|
| 1 | +# 2025-07-15 Triage Log |
| 2 | + |
| 3 | +A busy week with a lot of rollups containing perf. regressions and mixed results. Overall, regressions won slightly, but there were also a few impressive wins on a few primary and secondary benchmarks. A lot of perf. effects are caused by the current rework of attribute parsing, which will hopefully result in slightly improved performance once it's finished. |
| 4 | + |
| 5 | +Triage done by **@kobzol**. |
| 6 | +Revision range: [0d11be5a..a9fb6103](https://perf.rust-lang.org/?start=0d11be5aabe0cd49609fff5fce57c4691a22fe55&end=a9fb6103b05c6ad6eee6bed4c0bb5a2e8e1024c6&absolute=false&stat=instructions%3Au) |
| 7 | + |
| 8 | +**Summary**: |
| 9 | + |
| 10 | +| (instructions:u) | mean | range | count | |
| 11 | +|:----------------------------------:|:-----:|:---------------:|:-----:| |
| 12 | +| Regressions ❌ <br /> (primary) | 0.5% | [0.1%, 1.5%] | 62 | |
| 13 | +| Regressions ❌ <br /> (secondary) | 0.5% | [0.1%, 1.8%] | 78 | |
| 14 | +| Improvements ✅ <br /> (primary) | -0.4% | [-3.9%, -0.1%] | 40 | |
| 15 | +| Improvements ✅ <br /> (secondary) | -1.4% | [-11.6%, -0.0%] | 74 | |
| 16 | +| All ❌✅ (primary) | 0.1% | [-3.9%, 1.5%] | 102 | |
| 17 | + |
| 18 | + |
| 19 | +5 Regressions, 4 Improvements, 8 Mixed; 5 of them in rollups |
| 20 | +47 artifact comparisons made in total |
| 21 | + |
| 22 | +#### Regressions |
| 23 | + |
| 24 | +Rollup of 7 pull requests [#143645](https://github.com/rust-lang/rust/pull/143645) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=f838cbc06de60819faff3413f374706b74824ca2&end=ab68b0fb26485ab1fa6977b2d8b59cc8a171c4aa&stat=instructions:u) |
| 25 | + |
| 26 | +| (instructions:u) | mean | range | count | |
| 27 | +|:----------------------------------:|:----:|:------------:|:-----:| |
| 28 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 29 | +| Regressions ❌ <br /> (secondary) | 0.3% | [0.1%, 0.5%] | 16 | |
| 30 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 31 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 32 | +| All ❌✅ (primary) | - | - | 0 | |
| 33 | + |
| 34 | +- Most regressions come from [#143402](https://github.com/rust-lang/rust/pull/143402), which refactors attribute parsing to the new system. It does more validations now, so slight regressions are expected. |
| 35 | +- Marked as triaged. |
| 36 | + |
| 37 | +Instantiate auto trait/`Copy`/`Clone`/`Sized` before computing constituent types binder [#143538](https://github.com/rust-lang/rust/pull/143538) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=e3fccdd4a16bf3aa223749efef1fa981589e43ae&end=e43d139a82620a268d3828a73e12a8679339e8f8&stat=instructions:u) |
| 38 | + |
| 39 | +| (instructions:u) | mean | range | count | |
| 40 | +|:----------------------------------:|:----:|:------------:|:-----:| |
| 41 | +| Regressions ❌ <br /> (primary) | 0.2% | [0.2%, 0.3%] | 3 | |
| 42 | +| Regressions ❌ <br /> (secondary) | 0.0% | [0.0%, 0.0%] | 1 | |
| 43 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 44 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 45 | +| All ❌✅ (primary) | 0.2% | [0.2%, 0.3%] | 3 | |
| 46 | + |
| 47 | +- The tiny regression was deemed to be [acceptable](https://github.com/rust-lang/rust/pull/143538#issuecomment-3047103921). |
| 48 | +- Marked as triaged. |
| 49 | + |
| 50 | +Rollup of 12 pull requests [#143731](https://github.com/rust-lang/rust/pull/143731) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=119574f83576dc1f3ae067f9a97986d4e2b0826c&end=78a6e132984dba0303ebad7dcfd1305c93ad5835&stat=instructions:u) |
| 51 | + |
| 52 | +| (instructions:u) | mean | range | count | |
| 53 | +|:----------------------------------:|:----:|:------------:|:-----:| |
| 54 | +| Regressions ❌ <br /> (primary) | 0.2% | [0.1%, 0.4%] | 12 | |
| 55 | +| Regressions ❌ <br /> (secondary) | 0.3% | [0.1%, 0.9%] | 14 | |
| 56 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 57 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 58 | +| All ❌✅ (primary) | 0.2% | [0.1%, 0.4%] | 12 | |
| 59 | + |
| 60 | +- The regression was caused by [#136906](https://github.com/rust-lang/rust/pull/136906), continuing discussion on that PR. |
| 61 | +- Marked as triaged. |
| 62 | + |
| 63 | +Rollup of 11 pull requests [#143888](https://github.com/rust-lang/rust/pull/143888) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=7e310f4b9a3f166d833ed09cf1d1ff96ab84b72c&end=56835d7ac14da9f966e1ff39fd9ffd2e29b764d1&stat=instructions:u) |
| 64 | + |
| 65 | +| (instructions:u) | mean | range | count | |
| 66 | +|:----------------------------------:|:----:|:------------:|:-----:| |
| 67 | +| Regressions ❌ <br /> (primary) | 0.4% | [0.1%, 0.8%] | 15 | |
| 68 | +| Regressions ❌ <br /> (secondary) | 0.8% | [0.8%, 0.8%] | 1 | |
| 69 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 70 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 71 | +| All ❌✅ (primary) | 0.4% | [0.1%, 0.8%] | 15 | |
| 72 | + |
| 73 | +- Regression caused by [#143519](https://github.com/rust-lang/rust/pull/143519). |
| 74 | +- The regession was expected due to rustc doing more work. Created an [issue](https://github.com/rust-lang/rust/issues/143945) for tracking a potential improvement. |
| 75 | +- Marked as triaged. |
| 76 | + |
| 77 | +Port `#[cfg]` to the new attribute parsing infrastructure [#143460](https://github.com/rust-lang/rust/pull/143460) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=7f2065a4bae1faed5bab928c670964eafbf43b55&end=a9fb6103b05c6ad6eee6bed4c0bb5a2e8e1024c6&stat=instructions:u) |
| 78 | + |
| 79 | +| (instructions:u) | mean | range | count | |
| 80 | +|:----------------------------------:|:----:|:------------:|:-----:| |
| 81 | +| Regressions ❌ <br /> (primary) | 0.4% | [0.2%, 1.3%] | 6 | |
| 82 | +| Regressions ❌ <br /> (secondary) | 0.8% | [0.5%, 1.3%] | 3 | |
| 83 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 84 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 85 | +| All ❌✅ (primary) | 0.4% | [0.2%, 1.3%] | 6 | |
| 86 | + |
| 87 | +- The `syn` regressions are noise. |
| 88 | +- The rest of the regressions are real, but tiny, and we expect to improve them soon, with the rest of the attribute parsing rework (https://github.com/rust-lang/rust/pull/143963). |
| 89 | +- Marked as triaged. |
| 90 | + |
| 91 | +#### Improvements |
| 92 | + |
| 93 | +Use `join_with_double_colon` in `write_shared.rs`. [#142869](https://github.com/rust-lang/rust/pull/142869) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=2783fc43fd13a669f314742890acd64200ad0bbf&end=45b80ac21a454d343833aad763ef604510c88375&stat=instructions:u) |
| 94 | + |
| 95 | +| (instructions:u) | mean | range | count | |
| 96 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 97 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 98 | +| Regressions ❌ <br /> (secondary) | - | - | 0 | |
| 99 | +| Improvements ✅ <br /> (primary) | -2.3% | [-4.0%, -0.6%] | 2 | |
| 100 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 101 | +| All ❌✅ (primary) | -2.3% | [-4.0%, -0.6%] | 2 | |
| 102 | + |
| 103 | + |
| 104 | +`rustc_pattern_analysis`: always check that deref patterns don't match on the same place as normal constructors [#143472](https://github.com/rust-lang/rust/pull/143472) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=558d25371fe1cc3d907ebcfc4e12d3c27acbe2b7&end=6b3ae3f6e45a33c2d95fa0362c9b2593e567fd34&stat=instructions:u) |
| 105 | + |
| 106 | +| (instructions:u) | mean | range | count | |
| 107 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 108 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 109 | +| Regressions ❌ <br /> (secondary) | 0.1% | [0.1%, 0.1%] | 1 | |
| 110 | +| Improvements ✅ <br /> (primary) | -0.2% | [-0.2%, -0.2%] | 2 | |
| 111 | +| Improvements ✅ <br /> (secondary) | -2.7% | [-2.9%, -2.6%] | 6 | |
| 112 | +| All ❌✅ (primary) | -0.2% | [-0.2%, -0.2%] | 2 | |
| 113 | + |
| 114 | + |
| 115 | +Add opaque TypeId handles for CTFE [#143696](https://github.com/rust-lang/rust/pull/143696) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=32cd9114712a24010b0583624dc52ac302194128&end=cf3fb768db439825e3c8d327f6d9f46e02965668&stat=instructions:u) |
| 116 | + |
| 117 | +| (instructions:u) | mean | range | count | |
| 118 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 119 | +| Regressions ❌ <br /> (primary) | 0.4% | [0.4%, 0.4%] | 1 | |
| 120 | +| Regressions ❌ <br /> (secondary) | 0.2% | [0.1%, 0.2%] | 6 | |
| 121 | +| Improvements ✅ <br /> (primary) | -0.2% | [-0.3%, -0.1%] | 20 | |
| 122 | +| Improvements ✅ <br /> (secondary) | -0.3% | [-0.6%, -0.1%] | 37 | |
| 123 | +| All ❌✅ (primary) | -0.1% | [-0.3%, 0.4%] | 21 | |
| 124 | + |
| 125 | + |
| 126 | +Propagate from borrowed locals in CopyProp [#143624](https://github.com/rust-lang/rust/pull/143624) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=915e5352448afb3c24f89117468935283bc7d2cf&end=bfc046a4b8d6b57db02540182466e989a9b0fb40&stat=instructions:u) |
| 127 | + |
| 128 | +| (instructions:u) | mean | range | count | |
| 129 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 130 | +| Regressions ❌ <br /> (primary) | 0.5% | [0.5%, 0.5%] | 1 | |
| 131 | +| Regressions ❌ <br /> (secondary) | 0.0% | [0.0%, 0.0%] | 1 | |
| 132 | +| Improvements ✅ <br /> (primary) | -0.2% | [-0.6%, -0.1%] | 27 | |
| 133 | +| Improvements ✅ <br /> (secondary) | -0.3% | [-1.0%, -0.0%] | 33 | |
| 134 | +| All ❌✅ (primary) | -0.1% | [-0.6%, 0.5%] | 28 | |
| 135 | + |
| 136 | + |
| 137 | +#### Mixed |
| 138 | + |
| 139 | +Add `ty_span` query [#143433](https://github.com/rust-lang/rust/pull/143433) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=a2d45f73c70d9dec57140c9412f83586eda895f8&end=688ea65df6a47866d0f72a00f1e18b47a7edf83b&stat=instructions:u) |
| 140 | + |
| 141 | +| (instructions:u) | mean | range | count | |
| 142 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 143 | +| Regressions ❌ <br /> (primary) | 0.6% | [0.2%, 0.9%] | 11 | |
| 144 | +| Regressions ❌ <br /> (secondary) | 0.1% | [0.0%, 0.2%] | 2 | |
| 145 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 146 | +| Improvements ✅ <br /> (secondary) | -0.2% | [-0.3%, -0.1%] | 2 | |
| 147 | +| All ❌✅ (primary) | 0.6% | [0.2%, 0.9%] | 11 | |
| 148 | + |
| 149 | +- Small expected regression in libc, as rustc now stores more information per item in the dep graph. |
| 150 | +- Marked as triaged. |
| 151 | + |
| 152 | +Use lld by default on `x86_64-unknown-linux-gnu` stable [#140525](https://github.com/rust-lang/rust/pull/140525) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=ab68b0fb26485ab1fa6977b2d8b59cc8a171c4aa&end=34097a38afc9efdedf776d3f1c84a190ff334886&stat=instructions:u) |
| 153 | + |
| 154 | +| (instructions:u) | mean | range | count | |
| 155 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 156 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 157 | +| Regressions ❌ <br /> (secondary) | 0.1% | [0.1%, 0.1%] | 1 | |
| 158 | +| Improvements ✅ <br /> (primary) | -0.3% | [-0.3%, -0.2%] | 2 | |
| 159 | +| Improvements ✅ <br /> (secondary) | -0.2% | [-0.3%, -0.1%] | 2 | |
| 160 | +| All ❌✅ (primary) | -0.3% | [-0.3%, -0.2%] | 2 | |
| 161 | + |
| 162 | +- This should be noise, as there should be no observable perf. changes in this PR. |
| 163 | +- Marked as triaged. |
| 164 | + |
| 165 | +Apply effects to `otherwise` edge in dataflow analysis [#142707](https://github.com/rust-lang/rust/pull/142707) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=34097a38afc9efdedf776d3f1c84a190ff334886&end=d350797b7e1a5b6952f5035cbad2a21613b32f6c&stat=instructions:u) |
| 166 | + |
| 167 | +| (instructions:u) | mean | range | count | |
| 168 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 169 | +| Regressions ❌ <br /> (primary) | 0.5% | [0.4%, 0.7%] | 2 | |
| 170 | +| Regressions ❌ <br /> (secondary) | 0.6% | [0.1%, 1.0%] | 15 | |
| 171 | +| Improvements ✅ <br /> (primary) | -0.4% | [-1.3%, -0.1%] | 18 | |
| 172 | +| Improvements ✅ <br /> (secondary) | -0.3% | [-0.5%, -0.0%] | 10 | |
| 173 | +| All ❌✅ (primary) | -0.3% | [-1.3%, 0.7%] | 20 | |
| 174 | + |
| 175 | +- More improvements than regressions, and improvements are mostly in primary benchmarks, while regressions are in secondary benchmarks. [This comment](https://github.com/rust-lang/rust/pull/142707#issuecomment-3050525801) mentions a plan for possible future improvements. |
| 176 | +- Marked as triaged. |
| 177 | + |
| 178 | +Rollup of 9 pull requests [#143667](https://github.com/rust-lang/rust/pull/143667) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=d350797b7e1a5b6952f5035cbad2a21613b32f6c&end=558d25371fe1cc3d907ebcfc4e12d3c27acbe2b7&stat=instructions:u) |
| 179 | + |
| 180 | +| (instructions:u) | mean | range | count | |
| 181 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 182 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 183 | +| Regressions ❌ <br /> (secondary) | 0.1% | [0.0%, 0.3%] | 20 | |
| 184 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 185 | +| Improvements ✅ <br /> (secondary) | -0.8% | [-1.1%, -0.0%] | 8 | |
| 186 | +| All ❌✅ (primary) | - | - | 0 | |
| 187 | + |
| 188 | +- The small regressions are caused by [#143520](https://github.com/rust-lang/rust/pull/143520). |
| 189 | +- The regressions are only on secondary benchmarks, and this produces a larger win for CTFE benchmarks. |
| 190 | +- Marked as triaged. |
| 191 | + |
| 192 | +Rollup of 9 pull requests [#143810](https://github.com/rust-lang/rust/pull/143810) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=2f9c9cede68be26774ea44efc79d0391f1c58af2&end=915e5352448afb3c24f89117468935283bc7d2cf&stat=instructions:u) |
| 193 | + |
| 194 | +| (instructions:u) | mean | range | count | |
| 195 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 196 | +| Regressions ❌ <br /> (primary) | 0.2% | [0.1%, 0.3%] | 34 | |
| 197 | +| Regressions ❌ <br /> (secondary) | 0.4% | [0.0%, 1.0%] | 33 | |
| 198 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 199 | +| Improvements ✅ <br /> (secondary) | -0.1% | [-0.3%, -0.0%] | 3 | |
| 200 | +| All ❌✅ (primary) | 0.2% | [0.1%, 0.3%] | 34 | |
| 201 | + |
| 202 | +- The regression comes from [#143403](https://github.com/rust-lang/rust/pull/143403). |
| 203 | +- Continuing discussion on the PR. |
| 204 | +- Marked as triaged. |
| 205 | + |
| 206 | +compute all rpitit of a trait [#143783](https://github.com/rust-lang/rust/pull/143783) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=bfc046a4b8d6b57db02540182466e989a9b0fb40&end=288e94c4ba406d612a556520442683d0f1ef5dbb&stat=instructions:u) |
| 207 | + |
| 208 | +| (instructions:u) | mean | range | count | |
| 209 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 210 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 211 | +| Regressions ❌ <br /> (secondary) | 1.1% | [1.1%, 1.1%] | 1 | |
| 212 | +| Improvements ✅ <br /> (primary) | -0.2% | [-0.3%, -0.2%] | 9 | |
| 213 | +| Improvements ✅ <br /> (secondary) | -0.3% | [-0.3%, -0.3%] | 4 | |
| 214 | +| All ❌✅ (primary) | -0.2% | [-0.3%, -0.2%] | 9 | |
| 215 | + |
| 216 | +- Improvements outweigh regressions. |
| 217 | +- Marked as triaged. |
| 218 | + |
| 219 | +de-duplicate condition scoping logic between AST→HIR lowering and `ScopeTree` construction [#143213](https://github.com/rust-lang/rust/pull/143213) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=b1d2f2c64c6254b2a935d6a2d29b68f4511cf500&end=d2baa49a106fad06fbf6202fb6ea8a0b3d2767cc&stat=instructions:u) |
| 220 | + |
| 221 | +| (instructions:u) | mean | range | count | |
| 222 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 223 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 224 | +| Regressions ❌ <br /> (secondary) | 0.2% | [0.1%, 0.3%] | 5 | |
| 225 | +| Improvements ✅ <br /> (primary) | -0.3% | [-0.3%, -0.2%] | 3 | |
| 226 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 227 | +| All ❌✅ (primary) | -0.3% | [-0.3%, -0.2%] | 3 | |
| 228 | + |
| 229 | +- Kind of a wash, a few small wins on a primary benchmark and a few small losses on secondary benchmarks. |
| 230 | +- Marked as triaged. |
| 231 | + |
| 232 | +Retire hir::*ItemRef. [#143357](https://github.com/rust-lang/rust/pull/143357) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=e9182f195b8505c87c4bd055b9f6e114ccda0981&end=9c3064e131f4939cc95a29bb11413c49bbda1491&stat=instructions:u) |
| 233 | + |
| 234 | +| (instructions:u) | mean | range | count | |
| 235 | +|:----------------------------------:|:-----:|:---------------:|:-----:| |
| 236 | +| Regressions ❌ <br /> (primary) | 0.2% | [0.2%, 0.2%] | 1 | |
| 237 | +| Regressions ❌ <br /> (secondary) | 0.3% | [0.2%, 0.3%] | 6 | |
| 238 | +| Improvements ✅ <br /> (primary) | -0.1% | [-0.1%, -0.1%] | 1 | |
| 239 | +| Improvements ✅ <br /> (secondary) | -4.8% | [-11.6%, -0.2%] | 14 | |
| 240 | +| All ❌✅ (primary) | 0.0% | [-0.1%, 0.2%] | 2 | |
| 241 | + |
| 242 | +- More improvements than regressions, in particular an impressive win on `many-assoc-items`. |
| 243 | +- Marked as triaged. |
0 commit comments