Upgrade Rust toolchain to 2023-08-14

[zhassan-aws]

Description of changes:

Relevant upstream commits/PRs:

rust-lang/rust#114382: this required handling the new compare_bytes intrinsic

rust-lang/rust#114622: crate_types is no longer a method in Session

Resolved issues:

Resolves #ISSUE-NUMBER

Related RFC:

Optional #ISSUE-NUMBER.

Call-outs:

Testing:

• How is this change tested? Current regressions

• Is this a refactor change? No

Checklist

• Each commit message has a non-empty body, explaining why the change was made
• Methods or procedures are documented
• Regression or unit tests are included, or existing tests cover the modified code
• My PR is restricted to a single feature or bugfix

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 and MIT licenses.

[tautschnig]

It seems the fmt_i8 test is now taking substantially longer, but perhaps not in a surprising way: this test also has an change in the number of instructions. Perhaps this exercises the new intrinsic? IMHO ok to accept this performance change.

[zhassan-aws]

I'm a bit confused about the fmt_i8 test because in #2576, it was reported that its verification time increased from ~6s to 10 minutes. Looks like the time went down at some point, and then went up again with this toolchain upgrade? The increase is not as substantial as before though as it only went up to ~44 seconds (and solver runtime went up to ~19 seconds).

@celinval do you have any more information on the performance trend of this test?

[zhassan-aws]

Ah, looks like changing the default solver to cadical brought down the time from ~1059 seconds to ~42 seconds:

https://github.com/model-checking/kani/actions/runs/5744858062

But based on the perf-benchcomp action runs, there is some variance in the runtime of this benchmark, so we can safely ignore the increase in runtime seen on this PR.