Stream: git-wasmtime

Topic: wasmtime / PR #11128 x64: Add lowerings for the `bt` inst...

Wasmtime GitHub notifications bot (Jun 24 2025 at 17:34):

alexcrichton opened PR #11128 from alexcrichton:x64-bt to bytecodealliance:main:

This commit builds on the previous refactorings to leverage the bt instruction on x64 which moves a specific bit into the CF bit of EFLAGS which can be useful for comparisons/branches/selects/etc.

<!--
Please make sure you include the following information:

• If this work has been discussed elsewhere, please include a link to that
  conversation. If it was discussed in an issue, just mention "issue #...".

• Explain why this change is needed. If the details are in an issue already,
  this can be brief.

Our development process is documented in the Wasmtime book:
https://docs.wasmtime.dev/contributing-development-process.html

Please ensure all communication follows the code of conduct:
https://github.com/bytecodealliance/wasmtime/blob/main/CODE_OF_CONDUCT.md
-->

Wasmtime GitHub notifications bot (Jun 24 2025 at 17:34):

alexcrichton requested wasmtime-compiler-reviewers for a review on PR #11128.

Wasmtime GitHub notifications bot (Jun 24 2025 at 17:34):

alexcrichton requested wasmtime-core-reviewers for a review on PR #11128.

Wasmtime GitHub notifications bot (Jun 24 2025 at 17:34):

alexcrichton requested fitzgen for a review on PR #11128.

Wasmtime GitHub notifications bot (Jun 24 2025 at 19:38):

fitzgen submitted PR review.

Wasmtime GitHub notifications bot (Jun 24 2025 at 19:38):

fitzgen created PR review comment:

Do we expect this to use a bt instruction? If not, then why not?

Wasmtime GitHub notifications bot (Jun 24 2025 at 19:38):

fitzgen created PR review comment:

Slightly easier to understand, for me at least, like this:

        if val.count_ones() == 1 {
            Some(u8::try_from(val.trailing_zeros()).unwrap())
        } else {
            None
        }

Wasmtime GitHub notifications bot (Jun 24 2025 at 19:38):

fitzgen created PR review comment:

Same here?

Wasmtime GitHub notifications bot (Jun 24 2025 at 20:45):

alexcrichton updated PR #11128.

Wasmtime GitHub notifications bot (Jun 24 2025 at 20:46):

github-actions[bot] commented on PR #11128:

Subscribe to Label Action

cc @cfallin, @fitzgen

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

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

• cfallin: isle
• fitzgen: isle

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

Learn more.
</details>

Wasmtime GitHub notifications bot (Jun 24 2025 at 20:47):

alexcrichton submitted PR review.

Wasmtime GitHub notifications bot (Jun 24 2025 at 20:47):

alexcrichton created PR review comment:

It's intentional yeah, and corresponds to this lowering rule. My rationale is "LLVM is probably smarter than I" and I'm assuming LLVM did things like micro-op tests or something to figure out that test has a lower latency than bt or something like that. No idea if that's actually the case though, just mirroring them.

Wasmtime GitHub notifications bot (Jun 24 2025 at 20:58):

cfallin submitted PR review.

Wasmtime GitHub notifications bot (Jun 24 2025 at 20:58):

cfallin created PR review comment:

According to Agner Fog's tables, taking Skylake as a random reasonable baseline µarch, test has one cycle latency and a throughput of 4/cycle (four ALU ports can execute it) while bt has one cycle latency and a throughput of 2/cycle (two ALU ports can execute it). So probably reasonable to prefer test to bt where possible.

Wasmtime GitHub notifications bot (Jun 24 2025 at 21:19):

alexcrichton submitted PR review.

Wasmtime GitHub notifications bot (Jun 24 2025 at 21:19):

alexcrichton created PR review comment:

Nice that sounds like an excellent reference to put as documentation in the lowering rules so I'll do that.

Wasmtime GitHub notifications bot (Jun 24 2025 at 22:56):

alexcrichton updated PR #11128.

Wasmtime GitHub notifications bot (Jun 25 2025 at 15:06):

fitzgen submitted PR review:

Thanks!

Wasmtime GitHub notifications bot (Jun 25 2025 at 15:29):

fitzgen merged PR #11128.

Last updated: Dec 06 2025 at 06:05 UTC