On 22/05/19 6:08 PM, Anton Youdkevitch wrote:
They were. The difference was up to 50% in some cases. But this is the
data for bench-memcpy. Reports for bench-memcpy-walk and
bench-memcpy-random do not have alignment info in them.
So, unaligned accesses can be much slower that aligned ones. Or, at least,
this is how the benchmarks measure them. We already know that the results
are not very stable, though.
bench-memcpy-walk mixes in the misaligned sizes, but it shouldn't have a
big impact because 1) the source and destination are not mutually
misaligned and 2) the alignment code is usually a tiny portion of the
execution time and hence shouldn't show up in memcpy-walk.