2021-06-22 Triage Log

A few small regressions on smaller benchmarks (e.g., helloworld), likely centered around more IR being generated in a few cases.

Triage done by @simulacrum. Revision range: d192c80d2284ba6b5146bb3da586354c3762c72b..3912083821c5072f700a75589c8af6a9d3e20a21

2 Regressions, 1 Improvements, 0 Mixed

1 of them in rollups

Regressions

Rollup of 6 pull requests #86417

  • Moderate regression in instruction counts (up to 2.0% on incr-unchanged builds of helloworld-opt)
  • Not really clear what the cause here is, but the regression is concentrated on small benchmarks.

Change entry point to 🛡️ against 💥 💥-payloads #86034

  • Moderate regression in instruction counts (up to 4.5% on incr-full builds of helloworld-opt)
  • Seems to be more code getting generated into binaries, leading to regressions on the smaller benchmarks.

Improvements

optimize Eq implementation for paths #86179

  • Moderate improvement in instruction counts (up to -3.3% on incr-patched: println builds of cargo-opt)
  • Mostly wins on incremental benchmarks.

Nags requiring follow up

Backlog:

  • Follow up on regression in #85343