perf: use mimalloc as the runner's global allocator

[not-matthias]

TLDR: The released runner is a musl build, and musl's allocator is slow on many small allocations and serializes threads. Post-processing a memory run decodes the whole memtrack artifact with allocations per event, so it took 8-19 minutes on the largest platform shard. The runner now uses mimalloc, as memtrack already does.

• Adds mimalloc as the global allocator in the codspeed binary.
• Independent of test(runner-shared): benchmark finding module events in a memtrack artifact #576/perf(runner-shared): decode memtrack frames for module events in parallel #578; it speeds up the decode before and after them.

Time from memtrack's last log line to the end of post-processing, platform memory shards, musl build like the release, 3 runs each (medians, seconds):

Shard (events)	#576	#576 + mimalloc	#578	#578 + mimalloc
shard-2 (62M)	823	121 (2 runs)	522	24
shard-4 (39M)	335	69	209	16
shard-5 (7.8M)	58	17	68	6

Review note: affects every allocation in the runner, not just the memtrack decode.

[codspeed]

Merging this PR will improve performance by 32.52%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚠️ 8 benchmarks spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

• find_module_events[1000000]
• find_module_events_with_stacks[1000000]
• encode_events_realistic[16]
• sleep 1
• encode_events_realistic[2]
• encode_events_realistic[4]
• encode_events_realistic[8]
• encode_events_realistic[1]

⚡ 2 improved benchmarks
✅ 39 untouched benchmarks
⏩ 6 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	WallTime	encode_events_realistic[16]	130.5 ms	92.5 ms	+41.1%
⚡	WallTime	encode_events_realistic[8]	128.6 ms	103.3 ms	+24.46%

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing cod-3819-runner-mimalloc (5ee7e58) with main (5ff86d7)}

Footnotes

1. 6 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[GuillaumeLagrange]

@not-matthias this is already done as part of #531

[greptile-apps]

Confidence Score: 5/5

[High impact] The PR appears safe to merge; no actionable failure was found.

Summary

The PR switches the codspeed binary to mimalloc and shares its dependency declaration across the workspace.

• The runner uses mimalloc for its memory allocations.

The previous review commit is not an ancestor of this head, so the full PR diff was reviewed. No previous finding threads were supplied.

_{Reviews (3) · Last reviewed commit: "test(runner-shared): run the memtrack be..." · Reviewed by Greptile}

[GuillaumeLagrange]

olgtm