Improve benchmark test image

[akx]

The old image was highly channel-correlated, low-color, low-structure and non-representative of anything real. My bad in #9654 :)

The new image contains both low-energy, high-energy (full noise) and zero-energy segments.
If someone has an idea for an even better synthetic image, I'm all ears!

Let's see how the Codspeed numbers change - at least locally, test_quantize and test_histogram got a lot slower, because they actually need to do work 😄

Old	New

[codspeed-hq]

Merging this PR will not alter performance

✅ 12 untouched benchmarks
🆕 333 new benchmarks
⏩ 335 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
🆕	test_allocate[1237x811-L]	N/A	544.7 µs	N/A
🆕	test_allocate[1237x811-LA]	N/A	4.1 ms	N/A
🆕	test_allocate[1237x811-RGB]	N/A	4.1 ms	N/A
🆕	test_allocate[1237x811-RGBA]	N/A	4.1 ms	N/A
🆕	test_alpha_composite[1237x811-LA-mixed]	N/A	25.6 ms	N/A
🆕	test_alpha_composite[1237x811-LA-opaque]	N/A	25.6 ms	N/A
🆕	test_alpha_composite[1237x811-LA-transparent]	N/A	9.2 ms	N/A
🆕	test_alpha_composite[1237x811-RGBA-mixed]	N/A	25.6 ms	N/A
🆕	test_alpha_composite[1237x811-RGBA-opaque]	N/A	25.6 ms	N/A
🆕	test_alpha_composite[1237x811-RGBA-transparent]	N/A	9.2 ms	N/A
🆕	test_blend[1237x811-L]	N/A	2.7 ms	N/A
🆕	test_blend[1237x811-LA]	N/A	10.5 ms	N/A
🆕	test_blend[1237x811-RGB]	N/A	10.5 ms	N/A
🆕	test_blend[1237x811-RGBA]	N/A	10.5 ms	N/A
🆕	test_box_blur[1237x811-L-1]	N/A	24 ms	N/A
🆕	test_box_blur[1237x811-L-10]	N/A	24 ms	N/A
🆕	test_box_blur[1237x811-L-30]	N/A	23.9 ms	N/A
🆕	test_box_blur[1237x811-LA-1]	N/A	54.9 ms	N/A
🆕	test_box_blur[1237x811-LA-10]	N/A	55.1 ms	N/A
🆕	test_box_blur[1237x811-LA-30]	N/A	55.2 ms	N/A
...	...	...	...	...

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

---

_{Comparing akx:better-benchmark-image (854990f) with main (bdfa0e2)²}

Footnotes

1. 335 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. ↩

2. No successful run was found on main (204f2e9) during the generation of this report, so bdfa0e2 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[akx]

Merging this PR will degrade performance by 29.32%

As predicted! But no, code performance didn't change, just measurements for the more accurate...

[hugovk]

Good idea. Do you think making the image non-square would have any benefit?

[akx]

@hugovk Good point, it might. I'm a little afk right now (I'll let you know it's quite warm in Dubrovnik). And maybe also NPOT (non-power-of-2)... Honestly some tests (resizing, etc.) could benefit from running on NPOT and non-NPOT?

[hugovk]

Yes, sounds good.

[akx]

@hugovk Made it NPOT and non-square :)

[hugovk]

Thanks!