CUDA vs OpenCL
This cluster centers on discussions comparing CUDA's dominance in GPU programming to alternatives like OpenCL, SYCL, Vulkan compute, and efforts for cross-platform GPU compute solutions that enable running standard C/C++ code on GPUs.
Activity Over Time
Top Contributors
Keywords
Sample Comments
Running normal c directly on gpu has been the dream for a long time. this looks excellent
Shouldn't we have an API that can speak to both CUDA and opencl? Or is opencl sufficiently capable?
You're not aware of OpenCL, GLSL, or CUDA then!
Something like https://en.wikipedia.org/wiki/OpenCL ?
Why is CUDA so prevalent oppose to its alternatives?
Cool story. How supported is OpenCL compared to CUDA again?
No mention of OpenCL / VK compute?
Do they have the equivalent of CUDA, and what is it called?
Fuck CUDA, OpenMP supports GPUs
No there isn't, Google developed a CUDA frontend for clang/llvm which targets PTX (nvidias virtual machine instruction set), in principle someone could come along and implement a different backend for this frontend targeting for example AMDs gpu instruction architecture or something different. The other ingredient missing is a reimplementation of CUDAs runtime libraries.