Open llvmbot opened 12 years ago
Implementing a trivial float to int16_t conversion routine, Clang is unable to optimize 8 subsequent extract/store operations into one single unaligned store.
Code and x86_64 disasm in attachment.
Added similar bug. This one for int16_t -> float conversion.
Extended Description
Implementing a trivial float to int16_t conversion routine, Clang is unable to optimize 8 subsequent extract/store operations into one single unaligned store.
Code and x86_64 disasm in attachment.