Unable to get CUDA working on OSX 10.6.5 / Quadro 4000

Hello all,

We are trying to get a Quadro 4000 card running in a Mac Pro. This machine previously had a GTX 285 in it that ran correctly. This Mac Pro was purchased in May of 2010, so it should definitely be recent enough to support this card. We have tried/done the following:

  • Updated the CUDA drivers to the latest versions available.
  • Removed the drivers, then reinstalled them.
  • Reinstalled Mac OSX with the GT120 video card, then updated to 10.6.5, then installed the CUDA drivers, then installed the card.
  • Put in only the Quadro 4000 card.
  • Put in the Quadro and GT 120 card.

There have also been numerous reboots along the way. The card is always detected as a display adapter just fine in the System Information panel. The CUDA applet in System Preferences never gives us the ability to update the driver, so it thinks that we must be at the latest version. It outputs to the screen correctly and does that part just fine. The problem is that the second entry in the System Information panel will never say that the driver is loaded. We have also installed the toolkit, as well. All of these installations complete without any errors occurring.

Any help appreciated, as we appear to be running out of options quickly.

Additional information:

  • If we put the old GTX 285 back into the machine, everything works correctly again.

You need to install the specific NVDIA driver for Quadro 4000.
It is available at:

After you install this driver and the CUDA driver, everything should work.

I have a similar setup and it works:

/Developer/GPU\ Computing/C/bin/darwin/release/deviceQuery
/Developer/GPU Computing/C/bin/darwin/release/deviceQuery Starting…

CUDA Device Query (Runtime API) version (CUDART static linking)

There is 1 device supporting CUDA

Device 0: “Quadro 4000”
CUDA Driver Version: 3.20
CUDA Runtime Version: 3.20
CUDA Capability Major/Minor version number: 2.0
Total amount of global memory: 2147024896 bytes
Multiprocessors x Cores/MP = Cores: 8 (MP) x 32 (Cores/MP) = 256 (Cores)
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 49152 bytes
Total number of registers available per block: 32768
Warp size: 32
Maximum number of threads per block: 1024
Maximum sizes of each dimension of a block: 1024 x 1024 x 64
Maximum sizes of each dimension of a grid: 65535 x 65535 x 1
Maximum memory pitch: 2147483647 bytes
Texture alignment: 512 bytes
Clock rate: 0.95 GHz
Concurrent copy and execution: Yes
Run time limit on kernels: Yes
Integrated: No
Support host page-locked memory mapping: Yes
Compute mode: Default (multiple host threads can use this device simultaneously)
Concurrent kernel execution: Yes
Device has ECC support enabled: No
Device is using TCC driver mode: No

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 3.20, CUDA Runtime Version = 3.20, NumDevs = 1, Device = Quadro 4000


Press to Quit…

Thank you! We are in business now!

FWIW, this language at http://www.nvidia.com/object/quadro-macosx-256.01.00f03-driver.html is extremely incorrect and should be removed. It was the source of a lot of my confusion:

“Please note that if you are running the Quadro FX 4800 for Mac or the GeForce GTX 285, you do not need to upgrade to this driver”