Migrate iai-callgrind to the renamed successor gungraun

[gamma0987]

Migrate from iai-callgrind to gungraun

Summary

• Replace iai-callgrind with gungraun and update it from v0.16 to v0.18
• Rename all iai-callgrind benchmark files (*_iai.rs -> *_gungraun.rs) and their corresponding [[bench]] entries
• All imports updated: use iai_callgrind::{…} -> use gungraun::prelude::*
• Simplify CI workflow by replacing manual Valgrind installation and iai-callgrind-runner installation with the gungraun/setup-gungraun@v1 action.
• Update benchmark cache path from target/iai to target/gungraun

Other minor changes:

• Add .nvim.lua to .gitignore
• Update deny.toml advisory comment to reference gungraun instead of iai-callgrind

Advantages of using the setup-gungraun action in the CI:

The action always tries to install the latest valgrind version. The action automatically installs the gungraun version which is used in the project, so the CI workflow no longer needs updates when bumping the gungraun version. Note that gungraun outputs the same json format as iai-callgrind, so no changes required here.

I changed the cache key since just changing the path from target/iai to target/gungraun might not change the cache hit output from the actions/cache action. So, Baselines and cache will be rebuilt from scratch on the next CI run.

[gemini-code-assist]

Code Review

This pull request migrates the project's benchmarking infrastructure from iai-callgrind to gungraun. The changes include updating workspace and package-level Cargo.toml files, updating Cargo.lock, renaming benchmark targets and files, and adjusting imports in the benchmark source code. The .gitignore file was also updated to include .nvim.lua. I have no feedback to provide.

[cubic-dev-ai]

1 issue found across 11 files

Confidence score: 3/5

• There is a concrete CI regression risk: .github/workflows/comment-profiling-changes.yaml runs baseline benchmarks on master with renamed *_gungraun targets that are not present there.
• Because the issue is severity 7/10 with high confidence (9/10), this is more than a minor housekeeping concern and could cause benchmark baseline generation to fail in normal workflow runs.
• Pay close attention to .github/workflows/comment-profiling-changes.yaml - baseline benchmark commands need to match targets that actually exist on master to avoid CI failures.

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name=".github/workflows/comment-profiling-changes.yaml">

<violation number="1" location=".github/workflows/comment-profiling-changes.yaml:65">
P1: Baseline benchmarks are executed on `master` using renamed `*_gungraun` targets that do not exist on current `master`, causing CI benchmark baseline generation to fail.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[cubic-dev-ai]

P1: Baseline benchmarks are executed on master using renamed *_gungraun targets that do not exist on current master, causing CI benchmark baseline generation to fail.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At .github/workflows/comment-profiling-changes.yaml, line 65:

<comment>Baseline benchmarks are executed on `master` using renamed `*_gungraun` targets that do not exist on current `master`, causing CI benchmark baseline generation to fail.</comment>

<file context>
@@ -66,34 +53,34 @@ jobs:
         run: |
           # Compile benchmarks
-          cargo bench --bench compile_demo_art_iai -- --save-baseline=master
+          cargo bench --bench compile_demo_art_gungraun -- --save-baseline=master
 
           # Runtime benchmarks
</file context>

[TrueDoctor]

The code looks good, a bit of a shame we can't test the full ci workflow before merging the pr

[gamma0987]

I'm sorry I can't help with that.