5 ms·
CUDA Shared Memory Swizzling
- KeplerBoy 1mo agoLei Mao's blog is such an amazing resource for GPU performance engineering. I am stunned by the sheer amount of insight he puts out on his blog.
- johndough 1mo agoI was wondering why the author was using braced initialization like size_t i{0}; instead of the more common size_t i = 0; Apparently, braced initialization does not allow narrowing conversion, so you'd get a compiler error for e.g. casting double to float size_t i{0.0}; and a warning for double d = 0.0; size_t i{d}; which might silently overflow size_t otherwise, so this is a bit safer. In C++, you can get the same effect without the unusual syntax by passing -Wfloat-conversion to gcc/clang, but not sure how to do that with CUDA: