ra’s avatarra’s Twitter Archive—№ 32,714

  1. …in reply to @_vade
    @_vade @omershapira i mean that something is turning your source into GPU instructions and my understanding was that CUDA used an LLVM backend (might be wrong)