2012.04.14 10:10:36 (191075971674292225) from Daniel J. Bernstein, replying to "Samuel Neves (@sevenps)" (190786249735413762):
@sevenps @zooko @aumasson Linux now uses two incompatible little-endian ARM ABIs: soft-float and hard-float. You have to compile for both.
2012.04.13 08:48:13 (190692851200442368) from "zooko❤ⓩ🛡🦓🦓🦓 (@zooko)":
The SHA-3 contest has demonstrated that it is possible to build secure hash functions. Now can we build fast ones? https://bit.ly/HBZnkE
2012.04.13 09:15:30 (190699718639230976) from "JP Aumasson (@veorq)", replying to "zooko❤ⓩ🛡🦓🦓🦓 (@zooko)" (190692851200442368):
@zooko OTOH it's not true to say "SHA-3 won't be faster than SHA-2", unless you restrict to those platforms where SHA-2 is faster ;)
2012.04.13 14:40:37 (190781537338396673) from "zooko❤ⓩ🛡🦓🦓🦓 (@zooko)", replying to "JP Aumasson (@veorq)" (190699718639230976):
@aumasson You're right. Blake and Skein are faster than SHA-256 on most platforms for most message lengths. But SHA-256 is smaller in ASIC?
2012.04.13 14:59:21 (190786249735413762) from "Samuel Neves (@sevenps)", replying to "zooko❤ⓩ🛡🦓🦓🦓 (@zooko)" (190781537338396673):
@zooko @aumasson BLAKE should be performing much faster on h6dragon (NEON fails to build). See http://bench.cr.yp.to/web-impl/armeabi-h1mx515-crypto_hash.html vs http://bench.cr.yp.to/web-impl/armeabi-h6dragon-crypto_hash.html