Discover MakerZone

MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi

Learn more

Discover what MATLAB® can do for your career.

Opportunities for recent engineering grads.

Apply Today

Thread Subject:
CUDA Errors - All CUDA-capable devices are busy

Subject: CUDA Errors - All CUDA-capable devices are busy

From: Richard Biehl

Date: 9 Mar, 2012 02:27:34

Message: 1 of 2

I am having a problem getting the Matlab Parallel Computing toolbox to work with my Tesla M2070. When I submit the gpuDevice I get the following error:

Error using gpuDevice (line 22)
An unexpected error occurred during CUDA execution. The CUDA error was: all
CUDA-capable devices are busy or unavailable.

I get the following information from the nVidia SMI utility and I am using the card in TCC mode. I have it installed on a Windows Server 2008 R2 server with MatLab R2011b. The latest nVidia drivers are installed. The display adapter is a Matrox card and the Tesla is for GPU computing only. Any thoughts??

Richard

==============NVSMI LOG==============

Timestamp : Thu Mar 08 20:16:54 2012

Driver Version : 295.73

Attached GPUs : 1

GPU 0000:08:00.0
    Product Name : Tesla M2070
    Display Mode : Disabled
    Persistence Mode : N/A
    Driver Model
        Current : TCC
        Pending : TCC
    Serial Number : 0323710045651
    GPU UUID : GPU-896b9442dd076af3-ac3a92e7-8157433c-3d0de69
9-af9ee5af7da574b1866fa0d5
    VBIOS Version : 70.00.3E.00.03
    Inforom Version
        OEM Object : 1.0
        ECC Object : 1.0
        Power Management Object : 1.0
    PCI
        Bus : 0x08
        Device : 0x00
        Domain : 0x0000
        Device Id : 0x06D210DE
        Bus Id : 0000:08:00.0
        Sub System Id : 0x083010DE
        GPU Link Info
            PCIe Generation
                Max : 2
                Current : 2
            Link Width
                Max : 16x
                Current : 16x
    Fan Speed : N/A
    Performance State : P8
    Memory Usage

Subject: CUDA Errors - All CUDA-capable devices are busy

From: Ben Tordoff

Date: 12 Mar, 2012 09:00:13

Message: 2 of 2

Hi Richard,

This message appears either when no device is found or when the default device is locked out of use. In order to diagnose the problem, can you try the following and post the results:

>> gpuDeviceCount()
>> gpu = parallel.gpu.GPUDevice.getDevice(1)

Note that this latter call asks for information about the device without trying to select it, which should help in working out whether finding the device or selecting the device is the problem. Hopefully you will see the details for your card with "DeviceSupported" set to 1 and "ComputeMode" set to "Default".

Cheers

Ben


"Richard Biehl" wrote in message <jjbpqm$899$1@newscl01ah.mathworks.com>...
> I am having a problem getting the Matlab Parallel Computing toolbox to work with my Tesla M2070. When I submit the gpuDevice I get the following error:
>
> Error using gpuDevice (line 22)
> An unexpected error occurred during CUDA execution. The CUDA error was: all
> CUDA-capable devices are busy or unavailable.

[snip]

Tags for this Thread

No tags are associated with this thread.

What are tags?

A tag is like a keyword or category label associated with each thread. Tags make it easier for you to find threads of interest.

Anyone can tag a thread. Tags are public and visible to everyone.

Contact us