-claim.yml	86	# Check if issue already has assignees	COMMENT
LOW	.github/workflows/issue-claim.yml	301	# Check if user is now a collaborator (accepted invitation)	COMMENT
LOW	flashinfer/api_logging.py	2105	# Check if this parameter was provided	COMMENT
LOW	flashinfer/utils.py	391	# Check if it's a CUDA device	COMMENT
LOW	flashinfer/prefill.py	647	# Check if FP8 by presence of scale tensors	COMMENT
LOW	flashinfer/__main__.py	1106	# Check if this is a single dump or a session / sequence root	COMMENT
LOW	flashinfer/cuda_utils.py	20	# Check if cuda.cudart module is available and import accordingly	COMMENT
LOW	…e_attention/indexer_top_k/indexer_top_k_varlen_util.py	1470	# Check if the number of valid indices matches	COMMENT
LOW	…ention/prims_ts/kernels/fmha_context/fmha_resources.py	7103	# Check if we should skip correction (when old_max == new_max)	COMMENT
LOW	…utedsl_megamoe/src/moe_bf16_glu/mega_reference_bf16.py	564	# Check if input data types are compatible with MMA instruction	COMMENT
LOW	flashinfer/autotuner/autotuner.py	3083	# Set time_measured to inf to notify the failure of the tactic. This can happen when	COMMENT
LOW	flashinfer/triton/sm_constraint_gemm.py	66	# Set num_sms to be 100% of the available SMs	COMMENT
LOW	flashinfer/comm/mnnvl.py	884	# Check if cuda.cudart is available and import accordingly	COMMENT
LOW	flashinfer/comm/mnnvl.py	921	# Check if device supports multicasting	COMMENT
LOW	…hinfer/fused_moe/cute_dsl/blackwell/custom_pipeline.py	357	# Set mask to None if not using clusters (i.e. 1CTA kernels)	COMMENT
LOW	…hinfer/fused_moe/cute_dsl/blackwell/custom_pipeline.py	363	# Set mask to None if not using 2CTA instructions	COMMENT
LOW	flashinfer/gemm/gemm_base.py	10372	# Check if dimensions are aligned for cute-dsl (16-byte alignment)	COMMENT
LOW	flashinfer/gemm/kernels/dense_blockscaled_gemm_sm103.py	363	# Check if input data types are compatible with MMA instruction	COMMENT
LOW	flashinfer/gemm/kernels/dense_blockscaled_gemm_sm103.py	998	# Check if we're not at the last buffer of the last k_tile	COMMENT
LOW	flashinfer/gemm/kernels/dense_blockscaled_gemm_sm103.py	1083	# Check if we're not at the last stage of the last k_tile	COMMENT
LOW	flashinfer/gemm/kernels/epilogue_utils.py	241	# Check if tile_sched is StaticPersistentTileScheduler or any subclass inheriting from it	COMMENT
LOW	flashinfer/gemm/kernels/dense_blockscaled_gemm_sm107.py	754	# Check if prefetch is enabled (prefetch_dist > 0)	COMMENT
LOW	flashinfer/gemm/kernels/dense_blockscaled_gemm_sm107.py	2069	# Check if the current combination is valid	COMMENT
LOW	flashinfer/gemm/kernels/dense_blockscaled_gemm_sm107.py	2794	# Check if prefetch is enabled (prefetch_dist > 0)	COMMENT
LOW	flashinfer/gemm/kernels/blockwise_gemm_blackwell.py	437	# Check if input data types are compatible with MMA instruction	COMMENT
LOW	flashinfer/gemm/kernels/grouped_gemm_masked_rubin.py	694	# Check if input data types are compatible with MMA instruction	COMMENT
LOW	flashinfer/gemm/kernels/grouped_gemm_masked_rubin.py	1950	# Check if the current combination is valid	COMMENT
LOW	flashinfer/gemm/kernels/bmm_fp8_rubin.py	548	# Check if we should use Bkeep-Breuse pattern	COMMENT
LOW	flashinfer/gemm/kernels/bmm_fp8_blackwell.py	311	# Check if input data types are compatible with MMA instruction	COMMENT
LOW	flashinfer/gemm/kernels/blockwise_gemm_rubin.py	1434	# Check if we should use Bkeep-Breuse pattern	COMMENT
LOW	…ashinfer/gemm/kernels/grouped_gemm_masked_blackwell.py	907	# Check if input data types are compatible with MMA instruction	COMMENT
LOW	…er/gemm/kernels/dense_blockscaled_gemm_sm100_common.py	102	# Check if input data types are compatible with MMA instruction	COMMENT
LOW	…nfer/gemm/kernels/grouped_gemm_contiguous_blackwell.py	551	# Check if input data types are compatible with MMA instruction	COMMENT
LOW	flashinfer/gemm/kernels/cutile/ragged_bmm_cutile.py	780	# Check if autotune is enabled	COMMENT
LOW	…ashinfer/gemm/kernels/cutile/gemm_alpha_beta_cutile.py	411	# Check if autotune is requested	COMMENT
LOW	flashinfer/quantization/kernels/mxfp8_quantize.py	741	# Check if this thread's row is valid	COMMENT
10 more matches not shown…

Docstring Block Structure19 hits · 95 pts

Severity	File	Line	Snippet	Context
HIGH	flashinfer/green_ctx.py	133	Split the device into multiple `green contexts <https://docs.nvidia.com/cuda/cuda-driver-api/group__CUDA__GREEN__CO	STRING
HIGH	flashinfer/green_ctx.py	204	Split the device into multiple green contexts, each with a fixed number of SMs, and return the corresponding st	STRING
HIGH	flashinfer/green_ctx.py	348	Create one green context per locality domain (uGPU) of ``dev`` and return its stream and ``CUdevResource``, in	STRING
HIGH	flashinfer/utils.py	375	Get GPU memory bandwidth in GB/s for the specified CUDA device. Args: device: torch.device object, e.g	STRING
HIGH	…erimental/qwen_3_5_block/_cudnn_frost/QSA/selection.py	258	Select the ``top_k`` compressed blocks per query token (see the module docstring for the contract). Args: q	STRING
HIGH	flashinfer/autotuner/autotuner.py	113	Map x to the nearest bucket using floor or ceil semantics. Args: x: The value to map. buckets: Bu	STRING
HIGH	flashinfer/autotuner/autotuner.py	157	Create a mapper function for :class:`DynamicTensorSpec.map_to_tuning_buckets`. The returned callable maps any integ	STRING
HIGH	flashinfer/autotuner/autotuner.py	914	Context manager for autotuning with optional file-based caching. Controls how FlashInfer profiles and selects the b	STRING
HIGH	flashinfer/autotuner/autotuner.py	4512	Load autotuner configs from a JSON file. Populates the internal config lookup table so that ``search_cache()``	STRING
HIGH	…blockscaled_contiguous_grouped_gemm_finalize_fusion.py	380	Blockscaled contiguous grouped GEMM for MoE GEMM2 workloads. Fused mode applies routing weights and atomically redu	STRING
HIGH	…ockscaled_contiguous_gather_grouped_gemm_act_fusion.py	503	Blockscaled contiguous gather grouped GEMM with fused FC1 activation. Performs grouped matrix multiplication with f	STRING
HIGH	flashinfer/attn_scores/attn_scores.py	1610	FP8 paged MQA logits for Blackwell (SM100/SM103) and Rubin (SM107). Computes weighted per-head ReLU attention score	STRING
HIGH	flashinfer/attn_scores/attn_scores.py	2127	FP4 (MXFP4) paged MQA logits for Blackwell (SM100/SM103) and Rubin (SM107). Computes weighted per-head ReLU attenti	STRING
HIGH	flashinfer/gemm/gemm_bf16_fp4.py	177	Prepare FP4 weights for the bf16 x fp4 GEMM, for a specific backend. The caller is expected to start with weights i	STRING
HIGH	flashinfer/quantization/fp4_quantization.py	484	Quantize a batched tensor to FP4 (E2M1x2) with per-block scale factors. This function converts a float/bfloat16	STRING
HIGH	flashinfer/quantization/fp4_quantization.py	685	Quantize a silu and matmul with masked batched tensor to FP4 (E2M1x2) with per-block scale factors. This functi	STRING
HIGH	flashinfer/topk_varlen/kernels/filtered_topk_util.py	114	Return the filtered top-k resource policy for an architecture. Args: compute_capability: A ``(major, minor)	STRING
HIGH	flashinfer/parallel_attention/utils.py	356	Create a device mesh and return the Ring and Ulysses process groups. Builds a ``DeviceMesh`` with up to three dimen	STRING
HIGH	flashinfer/kda_kernels/kda_decode_wy_output_only.py	2070	Output-only (frozen-state) KDA decode over 1..16 tokens per sequence. Computes the KDA (Kimi Delta Attention) outpu	STRING

Modern AI Meta-Vocabulary28 hits · 85 pts

Severity	File	Line	Snippet	Context
MEDIUM	tests/experimental/test_cake_gdn_det_train_kcp.py	53	except KeyError: # the registry entry is added by the package scaffolding lane	CODE
MEDIUM	tests/autotuner/test_autotuner_core.py	2078	# TODO: Let AutoTuner orchestrate multiple same-shaped value samples within one tuning transaction.	COMMENT
MEDIUM	…hinfer/experimental/cake_gdn_det_train_kcp/__init__.py	24	# the byte queries; the partition / context plan, the host orchestration, the	COMMENT
MEDIUM	…hinfer/experimental/cake_kda_det_train_kcp/cake_kcp.py	503	# operand, state and output in FP64 so the orchestration can be checked against the global FP64 reference to ~1e-12.	COMMENT
MEDIUM	…l/qwen_3_5_block/_cudnn_frost/sdpa/bwd/config_sm100.py	894	# Rubin (SM107) usable dynamic SMEM: the 327 KiB cap minus the 2 KiB scaffold reserve	COMMENT
MEDIUM	…l/qwen_3_5_block/_cudnn_frost/sdpa/bwd/config_sm100.py	1227	f"bwd d512 2x2: SMEM {smem / 1024:.1f} KiB + {SMEM_SCAFFOLD_BYTES_2X2 // 1024} KiB scaffold over the {cfg.SM	CODE
MEDIUM	…l/qwen_3_5_block/_cudnn_frost/sdpa/bwd/config_sm107.py	184	# The honest bound is the device cap minus the declared scaffolding.	COMMENT
MEDIUM	…l/qwen_3_5_block/_cudnn_frost/sdpa/bwd/config_sm107.py	1669	# --- SMEM: the per-CTA cap, and the scaffold budget -----------------------------------------	COMMENT
MEDIUM⚡	…cutedsl_megakernel/src/moe_nvfp4_swapab/mega_runner.py	2414	# Top-level orchestration	COMMENT
MEDIUM	…cutedsl_megakernel/src/moe_nvfp4_swapab/mega_runner.py	2658	# orchestration above is the part being smoke-tested for now.	COMMENT
MEDIUM⚡	…cutedsl_megakernel/src/moe_nvfp4_swapab/mega_runner.py	2405	# Top-level orchestration	COMMENT
MEDIUM	…cutedsl_megakernel/src/moe_nvfp4_swapab/mega_runner.py	2648	# orchestration above is the part being smoke-tested for now.	COMMENT
MEDIUM⚡	…l_megakernel/src/moe_sm120_mxfp8_swapab/mega_runner.py	2151	# Top-level orchestration	COMMENT
MEDIUM	…l_megakernel/src/moe_sm120_mxfp8_swapab/mega_runner.py	2378	# orchestration above is the part being smoke-tested for now.	COMMENT
MEDIUM⚡	…00/cutedsl_megamoe/src/moe_nvfp4_swapab/mega_runner.py	2414	# Top-level orchestration	COMMENT
MEDIUM	…00/cutedsl_megamoe/src/moe_nvfp4_swapab/mega_runner.py	2658	# orchestration above is the part being smoke-tested for now.	COMMENT
MEDIUM	…infer/cake_causal_conv1d_det_train_backend/cake_jit.py	522	# Rendered by the exporter from the family adapter's scaffold: the backward	COMMENT
MEDIUM	…infer/cake_causal_conv1d_det_train_backend/cake_jit.py	609	# Rendered by the exporter from the family adapter's scaffold: the backward	COMMENT
MEDIUM	…infer/cake_causal_conv1d_det_train_backend/cake_jit.py	695	# Rendered by the exporter from the family adapter's scaffold: the backward	COMMENT
MEDIUM	…infer/cake_causal_conv1d_det_train_backend/cake_jit.py	1793	# Rendered by the exporter from the family adapter's scaffold: the backward	COMMENT
MEDIUM	…infer/cake_causal_conv1d_det_train_backend/cake_jit.py	1884	# Rendered by the exporter from the family adapter's scaffold: the backward	COMMENT
MEDIUM	…infer/cake_causal_conv1d_det_train_backend/cake_jit.py	1976	# Rendered by the exporter from the family adapter's scaffold: the backward	COMMENT
MEDIUM	…infer/cake_causal_conv1d_det_train_backend/cake_jit.py	2067	# Rendered by the exporter from the family adapter's scaffold: the backward	COMMENT
MEDIUM	flashinfer/fused_moe/da_runtime.py	770	# Strip orchestration-owned keywords before delegating dtype-specific allocation and	COMMENT
MEDIUM	flashinfer/mla/cake_dsv4.py	1568	# Launch orchestration #	COMMENT
MEDIUM⚡	flashinfer/cute_dsl/attention/roles/epilogue.py	74	# Prefill orchestration — proven-correct inline implementation	COMMENT
MEDIUM⚡	flashinfer/cute_dsl/attention/roles/loader_tma.py	282	# Prefill orchestration — proven-correct inline implementation	COMMENT
MEDIUM⚡	flashinfer/cute_dsl/attention/roles/mma.py	198	# Prefill orchestration — uses primitives for GEMMs and TMEM lifecycle	COMMENT

Example Usage Blocks20 hits · 30 pts

Severity	File	Line	Snippet	Context
LOW	csrc/fmha_v2/fmha/utils.h	66	// Example usage:	COMMENT
LOW	scripts/list_apis.sh	10	# Usage:	COMMENT
LOW	…_megakernel/src/moe_hopper_fp8/run_functional_tests.sh	13	# Usage:	COMMENT
LOW	…utedsl_megakernel/src/moe_hopper_fp8/run_mega_tests.sh	21	# Usage:	COMMENT
LOW	…cutedsl_megakernel/src/moe_hopper_fp8/run_perf_test.sh	18	# Usage:	COMMENT
LOW	…egakernel/src/moe_nvfp4_swapab/run_functional_tests.sh	10	# Usage:	COMMENT
LOW	…edsl_megakernel/src/moe_nvfp4_swapab/run_mega_tests.sh	17	# Usage:	COMMENT
LOW	…megakernel/src/moe_hopper_bf16/run_functional_tests.sh	13	# Usage:	COMMENT
LOW	…tedsl_megakernel/src/moe_hopper_bf16/run_mega_tests.sh	19	# Usage:	COMMENT
LOW	…utedsl_megakernel/src/moe_hopper_bf16/run_perf_test.sh	18	# Usage:	COMMENT
LOW	…egakernel/src/moe_nvfp4_swapab/run_functional_tests.sh	10	# Usage:	COMMENT
LOW	…edsl_megakernel/src/moe_nvfp4_swapab/run_mega_tests.sh	17	# Usage:	COMMENT
LOW	…l_megakernel/src/moe_mxfp8_glu/run_functional_tests.sh	18	# Usage:	COMMENT
LOW	…cutedsl_megakernel/src/moe_mxfp8_glu/run_mega_tests.sh	22	# Usage:	COMMENT
LOW	…l_megamoe/src/moe_nvfp4_swapab/run_functional_tests.sh	10	# Usage:	COMMENT
LOW	…cutedsl_megamoe/src/moe_nvfp4_swapab/run_mega_tests.sh	17	# Usage:	COMMENT
LOW	…edsl_megamoe/src/moe_mxfp8_glu/run_functional_tests.sh	18	# Usage:	COMMENT
LOW	…00/cutedsl_megamoe/src/moe_mxfp8_glu/run_mega_tests.sh	22	# Usage:	COMMENT
LOW	…tedsl_megamoe/src/moe_bf16_glu/run_functional_tests.sh	18	# Usage:	COMMENT
LOW	…100/cutedsl_megamoe/src/moe_bf16_glu/run_mega_tests.sh	33	# Usage:	COMMENT

AI Response Leakage3 hits · 22 pts

Severity	File	Line	Snippet	Context
HIGH	docs/tutorials/jax_tvm_ffi/flashinfer_jax_tvm_ffi.py	304	# In this example, you should learn how minimal the JAX-TVM bridge can be when the function signature and ``arg_spec`` a	COMMENT
HIGH	docs/tutorials/jax_tvm_ffi/flashinfer_jax_tvm_ffi.py	427	# In this example, you should learn how to handle two new complications that were absent in Example 1: a TVM function th	COMMENT
HIGH	docs/tutorials/jax_tvm_ffi/flashinfer_jax_tvm_ffi.py	644	# In this example, you should learn about three patterns that appear together for the first time: a kernel **compiled se	COMMENT

Overly Generic Function Names14 hits · 12 pts

Severity	File	Line	Snippet	Context
LOW	tests/cpu/utils/test_decorators.py	18	def my_function(x, y):	CODE
LOW	tests/utils/test_logging.py	186	def test_function(x, y):	CODE
LOW	tests/utils/test_logging.py	212	def test_function(tensor, value):	CODE
LOW	tests/utils/test_logging.py	250	def test_function(tensor):	CODE
LOW	tests/utils/test_logging.py	278	def test_function(mode: TestEnum, strategy: StringEnum):	CODE
LOW	tests/utils/test_logging.py	308	def test_function(x, y=10, z=20, mode=TestEnum.OPTION_A):	CODE
LOW	tests/utils/test_logging.py	339	def test_function(x, y=10, z=20):	CODE
LOW	tests/utils/test_logging.py	479	def test_function(	CODE
LOW	tests/utils/test_logging.py	517	def test_function(tensor):	CODE
LOW	tests/utils/test_logging.py	550	def test_function(nested):	CODE
LOW	tests/utils/test_logging.py	615	def test_function(x):	CODE
LOW	tests/utils/test_logging.py	640	def test_function(a, b, c):	CODE
LOW	tests/jit/test_cute_dsl_strict_key_closure.py	212	"from .impl import factor\n\ndef helper():\n return factor()\n",	CODE
LOW	flashinfer/api_logging.py	2451	... def my_function(x, y):	STRING

Synthetic Comment Markers1 hit · 5 pts

Severity	File	Line	Snippet	Context
HIGH	.github/workflows/update-codeowners.yml	76	Auto-generated CODEOWNERS update based on commit activity over the last ${{ env.DAYS_BACK }} days.	CODE

Dead Code2 hits · 4 pts

Severity	File	Line	Snippet	Context
MEDIUM	tests/experimental/b12x/preparation/test_session.py	62		CODE
MEDIUM	…infer/experimental/b12x/comm/pcie/pcie_vocab_argmax.py	300		CODE

Fake / Example Data4 hits · 3 pts

Severity	File	Line	Snippet	Context
LOW⚡	flashinfer/gdn_kernels/gdn_decode_mtp.py	3085	dummy_key = (q.device, h0_source.dtype)	CODE
LOW⚡	flashinfer/gdn_kernels/gdn_decode_mtp.py	3086	if dummy_key not in dummy_intermediate_states:	CODE
LOW⚡	flashinfer/gdn_kernels/gdn_decode_mtp.py	3087	dummy_intermediate_states[dummy_key] = torch.empty(	CODE
LOW⚡	flashinfer/gdn_kernels/gdn_decode_mtp.py	3090	intermediate_states = dummy_intermediate_states[dummy_key]	CODE

flashinfer-ai/flashinfer

Open-Source Code Quality, Security and Technical Debt Review