[jacopoabramo]

• Compile the extension for speed instead of size (NOMINSIZE)
• Run Ops when CMakeLists.txt changes

The current binary size itself is farly low (less than a MB), and the speed gains are considerable so it's a win in both cases anyway.

[codspeed]

Merging this PR will improve performance by 18.4%

⚡ 17 improved benchmarks
✅ 58 untouched benchmarks
⏩ 1 skipped benchmark¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	test_write_float_list[l64k-65536]	25 ms	18.4 ms	+36.03%
⚡	test_write_float_list[l1k-1024]	434.8 µs	329.2 µs	+32.07%
⚡	test_write_int_set[s64k-65536]	36.5 ms	29.1 ms	+25.73%
⚡	test_write_int_set[s1k-1024]	624.4 µs	505.7 µs	+23.48%
⚡	test_iterate_bounded_list[b64k]	24.7 ms	20.2 ms	+22.56%
⚡	test_write_int_dict[d1k-1024]	1.3 ms	1.1 ms	+20.64%
⚡	test_iterate_bounded_list[b1k]	434.2 µs	360.4 µs	+20.45%
⚡	test_write_int_dict[d64k-65536]	114.1 ms	95.9 ms	+19.02%
⚡	test_iterate_bounded_set[bs1k]	537.1 µs	463.9 µs	+15.77%
⚡	test_iterate_bounded_set[bs64k]	33.8 ms	29.2 ms	+15.44%
⚡	test_write_bounded_set[bs16-16]	28 µs	24.5 µs	+14.23%
⚡	test_write_bounded_list[b16-16]	28 µs	24.6 µs	+13.74%
⚡	test_read_long_list	42.1 µs	37.4 µs	+12.52%
⚡	test_write_int_dict[d16-16]	67.8 µs	60.4 µs	+12.3%
⚡	test_read_bounded_list[b16]	32.4 µs	29.1 µs	+11.14%
⚡	test_write_int_set[s16-16]	52.4 µs	47.2 µs	+10.94%
⚡	test_iterate_bounded_list[b16]	56.8 µs	51.4 µs	+10.45%

Tip

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

---

_{Comparing build/optimize-for-speed (d167ded) with main (47b081d)}

Footnotes

1. 1 benchmark was skipped, so the baseline result was used instead. If it was deleted from the codebase, click here and archive it to remove it from the performance reports. ↩