> 9 апр. 2019 г., в 22:12, Paul Ramsey <pramsey@cleverelephant.ca> написал(а):
>
> Wow, well beyond slicing, just being able to decompress 25% faster is a win for pretty much any TOAST use case. I
guessthe $100 question is: portability? The whole reason for the old-skool code that’s there now was concerns about
memcpy’ingoverlapping addresses and Bad Things happening.
Yeah, I've observed Bad Things (actually pretty cool and neat things) during memcpy of overlapping regions even on my
laptop.
But here we never copy overlapping addresses in one memcpy() call.
Though my tests are very synthetic. Do we have a natural way to demonstrate a performance improvement? Like reference
pileof bytes...
Best regards, Andrey Borodin.