Dynamic CUDA with NVIDIA's Runtime Compilation

[@UtahCppProgrammers]

Example code: github.com/LegalizeAdulthood/nvrtc-example

[@johnkost2514]

Awesome!

[@__hannibaalbarca__]

Is there asm-like for GPU as CPU.

[@UtahCppProgrammers]

PTX (Parallel Thread eXecution) is the assembly language for use with CUDA. When we compiled the CUDA C++ code with nvrtc without linking, the intermediate representation was PTX. The returned data is PTX text that you can examine in the debugger.
PTX reference documentation:
docs.nvidia.com/cuda/parallel-thread-execution/index.html