Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 8 Jul 2012 12:36:14 +0400
From: Solar Designer <>
Subject: Re: Rotate and bitselect investigation

On Sun, Jul 08, 2012 at 12:34:51PM +0530, Sayantan Datta wrote:
> There was a small typo in the optimized kernel. Apparently you didn't
> change from manual to bitselect in the SHA1_digest() function which is
> called in the 10K iteration.

Oh, indeed.  Good catch.

> Here's the new results with 1/4 of the KPC you used in the previous
> benchmark.
> Using device 0: GeForce GTX 570
> Benchmarking: M$ Cache Hash 2 (DCC2) PBKDF2-HMAC-SHA-1 [OpenCL]... DONE
> Raw:    30481 c/s real, 30481 c/s virtual

That's a 4% slowdown - is this solely because of the reduced KPC?

> Using device 0: Tahiti
> Benchmarking: M$ Cache Hash 2 (DCC2) PBKDF2-HMAC-SHA-1 [OpenCL]... DONE
> Raw:    99801 c/s real, 99296 c/s virtual

That's a 2% speedup, presumably due to bitselect().  I wonder what the
speedup would be with the same KPC as before.



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.