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 not alter performance

⚠️ 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.

⚠️ 6 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

• encode_events_realistic[16]
• encode_events_realistic[8]
• sleep 1
• encode_events_realistic[2]
• encode_events_realistic[4]
• encode_events_realistic[1]

✅ 31 untouched benchmarks
⏩ 6 skipped benchmarks¹

---

_{Comparing cod-3819-runner-mimalloc (5449177) with main (90cc8e9)}

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. ↩