[Roadmap] [QDP] GPU/Cuda kernel implementation

[ryankert01]

This is the mega roadmap for GPU related progress, feel free to create an issue and link it to the roadmap:

1. Implement kernel #677
2. Optimizations: #706
   • Parallel normalization kernel
   • Coalesced memory access patterns
   • Warp-level optimizations
   • Stream support for async execution
3. Implement cuda optional test @ryankert01
   • If no cuda device, the test skips
4. Implement suitable benchmark
   • The Scaling Test (Latency vs. Qubits)
   • The DataLoader Test (Batch Throughput) #687
5. gracefully handles OOM #688
6. Future encoding methods:
   • launch_angle_encode (angle encoding)
   • launch_basis_encode (basis encoding)
   • launch_iqp_encode (IQP encoding)
7. apache license pre-commit(cuda) #684
8. (after poc) move pre-processing from cpu to gpu