Date: Thu, 16 Jan 2003 00:37:00 +0300 From: Solar Designer <solar@openwall.com> To: Marcel Moolenaar <marcel@xcllnt.net> Cc: Dominic Marks <dom@cus.org.uk>, freebsd-ia64@FreeBSD.ORG Subject: Re: Request for performance measurement of john the ripper on ia64 Message-ID: <20030115213700.GA10813@openwall.com> In-Reply-To: <20030115212307.GA585@athlon.pn.xcllnt.net> References: <20030115054607.GB4457@openwall.com> <JOEOLBOCFFAPLGDIKKCBAEBKCBAA.dominic_marks@btinternet.com> <20030115195445.GA6527@openwall.com> <20030115200902.GA3078@cus.org.uk> <20030115203545.GA6881@openwall.com> <20030115204525.GA3746@cus.org.uk> <20030115210104.GA7052@openwall.com> <20030115212307.GA585@athlon.pn.xcllnt.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Jan 15, 2003 at 01:23:07PM -0800, Marcel Moolenaar wrote: > Machine: HP rx2600 > CPU: McKinley 900MHz (SMP machine with UP kernel) Thank you! > Version: 1.6 (port) > Perf. compile options: -O -fomit-frame-pointer -funroll-loops > > Benchmarking: Standard DES [64/64 BS]... DONE > Many salts: 181609 c/s real, 182349 c/s virtual > Only one salt: 161309 c/s real, 161698 c/s virtual > Version: 1.6.32 (development) > Perf. compile options: -O2 -fomit-frame-pointer -funroll-loops > > Benchmarking: Traditional DES [64/64 BS]... DONE > Many salts: 410439 c/s real, 412330 c/s virtual > Only one salt: 335115 c/s real, 336375 c/s virtual I think that some of the difference is due to -O vs. -O2. I'd expect a 20% difference for this benchmark between the versions of John on IA64. If you're interested and have the time, one other thing to try is setting DES_BS_VECTOR to a value greater than 1 (2, 3, 4), -- this might let John use more instruction-level parallelism. -- /sd To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-ia64" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030115213700.GA10813>