style: restore rustfmt cleanliness on main

[justinchuby]

What

cargo fmt --all -- --check currently fails on origin/main (bce03cabb) in five
places across crates/onnx-runtime-ep-cpu/src/kernels/gemm.rs and
crates/onnx-runtime-ep-cpu/src/kernels/matmul.rs. This is the cargo fmt --all output
and nothing else.

Why it happened

Nobody wrote badly-formatted code. #1073, #1079 and #1080 each touched these two files and
each was fmt-clean against its own base. Squash-merging them produced a combined text that
rustfmt formats differently:

• TRANSPOSE_TEST_LOCK.lock().unwrap_or_else(|e| e.into_inner()) is 61 characters, which
  exceeds rustfmt's default chain_width of 60 once it sits at test-body indentation (two
  sites).
• The onnx_runtime_ir import list grew past max_width and now wants braces on their own
  lines.
• One Gemm/transB attribute chain became short enough to fit on a single line after a
  neighbouring edit.
• One stray double blank line at end of a mod tests.

main is unprotected, so no required check re-ran fmt on the merge result and the breakage
landed silently.

How it was found

It blocked #1086: that PR's Fast (Linux x86_64) and Rust quality jobs failed on the
PR merge ref with diffs in files #1086 does not touch. Reproduced independently by
checking out origin/main into a clean worktree and running cargo fmt --all -- --check.

Verification

• cargo fmt --all -- --check -> clean (was: 5 diffs).
• Diff is whitespace/line-breaking only; git diff -w on the two files is empty apart from
  the import-brace move. No logic, no behaviour, no test changes.

Risk

None. Mechanical formatter output.

[codecov]

Codecov Report

❌ Patch coverage is 71.42857% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 79.80%. Comparing base (bce03ca) to head (904d15e).
⚠️ Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
crates/onnx-runtime-ep-cpu/src/kernels/gemm.rs	66.66%	0 Missing and 2 partials ⚠️

Additional details and impacted files

@@           Coverage Diff            @@
##             main    #1089    +/-   ##
========================================
  Coverage   79.79%   79.80%            
========================================
  Files         369      369            
  Lines      161558   161732   +174     
  Branches   161558   161732   +174     
========================================
+ Hits       128915   129068   +153     
- Misses      27898    27916    +18     
- Partials     4745     4748     +3     

Flag	Coverage Δ
cli-ort-linux	83.79% <ø> (ø)
cli-ort-windows	?
mlas	83.28% <ø> (+0.03%)	⬆️
offline	79.60% <71.42%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
crates/onnx-runtime-ep-cpu/src/kernels/matmul.rs	84.16% <100.00%> (-0.04%)	⬇️
crates/onnx-runtime-ep-cpu/src/kernels/gemm.rs	88.80% <66.66%> (+0.06%)	⬆️

... and 8 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

🔴 Benchmark Regression Detected

Comparison of criterion micro-benchmarks: PR head vs merge-base, measured on the same runner in the same job (base first → PR second).

ℹ️ Absolute times are informational only — they vary with runner load. The % change column is the reliable signal because both sides ran under identical conditions.

Status	Scenario	Base	PR	Change
🔴	matmul/small_generic_bf16_threads=8/1x256x256	29.91 µs	86.16 µs	+188.0%
🔴	matmul/medium_generic_f16_threads=8/32x512x512	33.79 µs	80.36 µs	+137.8%
🔴	matmul/small_generic_f16_threads=8/1x256x256	30.73 µs	72.27 µs	+135.2%
🔴	matmul/medium_generic_bf16_threads=8/32x512x512	434.75 µs	951.14 µs	+118.8%
🔴	matmul/small_generic_f32_threads=8/1x256x256	37.64 µs	75.23 µs	+99.9%
🔴	sampling_latency/top_k_per_token	51.88 µs	75.37 µs	+45.3%
🔴	gather/large_f32_threads=1-internal/131072	24.32 µs	34.66 µs	+42.5%
🔴	matmul/small_generic_bf16_threads=1/1x256x256	32.03 µs	41.81 µs	+30.5%
⚠️	matmul/medium_generic_f32_threads=1/32x512x512	2.32 ms	2.96 ms	+27.6%
⚠️	sampling_latency/greedy_per_token	3.21 µs	4.04 µs	+25.7%
⚠️	sampling_latency/top_p_per_token	383.21 µs	474.42 µs	+23.8%
⚠️	matmul/small_generic_f16_threads=1/1x256x256	31.46 µs	37.07 µs	+17.8%
⚠️	matmul/medium_generic_bf16_threads=1/32x512x512	566.13 µs	662.84 µs	+17.1%
⚠️	tokenization/decode_tokens_per_second	6.86 ms	7.92 ms	+15.4%
✅	sampling_latency/min_p_per_token	216.23 µs	248.13 µs	+14.8%
✅	qwen3_sampling_processors/top_k_full_sort_baseline	2.25 ms	2.55 ms	+13.4%
✅	add/medium_bf16_threads=1-internal/262144	101.12 µs	113.69 µs	+12.4%
✅	gather/medium_f32_threads=1-internal/32768	3.73 µs	4.15 µs	+11.3%
✅	grammar_masking/llguidance_compute_mask/32	81.65 µs	90.70 µs	+11.1%
✅	qwen3_sampling_processors/top_k_top_p_fast	688.67 µs	746.75 µs	+8.4%
✅	add/small_bf16_threads=1-internal/1024	433.0 ns	466.1 ns	+7.7%
✅	add/medium_f32_threads=1-internal/262144	24.93 µs	26.80 µs	+7.5%
✅	add/large_f32_threads=1-internal/4194304	668.14 µs	711.32 µs	+6.5%
✅	add/large_f16_threads=1-internal/4194304	1.67 ms	1.76 ms	+5.7%
✅	logit_processing/seven_processor_chain_per_step	347.89 µs	367.41 µs	+5.6%
✅	matmul/large_generic_bf16_threads=8/32x1024x1024	1.43 ms	1.51 ms	+5.3%
✅	matmul/large_generic_f32_threads=1/32x1024x1024	9.69 ms	10.15 ms	+4.8%
✅	add/medium_f16_threads=1-internal/262144	105.66 µs	110.44 µs	+4.5%
✅	add/small_f16_threads=1-internal/1024	460.5 ns	481.0 ns	+4.5%
✅	matmul/medium_generic_f16_threads=1/32x512x512	32.67 µs	34.13 µs	+4.5%
✅	kv_cache/alloc_dealloc_pages	41.57 µs	43.32 µs	+4.2%
✅	reduce_mean/large_f32_threads=1-internal/262144	980.30 µs	1.02 ms	+4.0%
✅	gather/small_f32_threads=1-internal/4096	678.2 ns	698.2 ns	+3.0%
✅	gather/small_f16_threads=1-internal/4096	476.1 ns	489.6 ns	+2.8%
✅	gather/medium_f16_threads=1-internal/32768	2.50 µs	2.57 µs	+2.8%
✅	gather/medium_bf16_threads=1-internal/32768	2.53 µs	2.60 µs	+2.7%
✅	reduce_mean/medium_f32_threads=1-internal/65536	254.05 µs	260.37 µs	+2.5%
✅	tokenization/encode_tokens_per_second	394.79 µs	404.45 µs	+2.4%
✅	qwen3_sampling_processors/top_p_full_sort_after_top_k_baseline	3.98 ms	4.07 ms	+2.3%
✅	matmul/medium_generic_f32_threads=8/32x512x512	1.07 ms	1.08 ms	+1.5%
✅	matmul/large_generic_bf16_threads=1/32x1024x1024	2.07 ms	2.08 ms	+0.4%
✅	qwen3_sampling_processors/top_p_fast_after_top_k	544.82 µs	544.44 µs	-0.1%
✅	add/small_f32_threads=1-internal/1024	203.7 ns	203.5 ns	-0.1%
✅	qwen3_sampling_processors/top_k_top_p_full_sort_baseline	6.21 ms	6.16 ms	-0.7%
✅	matmul/small_generic_f32_threads=1/1x256x256	39.94 µs	39.62 µs	-0.8%
✅	gather/small_bf16_threads=1-internal/4096	494.9 ns	488.0 ns	-1.4%
✅	add/large_bf16_threads=1-internal/4194304	1.71 ms	1.69 ms	-1.5%
✅	reduce_mean/small_f32_threads=1-internal/4096	16.13 µs	15.74 µs	-2.5%
✅	matmul/large_generic_f16_threads=1/32x1024x1024	85.95 µs	82.29 µs	-4.3%
✅	matmul/large_generic_f16_threads=8/32x1024x1024	92.05 µs	88.12 µs	-4.3%
✅	block_quantized_matmul_cached_dense/mxfp4_uncached_dequant_each_call/1x1024x1024	740.92 µs	698.89 µs	-5.7%
✅	matmul/large_generic_f32_threads=8/32x1024x1024	4.31 ms	3.98 ms	-7.7%
✅	qwen3_sampling_processors/top_k_partial_selection	186.91 µs	161.77 µs	-13.5%
✅	gather/large_bf16_threads=1-internal/131072	14.53 µs	12.47 µs	-14.2%
🟢	gather/large_f16_threads=1-internal/131072	16.08 µs	12.56 µs	-21.9%
🟢	block_quantized_moe_cached_dense/mxfp4_cached_dense_expert_repeated_call/rows=1,H=256,I=256,E=4,top_k=1	228.69 µs	167.24 µs	-26.9%
🟢	block_quantized_moe_cached_dense/mxfp4_uncached_expert_dequant_each_call/rows=1,H=256,I=256,E=4,top_k=1	823.65 µs	492.13 µs	-40.3%
🟢	block_quantized_matmul_cached_dense/mxfp4_preexpanded_dense_oncelock_like_proxy/1x1024x1024	120.34 µs	61.42 µs	-49.0%
🟢	block_quantized_matmul_cached_dense/mxfp4_cached_dense_repeated_call/1x1024x1024	146.92 µs	59.07 µs	-59.8%

Visual flags: ⚠️ ≥ 15% slower, 🔴 ≥ 30% slower — calibrated against measured runner noise (~27% worst-case on multi-threaded matmul)

Host info

CPU: Apple M1 (Virtual)
Cores: 3
OS: Darwin 25.5.0 arm64
Rust: rustc 1.97.1 (8bab26f4f 2026-07-14)
Load avg: { 2.53 3.40 4.84 }

What this cannot catch

• Regressions in code paths not covered by these benchmarks (e.g., end-to-end decode with a real model)
• Sub-threshold regressions that compound over multiple PRs
• Performance changes that only manifest under GPU execution
• Latency changes in the ORT integration path (these benchmarks exercise the native Rust kernels)