Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 4 Mar 2012 05:55:05 +0400
From: Solar Designer <>
Subject: Re: CUDA & OpenCL status

On Sun, Mar 04, 2012 at 02:26:04AM +0100, magnum wrote:
> It sure was my fault and it had little to do with GPU, it was the fact I
> implicitly use pre-built intrinsics for GPU builds but forgot an
> important #define...
> Fixed in git trees now. Thanks for reporting.

Thanks for fixing.  BTW, should we possibly add these lines:

NVIDIA_CUDA = /usr/local/cuda

to the Makefile?

If not, then where are these settings supposed to come from?

> Using gcc 4.6, I never hit it.

Oh, I have to use gcc 4.5 for CUDA-enabled builds, because nvcc says it
does not like 4.6.  Luckily, Ubuntu still provides that version as well,
so I did:

apt-get install gcc-4.5 g++-4.5

Then created symlinks named just gcc and g++ in my ~/john directory and
pointing to /usr/bin/gcc-4.5 and /usr/bin/g++-4.5.  I prepend ~/john to
PATH for those builds.

We'll need to document known issues and workarounds like this on some
wiki page, even though they're not specific to JtR.

BTW, one reason why I chose to use Ubuntu on this machine was to stay
close to mainstream - likely run into the same issues that others would
and avoid issues that others would also not run into.


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.