That is a very valid concern and a key motivation for the proposed StreamExecutor project (http://lists.llvm.org/pipermail/llvm-dev/2016-March/096576.h...).
Thanks for your interest, and hope you like it! Yes, it is currently incomplete, but I'd say at least 80% of the optimizations are upstreamed already. Also, folks in the LLVM community are actively working on that. For…
http://llvm.org/docs/CompileCudaWithLLVM.html
You are right that gpucc still depends on NVIDIA's ptxas tool that translates PTX to native binaries. NVIDIA does not publish the specification of their native binaries. Besides that, it is fully open-source.
It currently generates NVIDIA's PTX only.
That is a very valid concern and a key motivation for the proposed StreamExecutor project (http://lists.llvm.org/pipermail/llvm-dev/2016-March/096576.h...).
Thanks for your interest, and hope you like it! Yes, it is currently incomplete, but I'd say at least 80% of the optimizations are upstreamed already. Also, folks in the LLVM community are actively working on that. For…
http://llvm.org/docs/CompileCudaWithLLVM.html
You are right that gpucc still depends on NVIDIA's ptxas tool that translates PTX to native binaries. NVIDIA does not publish the specification of their native binaries. Besides that, it is fully open-source.
It currently generates NVIDIA's PTX only.