vortex-cuda format in benchmarks

[onursatici]

Does this PR closes an open issue or discussion?

• Closes #.

What changes are included in this PR?

allows benchmarks to create and ran against vortex-cuda files, which are vortex files that only include cuda compatible encodings

What is the rationale for this change?

How is this change tested?

Are there any user-facing changes?

[codspeed-hq]

Merging this PR will degrade performance by 21.88%

❌ 16 regressed benchmarks
✅ 1119 untouched benchmarks
⏩ 1268 skipped benchmarks¹

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

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	old_raw_prim_test_between[f32, 2048]	113.4 µs	144.7 µs	-21.59%
❌	Simulation	old_raw_prim_test_between[f64, 16384]	267.6 µs	298.6 µs	-10.39%
❌	Simulation	old_raw_prim_test_between[f64, 2048]	118.3 µs	149.8 µs	-21.04%
❌	Simulation	old_raw_prim_test_between[f32, 16384]	232.3 µs	262.4 µs	-11.49%
❌	Simulation	old_raw_prim_test_between[i32, 16384]	233.3 µs	263.3 µs	-11.39%
❌	Simulation	old_raw_prim_test_between[i64, 2048]	120.2 µs	151.2 µs	-20.46%
❌	Simulation	old_raw_prim_test_between[u32, 16384]	236.4 µs	266.1 µs	-11.16%
❌	Simulation	old_raw_prim_test_between[i32, 2048]	113.2 µs	144.9 µs	-21.88%
❌	Simulation	old_raw_prim_test_between[u64, 2048]	121 µs	151.3 µs	-20.03%
❌	Simulation	old_alp_prim_test_between[f32, 2048]	139.7 µs	178 µs	-21.5%
❌	Simulation	old_raw_prim_test_between[u32, 2048]	114.3 µs	145.3 µs	-21.35%
❌	Simulation	old_alp_prim_test_between[f64, 2048]	146.2 µs	178 µs	-17.87%
❌	Simulation	old_bp_prim_test_between[i16, 2048]	133 µs	161.9 µs	-17.82%
❌	Simulation	old_bp_prim_test_between[i16, 16384]	267.9 µs	297.8 µs	-10.04%
❌	Simulation	old_bp_prim_test_between[i32, 2048]	133 µs	162.6 µs	-18.2%
❌	Simulation	old_bp_prim_test_between[i64, 2048]	140.4 µs	170.9 µs	-17.84%

---

_{Comparing os/vortex-cuda-files (e878f8e) with develop (18b5846)²}

Footnotes

1. 1268 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 develop (001c920) during the generation of this report, so 18b5846 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[robert3005]

If we want to use this we should also add GPU runners to benchmarks. I think it's too early to focus on benchmarking these

[joseph-isaacs]

Can you clean this up once its gone

[onursatici]

If we have different strategies we expose, we should be able to generate benchmark data with all of those. This one is named cuda compatible but it is just a subset of encodings like the compact encodings, and I should be able to generate those and run my existing benchmark code on it. If I choose to run the query benchmark as is, it will run but will be executed on the cpu. If I have a gpu benchmark in the future, I still can use the data-gen binary to generate the files and now run on the gpu.

That being said, I am happy to keep this on my branch