Version 0.5.1 Beta 2: Fully header-only, launch config builder, multi-device fixes

[eyalroz]

Build mechanism

• Make the nvtx library header-only #307 The library is now entirely header-only (the NVTX wrappers, which used to be compiled, are now all within headers).

New supported features

• In NVTX, be willing to accept both narrow and wide character strings #308 Supporting both narrow/regular and wide character inputs for NVRTC compilation.
• Add support for naming streams, events and devices in NVTX #309 Support for naming streams, devices and events with NVTX

Concepts/facilities introduced

• Make launch config construction more convenient with a builder class #311 A Builder-pattern class for building launch configurations more easily.

Compatibility

• Restore CUDA 9.x compatibility #304 : Now compatible will all CUDA versions between 9.0 and 11.6

Bug fixes

• Leaking a current context and a primary context reference unit in scoped_current_device_fallback_t #317 No longer "leaking" references to device primary contexts which made them never be destroyed after some point. Fixing this exposed a few other latent issues involving non-existence of primary contexts: Cannot rely on scoped_current_device_fallback_t #316.
• Potential failure to set context when recording an event #314 No longer failing to enqueue events when there is no current context.
• Add missing named errors #305, Use driver API error enum values whenever applicable #306 :
  • Added missing named errors to cuda::status
  • Now using driver error codes wherever applicable (they only started to coincide with Runtime API error codes in a recent CUDA version)
  • Renamed mis-named error: cuda::status::not_ready -> cuda::status::async_operations_not_yet_completed.
• Launching on the current device instead of the selected device #315 In one of the example programs, we were launching a kernel on the current device rather than the one the user had chosen.

Miscellaneous and internal issues

• Switch the NVTX wrapper to use driver-API-style calls #310 NVTX wrapper now uses driver-API-style
• Use CUDA_VERSION instead of CUDART_VERSION #303 Using CUDA_VERSION instead of CUDA_RT_VERSION where relevant.

---

This discussion was created from the release Version 0.5.1 Beta 2: Fully header-only, launch config builder, multi-device fixes.