Do we have any plans to support the cudaFuncGetName function?

[pipul]

cat ./cuda_bindings/docs/source/release/12.6.0-notes.md

CUDA Python 12.6.0 Release notes

Released on August 1, 2024

Highlights

• Rebase to CUDA Toolkit 12.6
• Resolve Issue #32: Add 'pywin32' as Windows requirement
• Resolve Issue #72: Allow both lists and tuples as parameter
• Resolve Issue #73: Fix 'cuLibraryLoadData' processing of parameters

Limitations

CUDA Functions Not Supported in this Release

• Symbol APIs
  • cudaGraphExecMemcpyNodeSetParamsFromSymbol
  • cudaGraphExecMemcpyNodeSetParamsToSymbol
  • cudaGraphAddMemcpyNodeToSymbol
  • cudaGraphAddMemcpyNodeFromSymbol
  • cudaGraphMemcpyNodeSetParamsToSymbol
  • cudaGraphMemcpyNodeSetParamsFromSymbol
  • cudaMemcpyToSymbol
  • cudaMemcpyFromSymbol
  • cudaMemcpyToSymbolAsync
  • cudaMemcpyFromSymbolAsync
  • cudaGetSymbolAddress
  • cudaGetSymbolSize
  • cudaGetFuncBySymbol
• Launch Options
  • cudaLaunchKernel
  • cudaLaunchCooperativeKernel
  • cudaLaunchCooperativeKernelMultiDevice
• cudaSetValidDevices
• cudaVDPAUSetVDPAUDevice
• cudaFuncGetName
• cudaFuncGetParamInfo

[leofang]

What is your use case? Note that cudaKernel_t from CUDA Runtime is interchangeable with CUkernel from CUDA Driver (see here), so depending on how you get your kernel, you might be able to use the driver API cuKernelGetName.