Hi Bryan,

The driver is already upgraded to CUDA 2.3a on my test machine so i don't know what's happening under the hood. Any suggestions?

Ray

On Wed, Dec 30, 2009 at 4:24 PM, Bryan Catanzaro <bryan.catanzaro@gmail.com> wrote:
I run CUDA code on my 9400m under OS X all the time, although my Macbook doesn't have the 9600m.  Device query should be showing both your cards, if it's not, perhaps you should try upgrading your driver.

- bryan

On Dec 29, 2009, at 10:15 PM, Raymond Tay <ray@lindenlab.com> wrote:

Good suggestion Andreas. I've found the problem.

It slipped my earlier observations that whenever i ran the SDK deviceQueryDrv, the clock rate was always displaying as "0.0 Ghz" and it didn't occur to me that my Mac Book Pro (has 2 GeForce cards - 9400m and 9600m) was choosing the 9400m low powered chip and for some reason the 9400m device wasn't returning properly but i dont know enough to know why this is occurring

Is this a bug with nvidia or with mac os x? I don't know

Ray

On Tue, Dec 29, 2009 at 9:35 AM, Andreas Klöckner <lists@informa.tiker.net> wrote:
On Montag 28 Dezember 2009, Raymond Tay wrote:
> Hi Andreas,
>
> They are one and the same and it was the same card at both times. Any other
> ideas?

Try inserting the query right after PyCUDA's call to cuInit (see
src/cpp/cuda.hpp, l. 316 or thereabouts--look for cuInit). If it fails there,
too, then it's likely an Apple or Nvidia bug.

Andreas


_______________________________________________
PyCUDA mailing list
PyCUDA@tiker.net
http://tiker.net/mailman/listinfo/pycuda_tiker.net


_______________________________________________
PyCUDA mailing list
PyCUDA@tiker.net
http://tiker.net/mailman/listinfo/pycuda_tiker.net