Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20110923014235.GK21439@openwall.com>
Date: Fri, 23 Sep 2011 05:42:35 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: Fork=n

On Wed, Sep 21, 2011 at 10:18:33PM +0200, magnum wrote:
> I just recently got hold of that code and I don't even have any 
> documentation at all. Is there any? How did you use -fork?

(I think we've addressed this in other messages already.)

> BTW I have seen an MPI problem with j5c4 (neither using -fork nor -node, 
> just -inc mode under MPI) but I've had no time to look into it. I got 
> "Unexpected count: x != y" error.

Obviously, I re-worked the MPI code in inc.c while adding --node, and I
never tested the effect of my changes on MPI (we didn't intend to use
MPI in the contest, and I think we didn't use it).  Someone reported
that MPI worked for him anyway (after the contest), but that might not
be reliable info.

> I guess the -node option and MPI 
> should be unified and use the same code in most places. This will help 
> debugging MPI a lot too.

I agree, and in fact they are unified in j5c4's inc.c.  So I don't know
why MPI broke for you.

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.