I thought that although it works like a charm on 480, both Todd and
Mark said that it doesn't always work - it just makes the crashes
happen vastly less often - esp. on 580.
On Wed, Jun 1, 2011 at 10:20 AM, Ross Walker <ross.rosswalker.co.uk> wrote:
>
>
>> Since when was that not needed for GeForce Ross? Or did you default
>> it to being on? The HW bug hits every GeForce card in existence. If
>> you're not seeing it, you're getting lucky.
>
> Amber 11 with bugfix.12 has
>
> static const int SM_2X_LOCALFORCES_THREADS_PER_BLOCK = 256;
>
> instead of 256 which provides the stability on GeForce cards (no lockups
> reported since this change) and has minimal impact on performance. For the
> development code however, the ifdef is required and will (I hope) be
> automated for the main release.
>
> All the best
> Ross
>
> /\
> \/
> |\oss Walker
>
> ---------------------------------------------------------
> | Assistant Research Professor |
> | San Diego Supercomputer Center |
> | Adjunct Assistant Professor |
> | Dept. of Chemistry and Biochemistry |
> | University of California San Diego |
> | NVIDIA Fellow |
> | http://www.rosswalker.co.uk | http://www.wmd-lab.org/ |
> | Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
> ---------------------------------------------------------
>
> Note: Electronic Mail is not secure, has no guarantee of delivery, may not
> be read every day, and should not be used for urgent or sensitive issues.
>
>
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jun 01 2011 - 10:30:06 PDT
