Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 11 Apr 2015 18:43:09 +0300
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: [GSoC] John the Ripper support for PHC finalists

On Sat, Apr 11, 2015 at 05:35:53PM +0200, Agnieszka Bielec wrote:
> again in well when I type configure I get:
> a@...l:~/sse/src$ ./configure | grep -i avx
> checking for AVX... yes
> checking for AVX2... no

Oh, I'm sorry I forgot to mention where to find the AVX2-enabled newer
version of gcc on that machine.  You need to do:

export PATH=~gcc/gcc-4.9.2/bin:$PATH LD_LIBRARY_PATH=~gcc/gcc-4.9.2/lib64:$LD_LIBRARY_PATH

> I wanted to test my sse2 implementation but on well when I want to
> compile johntheripper I have
> 
> /tmp/ccsm0C9J.s: Assembler messages:
> /tmp/ccsm0C9J.s:434: Error: no such instruction: `vfmadd312sd
> .LC5(%rip),%xmm0,%xmm2'
> make[1]: *** [formats.o] Error 1
> make[1]: Leaving directory `/space/home/a/sse/src'
> make: *** [default] Error 2

The version of binutils on well might not support all of the
instructions that recent gcc may generate for that platform, but in
practice it's been sufficient for us so far.

Why would the vfmadd312sd instruction appear in code generated for
POMELO?  This is a floating-point instruction.  Do you possibly
inadvertently use floating-point math somewhere in your code?

Alexander

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.