phc-discussions - Doubts about AntCrypt floating-point use

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAK9dnSxe1WppG-dL09zDVymn-AnDrher2d8GXbjfczzRncKp7A@mail.gmail.com>
Date: Tue, 1 Apr 2014 15:44:52 +0200
From: CodesInChaos <codesinchaos@...il.com>
To: discussions@...sword-hashing.net
Subject: Doubts about AntCrypt floating-point use

I expect the floating point use to be very non-portable.

Consider  F07(33843633) which before casting to int evaluates to
72664774.99999999931. So the cast to int should truncate it to
72664774, but my implementation actually rounds it to 72664775.

Double only has a 53 bit mantissa, but your code expects it to have
more than 64 or to be completely reproducible across implementations
(it's not).

https://www.wolframalpha.com/input/?i=sin%2833843633%2F1000000000.0%29*2%5E31