Merge pull request #725 from Mark-Simulacrum/triage-up

Update triage log to account for performance investigation

Author: nnethercote

triage/2020-07-21.md

@@ -9,7 +9,7 @@ regressions/improvements. 3 improvements, one of which was a reversion of a
 regression. Thanks for Mark-Simulacrum and eddyb for follow-up measurements and
 adding new tooling to help investigate regressions in rollups. A follow-up
 thread on Zulip is
-[here](https://rust-lang.zulipchat.com/#narrow/stream/247081-t-compiler.2Fperformance/topic/Follow-up.20to.20the.20bad.20perf.20triage.20of.202020-07-21)
+[here](https://rust-lang.zulipchat.com/#narrow/stream/247081-t-compiler.2Fperformance/topic/Follow-up.20to.20the.20bad.20perf.20triage.20of.202020-07-21).
 
 In better news, rustdoc performance is now being benchmarked, thanks to the
 efforts of [Joshua Nelson](https://github.com/rust-lang/rustc-perf/pull/675).
@@ -18,7 +18,7 @@ Triage done by njn. Revision range: [9d09331e00b02f81c714b0c41ce3a38380dd36a2..7
 
 Regressions
 - [Rollup of 18 pull requests #74461](https://github.com/rust-lang/rust/pull/74461#issuecomment-660492867)
-  ([instructions](https://perf.rust-lang.org/compare.html?start=39d5a61f2e4e237123837f5162cc275c2fd7e625&end=d3df8512d2c2afc6d2e7d8b5b951dd7f2ad77b02&stat=instructions:u)):
+  ([instructions](https://perf.rust-lang.org/compare.html?start=39d5a61f2e4e237123837f5162cc275c2fd7e625&end=d3df8512d2c2afc6d2e7d8b5b951dd7f2ad77b02-v1&stat=instructions:u)):
   Up to 73.0% losses on many benchmarks.
   [#74416](https://github.com/rust-lang/rust/pull/74416#issuecomment-660545238)
   is at fault; it was reverted in
@@ -27,9 +27,6 @@ Regressions
   ([instructions](https://perf.rust-lang.org/compare.html?start=7d31ffc1ac9e9ea356e896e63307168a64501b9d&end=1fa54ad9680cc82e7301f8ed4e9b7402dfd6ce0e&stat=instructions:u),
    [max-rss](https://perf.rust-lang.org/compare.html?start=7d31ffc1ac9e9ea356e896e63307168a64501b9d&end=1fa54ad9680cc82e7301f8ed4e9b7402dfd6ce0e&stat=max-rss)):
   Up to 44.8% instruction increases on many benchmarks, and 5-10% max-rss increases on many benchmarks. Being reverted in [#74613](https://github.com/rust-lang/rust/pull/74613).
-- [Rollup of 11 pull requests #74468](https://github.com/rust-lang/rust/pull/74468#issuecomment-661371815)
-  ([instructions](https://perf.rust-lang.org/compare.html?start=d3df8512d2c2afc6d2e7d8b5b951dd7f2ad77b02&end=7d31ffc1ac9e9ea356e896e63307168a64501b9d&stat=instructions:u)):
-  Up to 10.2% losses on many benchmarks. [#74069](https://github.com/rust-lang/rust/pull/74069) [responsible](https://perf.rust-lang.org/compare.html?start=d3df8512d2c2afc6d2e7d8b5b951dd7f2ad77b02&end=cfade73820883adf654fe34fd6b0b03a99458a51), revert [posted](https://github.com/rust-lang/rust/pull/74611).
 - [Support const args in type dependent paths (Take 2) #74113](https://github.com/rust-lang/rust/pull/74113#issuecomment-661470398)
   ([instructions](https://perf.rust-lang.org/compare.html?start=d9e8d6290745a65025a3e082aea72fbe372292c6&end=7e11379f3b4c376fbb9a6c4d44f3286ccc28d149&stat=instructions:u)):
   Up to 1.5% losses on numerous benchmarks, which was expected.
@@ -59,3 +56,27 @@ Improvements
   Wins of up to 62.6% and losses of up to 8.5%, all on rustdoc builds. Overall,
   the improvements greatly outweigh the losses. (Landed in rollup
   [#74408](https://github.com/rust-lang/rust/pull/74408).)
+
+## Notes
+
+[Rollup of 11 pull requests #74468](https://github.com/rust-lang/rust/pull/74468#issuecomment-661371815)
+ * ([instructions](https://perf.rust-lang.org/compare.html?start=d3df8512d2c2afc6d2e7d8b5b951dd7f2ad77b02-v1&end=7d31ffc1ac9e9ea356e896e63307168a64501b9d&stat=instructions:u))
+
+This rollup was originally judged as responsible for a 10% regression in
+instrutions:u. However, since then, it has been determined that the likely cause
+of this regression is actually perf's
+[move](https://github.com/rust-lang/rustc-perf/commit/048360c77a74244ba6e70d9b3f2bcd7779b8129f) to using `x.py dist`-shipped std's
+rather than building one locally. Investigation into *why* this move caused a
+regression is as yet not done, but is being tracked in [rustc-perf#724](https://github.com/rust-lang/rustc-perf/issues/724).
+
+Initially [#74069](https://github.com/rust-lang/rust/pull/74069) and/or
+[#72414](https://github.com/rust-lang/rust/pull/72414) were suspected
+as the cause of the regression, but further testing showed that to not be the
+case.
+
+ * The original rollup, #74468, landed with a [5-10% performance regression](https://perf.rust-lang.org/compare.html?start=d3df8512d2c2afc6d2e7d8b5b951dd7f2ad77b02-v1&end=7d31ffc1ac9e9ea356e896e63307168a64501b9d&stat=instructions:u)
+ * In #74611, we tested as-if a rollup of #74069 and #72414 landed. ([This](https://gist.github.com/Mark-Simulacrum/6893dff239d116947bca647f4f8128c7) is the diff between d3df85 and cfade73) That yielded [identical results to the rollup](https://perf.rust-lang.org/compare.html?start=d3df8512d2c2afc6d2e7d8b5b951dd7f2ad77b02-v1&end=cfade73820883adf654fe34fd6b0b03a99458a51).
+ * We reverted #74069 in #74611, this yielded [neutral performance results](https://perf.rust-lang.org/compare.html?start=e8b55a4ad230ebec762fdfc4f241ba98a98560af&end=fcac11993ca055bbdc7683a2f6ed7b88a838fb0f&stat=instructions:u).
+ * #74716 then tried a revert of #72414 but this also yielded [neutral performance results](https://perf.rust-lang.org/compare.html?start=900869371e13cead086f4f9809419daa6a63cfaf&end=193b2f77c9463ed7378c20bad843a9031489e215)
+
+We have since opened a PR to re-land #74069, as well: #74802.