Parallelize codspeed

[seddonym]

Parallelizes the codspeed benchmarks.

Benchmarking in CI was taking about 12 minutes, parallelizing it speeds up the time to completion.

[codspeed-hq]

CodSpeed Instrumentation Performance Report

Merging #204 will create unknown performance changes

_{Comparing parallelize-codspeed (9c4b35a) with master (7161f96)}

Summary

🆕 21 new benchmarks
⁉️ 21 dropped benchmarks

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

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
⁉️	test_deep_layers_large_graph_kept	16.3 ms	N/A	N/A
⁉️	test_deep_layers_large_graph_violated	10.4 ms	N/A	N/A
⁉️	test_top_level_large_graph_kept	46.6 ms	N/A	N/A
⁉️	test_top_level_large_graph_violated	234.1 ms	N/A	N/A
⁉️	test_chain_found	61.8 µs	N/A	N/A
⁉️	test_no_chain	1.1 ms	N/A	N/A
⁉️	test_chains_found	80.8 µs	N/A	N/A
⁉️	test_no_chains	1.1 ms	N/A	N/A
⁉️	test_build_django_from_cache_a_few_misses[15]	132.6 ms	N/A	N/A
⁉️	test_build_django_from_cache_a_few_misses[2]	131.3 ms	N/A	N/A
⁉️	test_build_django_from_cache_no_misses	131.1 ms	N/A	N/A
⁉️	test_build_django_uncached	6.2 s	N/A	N/A
⁉️	test_copy_graph	56.1 ms	N/A	N/A
⁉️	test_find_descendants	36.3 ms	N/A	N/A
⁉️	test_find_downstream_modules	175 µs	N/A	N/A
⁉️	test_find_matching_direct_imports	35.7 ms	N/A	N/A
⁉️	test_find_matching_modules	9.1 ms	N/A	N/A
⁉️	test_find_upstream_modules	3.4 ms	N/A	N/A
⁉️	test_get_import_details	249.1 µs	N/A	N/A
⁉️	test_modules_property_first_access	34.2 ms	N/A	N/A
...	...	...	...	...

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

[seddonym]

Not doing it at the moment, as we should find a way of parallelizing across the build graph ones which take the longest. It looks like pytest test groups doesn't work well with parametrized tests.