GPU Platform
## Kernel creation for GPUs
Merged at pycodegen/pystencils!384.
- [x] Implement GPU platforms into `pystencils.backend.platforms`
- [x] CUDA
- [x] SYCL
- [x] Implement GPU indexing and iteration space materialization
## Parameterization
- ~~[ ] Implement automatic GPU detection for the `CurrentGPU` target~~
- [x] Add GPU options to `CreateKernelOptions`
## JIT and external integrations
- [x] Implement launch grid size deduction
- [x] Implement CUDA just-in-time compilation into `pystencils.backend.jit`
Other than in the old code generator, the GPU launch configuration will be detached from the kernel function in the new backend.
The kernel function will only export its iteration range;
determination of GPU block and grid sizes is up to the JIT and the runtime environment.
task