feat: add perf walltime support

[not-matthias]

Depends on CodSpeedHQ/runner#94

Currently blocked by: CodSpeedHQ/instrument-hooks#6

[GuillaumeLagrange]

@not-matthias can you fix CI please before I review ? 👼

Bonus question: any idea why the diff is so huge in gh display?

[not-matthias]

@not-matthias can you fix CI please before I review ? 👼

Yeah, of course!

Bonus question: any idea why the diff is so huge in gh display?

I think it's due to the clang-format pre-commit-hook (mostly valgrind.h). Should I add it to the blacklist? It's not really created by us

[codspeed-hq]

CodSpeed WallTime Performance Report

Merging #13 will degrade performances by 33.33%

_{Comparing cod-918-perf-doesnt-work-for-cpp-codspeed (da6c95f) with main (719c41f)}

Summary

⚡ 3 improvements
❌ 6 regressions
✅ 33 untouched benchmarks
🆕 16 new benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
⚡	BarTest[MyFixture]	< 1 ns	< 1 ns	+25%
❌	DoubleTest[MyTemplatedFixture, double]	< 1 ns	< 1 ns	-33.33%
⚡	FooTest[MyFixture]	< 1 ns	< 1 ns	+50%
❌	IntTest[MyTemplatedFixture, int]	< 1 ns	< 1 ns	-16.67%
❌	TestA[MyTemplate1, int]	< 1 ns	< 1 ns	-16.67%
❌	TestB[MyTemplate2, int, double]	< 1 ns	< 1 ns	-20%
❌	BM_Capture[int_string_test]	< 1 ns	< 1 ns	-33.33%
❌	BM_rand_vector	< 1 ns	< 1 ns	-20%
🆕	BM_sleep_100ms	N/A	100.1 ms	N/A
🆕	BM_sleep_100ns	N/A	58.6 µs	N/A
🆕	BM_sleep_100us	N/A	158.7 µs	N/A
🆕	BM_sleep_10ms	N/A	10.1 ms	N/A
🆕	BM_sleep_1ms	N/A	1.1 ms	N/A
🆕	BM_sleep_1ns	N/A	58.6 µs	N/A
🆕	BM_sleep_1us	N/A	59.6 µs	N/A
🆕	BM_sleep_50ms	N/A	50.1 ms	N/A
⚡	BM_Template1_Capture[two_type_test, int, double]	< 1 ns	< 1 ns	+80%
🆕	BM_sleep_100ms	N/A	100.1 ms	N/A
🆕	BM_sleep_100ns	N/A	58.6 µs	N/A
🆕	BM_sleep_100us	N/A	158.7 µs	N/A
...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

[GuillaumeLagrange]

@not-matthias can you fix CI please before I review ? 👼

Yeah, of course!

Bonus question: any idea why the diff is so huge in gh display?

I think it's due to the clang-format pre-commit-hook (mostly valgrind.h). Should I add it to the blacklist? It's not really created by us

Absolutely, rule of thumb should be that we never format external files like this, and if possible when we edit external files we make it so our formatter matches the initial style of the file to avoid unneccessary diffs, same for callgrind.h.

[GuillaumeLagrange]

@not-matthias could you make sure to rebase this on top of the windows PR once it's merged ? And then I'll review for real to avoid rework

[art049]

Looks good overall! I think we'll need to be extra careful to make sure this still builds on Windows after #15; especially since we didn't test the instrument hooks on windows yet. I'll let @GuillaumeLagrange do the more thorough review once rebased on the initial PR