feat: cuda gpu skeleton

[0ax1]

Provides the initial skeleton to register and run CUDA kernels.

Note that cudarc does not need to be compiled conditionally as we defer symbol loading via dynamic-loading such that vortex-cuda can also be compiled on macOS.

[codspeed-hq]

CodSpeed Performance Report

Merging this PR will degrade performance by 40.62%

_{Comparing ad/cuda-api-draft (03b7d73) with develop (293026a)}

Summary

⚡ 5 improved benchmarks
❌ 2 regressed benchmarks
✅ 1247 untouched benchmarks

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

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	bitwise_and_arrow_buffer[1024]	3.3 µs	5.5 µs	-40.62%
⚡	bitwise_not_arrow_buffer[128]	5.6 µs	2.7 µs	×2.1
⚡	bitwise_not_arrow_buffer[2048]	5.7 µs	3 µs	+92.67%
⚡	bitwise_not_arrow_buffer[1024]	5.5 µs	2.8 µs	+93.75%
⚡	bitwise_not_arrow_buffer[16384]	10.6 µs	6.1 µs	+73.02%
⚡	bitwise_not_arrow_buffer[65536]	30.8 µs	12 µs	×2.6
❌	bitwise_or_arrow_buffer[1024]	3.3 µs	5.5 µs	-40.1%

[codecov]

Codecov Report

❌ Patch coverage is 18.66667% with 61 lines in your changes missing coverage. Please review.
✅ Project coverage is 83.41%. Comparing base (293026a) to head (03b7d73).

Files with missing lines	Patch %	Lines
vortex-cuda/src/executor.rs	0.00%	37 Missing ⚠️
vortex-cuda/src/session.rs	0.00%	12 Missing ⚠️
vortex-cuda/src/kernel.rs	60.86%	9 Missing ⚠️
vortex-cuda/src/lib.rs	0.00%	3 Missing ⚠️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[joseph-isaacs]

can we have for?

[0ax1]

yaeh

[gatesn]

new should do no work. The CudaSession should hold the CudaContext, and CudaSession::new_execution_ctx() should create the stream