Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Thu, 6 Jan 2011 01:50:56 -0800
From: Dhiru Kholia <>
Subject: Re: error in the topic -> my card is 4850 and not 4870

On Wed, Jan 5, 2011 at 1:33 AM,
<> wrote:
> Benchmarking: Raw MD5 [raw-md5-opencl]...
> OpenCL Platform: <<<Apple>>> and device: <<<Radeon HD 4850>>>
> Abort trap

This is a bit puzzling. Does your platform support
cl_khr_byte_addressable_store  extension (run CLInfo?). Even in case
of no support, the patch should have aborted with compilation
failures. It would be interesting to enable debugging (add -g to JtR's
Makefile), rebuild  and run JtR under gdb to get a backtrace.

My patches require support for cl_khr_byte_addressable_store to be
present in your OpenCL implementation (ATI 48XX series don't have that
extension AFAIK). A compile time option could be added to enable
alternate code which works for 48xx series but I am not 100% sure if
"it is worth it?" at this time. Thanks for compiling and running the
patches. :-)


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.