Apply hoist optimizations to Geometry.c's transforms

[akx]

Like the other recent optimization PRs: hoist im*->xsize and im*->ysize into locals so autovectorization can do its thing.

[codspeed-hq]

Merging this PR will improve performance by 13.77%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 16 improved benchmarks
✅ 539 untouched benchmarks
⏩ 335 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	test_rotate_right[1237x811-L-ROTATE_180]	4.3 ms	3.5 ms	+24.07%
⚡	test_rotate_right[1237x811-L-FLIP_LEFT_RIGHT]	4.3 ms	3.5 ms	+24.07%
⚡	test_rotate_right[1237x811-LA-ROTATE_180]	7.1 ms	6.2 ms	+13.39%
⚡	test_rotate_right[1237x811-RGB-ROTATE_180]	7.1 ms	6.2 ms	+13.39%
⚡	test_rotate_right[1237x811-RGBA-ROTATE_180]	7.1 ms	6.2 ms	+13.38%
⚡	test_rotate_right[1237x811-RGBA-FLIP_LEFT_RIGHT]	7.1 ms	6.2 ms	+13.38%
⚡	test_rotate_right[1237x811-RGB-FLIP_LEFT_RIGHT]	7.1 ms	6.2 ms	+13.37%
⚡	test_rotate_right[1237x811-LA-FLIP_LEFT_RIGHT]	7.1 ms	6.2 ms	+13.37%
⚡	test_scale[1237x811-RGB-2.14-NEAREST]	36.1 ms	32.2 ms	+11.89%
⚡	test_scale[1237x811-F-2.14-NEAREST]	36.1 ms	32.2 ms	+11.89%
⚡	test_scale[1237x811-RGBA-2.14-NEAREST]	36.1 ms	32.2 ms	+11.89%
⚡	test_scale[1237x811-I-2.14-NEAREST]	36.1 ms	32.2 ms	+11.88%
⚡	test_scale[1237x811-LA-2.14-NEAREST]	36.1 ms	32.2 ms	+11.85%
⚡	test_transform[1237x811-LA-BICUBIC-PERSPECTIVE]	192.7 ms	172.8 ms	+11.53%
⚡	test_transform[1237x811-LA-BICUBIC-AFFINE]	175.2 ms	157.4 ms	+11.29%
⚡	test_transform[1237x811-LA-BICUBIC-QUAD]	166.7 ms	150.6 ms	+10.68%

Tip

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

---

_{Comparing akx:faster-geometry-1 (cb14785) with main (c8c74c8)}

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. ↩

[radarhere]

Could you explain how this is an improvement? I mean, it wasn't using the values within a loop. There was no repetition at all.

[akx]

You mean out_xsize/out_ysize? Yes, they're not used in the loops, but I think it reads more cleanly for the human reader if it's

if (in_xsize != out_xsize || in_ysize != out_ysize) {

and not

if (in_xsize != imOut->xsize || in_ysize != imOut->ysize) {

-- with the latter, at least my code eye starts wondering why there's two different idioms for the sizes.