Stream: git-wasmtime

Topic: wasmtime / issue #5392 cranelift-isle: Add "partial" flag...

view this post on Zulip Wasmtime GitHub notifications bot (Dec 07 2022 at 07:24):

github-actions[bot] commented on issue #5392:

Subscribe to Label Action

cc @cfallin, @fitzgen

<details>
This issue or pull request has been labeled: "cranelift", "cranelift:area:aarch64", "cranelift:area:x64", "isle"

Thus the following users have been cc'd because of the following labels:

To subscribe or unsubscribe from this label, edit the <code>.github/subscribe-to-label.json</code> configuration file.

Learn more.
</details>

view this post on Zulip Wasmtime GitHub notifications bot (Dec 08 2022 at 01:14):

jameysharp commented on issue #5392:

On a suggestion from Chris, I ran wasmtime compile on the Spidermonkey benchmark from Sightglass, both with and without this PR, and compared the compiled .cwasm files. They were byte-for-byte identical. In conjunction with the fact that none of the precise-output filetests failed on any backend, I have pretty good confidence that this didn't change the results of instruction selection.

I had thought this PR might marginally improve performance in the instruction selection phase of compiling, due to fewer checks for whether constructors returned Option::Some.

When measuring time spent in wasmtime compile on Spidermonkey with DHAT, this PR reduces instructions retired by 0.06%, at the cost of 0.0007% more bytes read and 0.0001% more bytes written. Somehow this PR increased total heap allocations by 8 bytes.

Sightglass/perf agrees with DHAT that instructions retired during compilation are improved by a statistically significant amount but with such a small effect size that it reports this PR is "1.00x faster" on the bz2, pulldown-cmark, and spidermonkey benchmarks.

Sightglass' cpu-cycles measure and Hyperfine's wall-clock time both report that this PR is slightly slower when compiling Spidermonkey: the baseline version is "1.00x to 1.01x faster" according to Sightglass and "1.00 ± 0.02 times faster" according to Hyperfine. On the smaller benchmarks (bz2 and pulldown-cmark), Sightglass found no significant difference in cpu-cycles.

In short, this PR has almost no effect on performance by any measure.

Last updated: Nov 22 2024 at 17:03 UTC