|
| 1 | +# 2023-02-21 Triage Log |
| 2 | + |
| 3 | +Overall a fairly positive week, with few noise-related regressions or |
| 4 | +improvements and many benchmarks showing significant improvements. The one |
| 5 | +large regression is limited to documentation builds and has at least a partial |
| 6 | +fix already planned. |
| 7 | + |
| 8 | +Other wins this week include an average [improvement][memopt] of around 1% in |
| 9 | +maximum memory usage of optimized builds, and a 2% average [reduction][sizeopt] |
| 10 | +in compiled binary sizes. These are fairly significant wins for these metrics. |
| 11 | + |
| 12 | +[memopt]: https://perf.rust-lang.org/?start=9bb6e60d1f1360234aae90c97964c0fa5524f141&end=3fee48c161a48b0c142d3998fff56faee96bd56c&absolute=false&stat=max-rss&kind=percentfromfirst |
| 13 | +[sizeopt]: https://perf.rust-lang.org/?start=9bb6e60d1f1360234aae90c97964c0fa5524f141&end=3fee48c161a48b0c142d3998fff56faee96bd56c&absolute=false&stat=size%3Alinked_artifact&kind=percentfromfirst |
| 14 | + |
| 15 | +Triage done by **@simulacrum**. |
| 16 | +Revision range: [9bb6e60..3fee48c1](https://perf.rust-lang.org/?start=9bb6e60d1f1360234aae90c97964c0fa5524f141&end=3fee48c161a48b0c142d3998fff56faee96bd56c&absolute=false&stat=instructions%3Au) |
| 17 | + |
| 18 | +**Summary**: |
| 19 | + |
| 20 | +| (instructions:u) | mean | range | count | |
| 21 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 22 | +| Regressions ❌ <br /> (primary) | 9.0% | [0.3%, 79.8%] | 10 | |
| 23 | +| Regressions ❌ <br /> (secondary) | 0.9% | [0.4%, 2.1%] | 14 | |
| 24 | +| Improvements ✅ <br /> (primary) | -1.7% | [-5.0%, -0.2%] | 128 | |
| 25 | +| Improvements ✅ <br /> (secondary) | -1.6% | [-5.2%, -0.2%] | 61 | |
| 26 | +| All ❌✅ (primary) | -0.9% | [-5.0%, 79.8%] | 138 | |
| 27 | + |
| 28 | + |
| 29 | +3 Regressions, 3 Improvements, 3 Mixed; 2 of them in rollups |
| 30 | +45 artifact comparisons made in total |
| 31 | + |
| 32 | +#### Regressions |
| 33 | + |
| 34 | +Rollup of 10 pull requests [#108052](https://github.com/rust-lang/rust/pull/108052) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=9bb6e60d1f1360234aae90c97964c0fa5524f141&end=6e0115778b0aedc90b59e035476c38e1b8c5c29b&stat=instructions:u) |
| 35 | + |
| 36 | +| (instructions:u) | mean | range | count | |
| 37 | +|:----------------------------------:|:----:|:-------------:|:-----:| |
| 38 | +| Regressions ❌ <br /> (primary) | 9.5% | [0.3%, 80.7%] | 10 | |
| 39 | +| Regressions ❌ <br /> (secondary) | 2.7% | [2.7%, 2.7%] | 1 | |
| 40 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 41 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 42 | +| All ❌✅ (primary) | 9.5% | [0.3%, 80.7%] | 10 | |
| 43 | + |
| 44 | +This regression is limited to doc builds and is likely caused by the tooltip |
| 45 | +expansion in "rustdoc: add more tooltips to intra-doc links |
| 46 | +[#108025](https://github.com/rust-lang/rust/pull/108025)". An initial reduction |
| 47 | +in overhead is proposed in |
| 48 | +[#108098](https://github.com/rust-lang/rust/pull/108098). |
| 49 | + |
| 50 | +use semantic equality for const param type equality assertion [#107940](https://github.com/rust-lang/rust/pull/107940) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=0416b1a6f6d5c42696494e1a3a33580fd3f669d8&end=068161ea483b1a80a959476cb3e31e6619a72737&stat=instructions:u) |
| 51 | + |
| 52 | +| (instructions:u) | mean | range | count | |
| 53 | +|:----------------------------------:|:----:|:------------:|:-----:| |
| 54 | +| Regressions ❌ <br /> (primary) | 0.3% | [0.2%, 0.5%] | 5 | |
| 55 | +| Regressions ❌ <br /> (secondary) | - | - | 0 | |
| 56 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 57 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 58 | +| All ❌✅ (primary) | 0.3% | [0.2%, 0.5%] | 5 | |
| 59 | + |
| 60 | +Relatively small significance threshold in a single benchmark -- likely a real |
| 61 | +regression, but not worth intense scrutiny. |
| 62 | + |
| 63 | +give the resolver access to TyCtxt [#105462](https://github.com/rust-lang/rust/pull/105462) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=8f55d6025fec5fb648948fbe4fb6dcb4184c67bc&end=2deff71719a32af76eb4493ddb1dc4653fb88cad&stat=instructions:u) |
| 64 | + |
| 65 | +| (instructions:u) | mean | range | count | |
| 66 | +|:----------------------------------:|:----:|:------------:|:-----:| |
| 67 | +| Regressions ❌ <br /> (primary) | 0.4% | [0.2%, 1.4%] | 94 | |
| 68 | +| Regressions ❌ <br /> (secondary) | 0.6% | [0.2%, 2.6%] | 25 | |
| 69 | +| Improvements ✅ <br /> (primary) | - | - | 0 | |
| 70 | +| Improvements ✅ <br /> (secondary) | - | - | 0 | |
| 71 | +| All ❌✅ (primary) | 0.4% | [0.2%, 1.4%] | 94 | |
| 72 | + |
| 73 | +Regressions are being pursued in follow-up PRs, this architectural change |
| 74 | +benefits incremental compilation and in general is on the right path even if in |
| 75 | +the short term we are seeing some regressions due to increased locking. |
| 76 | + |
| 77 | +#### Improvements |
| 78 | + |
| 79 | +Avoid accessing HIR when it can be avoided [#108006](https://github.com/rust-lang/rust/pull/108006) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=999ac5f7770bff68bd65f490990d32c3ec1faaa6&end=2d14db321b043ffc579a7461464c88d7e3f54f83&stat=instructions:u) |
| 80 | + |
| 81 | +| (instructions:u) | mean | range | count | |
| 82 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 83 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 84 | +| Regressions ❌ <br /> (secondary) | - | - | 0 | |
| 85 | +| Improvements ✅ <br /> (primary) | -0.4% | [-0.4%, -0.4%] | 3 | |
| 86 | +| Improvements ✅ <br /> (secondary) | -0.2% | [-0.2%, -0.2%] | 1 | |
| 87 | +| All ❌✅ (primary) | -0.4% | [-0.4%, -0.4%] | 3 | |
| 88 | + |
| 89 | + |
| 90 | +Rollup of 6 pull requests [#108211](https://github.com/rust-lang/rust/pull/108211) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=3701bdc6333145410f009c83bd03f424eca05009&end=3eb5c4581a386b13c414e8c8bd73846ef37236d1&stat=instructions:u) |
| 91 | + |
| 92 | +| (instructions:u) | mean | range | count | |
| 93 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 94 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 95 | +| Regressions ❌ <br /> (secondary) | - | - | 0 | |
| 96 | +| Improvements ✅ <br /> (primary) | -0.2% | [-0.3%, -0.2%] | 3 | |
| 97 | +| Improvements ✅ <br /> (secondary) | -0.3% | [-0.3%, -0.3%] | 1 | |
| 98 | +| All ❌✅ (primary) | -0.2% | [-0.3%, -0.2%] | 3 | |
| 99 | + |
| 100 | + |
| 101 | +Use `ThinVec` more in the AST [#104754](https://github.com/rust-lang/rust/pull/104754) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=f715e430aac0de131e2ad21804013ea405722a66&end=3fee48c161a48b0c142d3998fff56faee96bd56c&stat=instructions:u) |
| 102 | + |
| 103 | +| (instructions:u) | mean | range | count | |
| 104 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 105 | +| Regressions ❌ <br /> (primary) | 0.5% | [0.4%, 0.6%] | 9 | |
| 106 | +| Regressions ❌ <br /> (secondary) | - | - | 0 | |
| 107 | +| Improvements ✅ <br /> (primary) | -0.5% | [-1.5%, -0.2%] | 80 | |
| 108 | +| Improvements ✅ <br /> (secondary) | -0.8% | [-3.3%, -0.2%] | 44 | |
| 109 | +| All ❌✅ (primary) | -0.4% | [-1.5%, 0.6%] | 89 | |
| 110 | + |
| 111 | + |
| 112 | +#### Mixed |
| 113 | + |
| 114 | +Enable CopyProp [#107449](https://github.com/rust-lang/rust/pull/107449) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=dc7a676778706bde3b50ff6d4fe81e2955bd4847&end=639377ed737b25830ec44dc6acf93467c980316a&stat=instructions:u) |
| 115 | + |
| 116 | +| (instructions:u) | mean | range | count | |
| 117 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 118 | +| Regressions ❌ <br /> (primary) | 0.3% | [0.3%, 0.4%] | 2 | |
| 119 | +| Regressions ❌ <br /> (secondary) | 1.2% | [0.1%, 2.7%] | 5 | |
| 120 | +| Improvements ✅ <br /> (primary) | -1.4% | [-4.4%, -0.3%] | 151 | |
| 121 | +| Improvements ✅ <br /> (secondary) | -1.7% | [-5.5%, -0.3%] | 57 | |
| 122 | +| All ❌✅ (primary) | -1.4% | [-4.4%, 0.4%] | 153 | |
| 123 | + |
| 124 | +The improvements here massively outweigh the tiny number of regressions. |
| 125 | + |
| 126 | +(https://github.com/rust-lang/rust/pull/107449#issuecomment-1432772163) |
| 127 | + |
| 128 | +Factor query arena allocation out from query caches [#107833](https://github.com/rust-lang/rust/pull/107833) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=9a7cc6c32f1a690f86827e4724bcda85e506ef35&end=947b696ce0ce42c98b8fb82ffa0735ade051466c&stat=instructions:u) |
| 129 | + |
| 130 | +| (instructions:u) | mean | range | count | |
| 131 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 132 | +| Regressions ❌ <br /> (primary) | - | - | 0 | |
| 133 | +| Regressions ❌ <br /> (secondary) | 1.0% | [0.2%, 1.9%] | 4 | |
| 134 | +| Improvements ✅ <br /> (primary) | -0.3% | [-0.3%, -0.3%] | 1 | |
| 135 | +| Improvements ✅ <br /> (secondary) | -0.3% | [-0.4%, -0.2%] | 3 | |
| 136 | +| All ❌✅ (primary) | -0.3% | [-0.3%, -0.3%] | 1 | |
| 137 | + |
| 138 | +Enable instcombine for mutable reborrows [#105274](https://github.com/rust-lang/rust/pull/105274) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=9aa5c24b7d763fb98d998819571128ff2eb8a3ca&end=231bcd131d109d41e5705f48b9c5a6e70c98ff80&stat=instructions:u) |
| 139 | + |
| 140 | +| (instructions:u) | mean | range | count | |
| 141 | +|:----------------------------------:|:-----:|:--------------:|:-----:| |
| 142 | +| Regressions ❌ <br /> (primary) | 0.6% | [0.2%, 3.1%] | 16 | |
| 143 | +| Regressions ❌ <br /> (secondary) | 0.4% | [0.2%, 0.5%] | 11 | |
| 144 | +| Improvements ✅ <br /> (primary) | -0.7% | [-2.3%, -0.3%] | 32 | |
| 145 | +| Improvements ✅ <br /> (secondary) | -1.0% | [-1.7%, -0.3%] | 17 | |
| 146 | +| All ❌✅ (primary) | -0.3% | [-2.3%, 3.1%] | 48 | |
| 147 | + |
| 148 | +This also provides significant wins for binary size, up to 8.8% in stm32f4. |
0 commit comments