Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20120411214249.GA18941@openwall.com>
Date: Thu, 12 Apr 2012 01:42:49 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: MSCash2 OpenCL (was: OpenCL tests on HD 7970)

Hi Sayantan,

Please do not unnecessarily reuse existing threads on this mailing list
for new topics.  Doing so results in messages on different topics being
mixed up in the threads as seen in the list archives.  When you need to
post something on a new topic, please send the message to the list
address anew (not as a "reply" to anything).  When you need to add
something to an existing topic, please locate the proper thread to add
to (then you may "reply" to it).  If you do not feel the need to quote
anything from a message you're replying to, this means that you probably
should not be replying to that message, but should be starting a new
thread or posting to another (more relevant) thread (where you actually
have/need to quote some context for your reply).  Thanks!  (And yes,
others on this list, including me, are not always very careful about
this, yet I felt the advice could be helpful.)

On Thu, Apr 12, 2012 at 12:11:22AM +0530, SAYANTAN DATTA wrote:
> I will post a new patch soon which will incrase the speeds furthur by
> around 13%. Currently I'm trying to squeez in as many optimizations as
> possible and will be ready in a day or two. Here's a sample benchmark for
> my new codes:
> 
> OpenCL platform 0: AMD Accelerated Parallel Processing, 2 device(s).
> Using device 0: ATI RV770
> Benchmarking: MSCASH2-OPENCL [PBKDF2_HMAC_SHA1]... DONE
> Raw:    19277 c/s real, 19306 c/s virtual

Looks good.  Please do post the patch.

Also, please start cleaning up your code - as separate patches.  Please
do not include any functional changes or optimizations in those patches,
limit them just to code cleanups.  (Generate separate patches for the
functional changes and optimizations.)  You may use this indent command
to format your source files:

indent -kr -i8 -nlp -nbbo -ncs -l79 -lc79

Please note that indent fails on our struct fmt_main initializers (it
indents them in a weird way).  So please only run indent on your files
once, then re-format the struct fmt_main initializer manually (look at
the one in DES_fmt.c as a sample of desirable indentation), and from
that point on just be careful with further edits to keep them in the
same style.

Besides indentation, you may want to make other kinds of cleanups as
well.  For example, you have many functions that are not declared
static, yet are probably only meant for use from the same source file -
you need to declare them static to avoid accidental use from elsewhere.

Thanks,

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.