Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 18 Sep 2015 19:13:49 +0300
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: multi-threaded hash table initialization

On Thu, Sep 17, 2015 at 12:53:23PM +0200, Frank Dittrich wrote:
> It also causes a warning for non-OMP builds:

The attached patch fixes this and other warnings you reported (some came
from another patch).  However, I now think none of them are related to
the runtime issues you saw.

Performance wise, the "multi-threaded hash table initialization" patch
appears to make things slightly worse on our "super" machine (roughly
10.2 to 10.5 seconds for loading of the 29M testcase, although
occasional lucky run with the patch do get below 10.2).  So I think we
should revert it for that reason (no clear performance benefit, and
potential severe performance regressions on smaller machines),
regardless of any bugs it might or might not contain.

> ... and tons of jtrts.pl failures

Is this in an OpenMP build or not?

Alexander

View attachment "john-loader-bugs1.diff" of type "text/plain" (2416 bytes)

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.