Lines Matching full:cycles
18 # non-parallelizable modes such as CBC encrypt is 3.75 cycles per byte
20 # performance for parallelizable modes is 1.25 cycles per byte. Being
84 # While Westmere processor features 6 cycles latency for aes[enc|dec]
86 # Bridge spends 8 cycles per instruction, but it can schedule them
92 # but in absolute terms it's 3.77 cycles per byte encrypted with
122 # (numbers are cycles per byte processed with 128-bit key):
142 # of additional cycles, but in 8x case number of instructions must be
156 # Add aesni_xts_[en|de]crypt. Westmere spends 1.33 cycles processing
209 # cycles which take care of loop variables...
