Problem with run CUDA on GeForce GT 755M

Accelerated Computing CUDA CUDA Programming and Performance
#1

Hello,

I have the same problem with using GeForce GT 755M for machine learning (CAFFA framework)
My platform:
ubuntu 16.04, Cuda V8.0.44, CuDNN

I suggest my version is too new and I have to downgrade.
Could you advise the best way for solve my problem?

Find in following more details:

nvcc is warning about deprecation. But it not error and as I know it is about future.

nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release

CAFFE install with no error, but on testing make runtest I recieve next massage:

Current device name: 
F0109 19:47:58.676647 25551 gpu_memory.cpp:159] Check failed: error == cudaSuccess (30 vs. 0)  unknown error
*** Check failure stack trace: ***
    @     0x2abdf1fd9daa  (unknown)
    @     0x2abdf1fd9ce4  (unknown)
    @     0x2abdf1fd96e6  (unknown)
    @     0x2abdf1fdc687  (unknown)
    @     0x2abdf3fba3e6  caffe::GPUMemory::Manager::update_dev_info()
    @     0x2abdf3fba7d9  caffe::GPUMemory::Manager::init()
    @           0x46673a  main
    @     0x2abdf4fbdf45  (unknown)
    @           0x46e2c9  (unknown)
    @              (nil)  (unknown)

If I disable CUDU in CAFFA Makefile.config, all had tested with no errors.

More outputs:

~/caffe# nvcc --version
nvcc: NVIDIA ® Cuda compiler driver
Copyright © 2005-2016 NVIDIA Corporation
Built on Sun_Sep__4_22:14:01_CDT_2016
Cuda compilation tools, release 8.0, V8.0.44

~/caffe# nvidia-smi
Mon Jan 9 19:51:56 2017
±----------------------------------------------------------------------------+
| NVIDIA-SMI 367.57 Driver Version: 367.57 |
|-------------------------------±---------------------±---------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GT 755M Off | 0000:01:00.0 N/A | N/A |
| N/A 60C P0 N/A / N/A | 823MiB / 2000MiB | N/A Default |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 Not Supported |
±----------------------------------------------------------------------------+

#2
  1. One possibility is that your 367.57 driver is not installed correctly. You should carefully follow the instructions in the linux install guide to verify your CUDA install.

Did you verify your CUDA install?

  1. CUDNN requires a compute capability 3.0+ GPU and I would suggest that you also compile your codes for a compute capability 3.0+ target (even better - match the target of your GT755M gpu). Its possible you are not doing that…

  2. Another possibility is that this is an optimus laptop and you will need to make sure the NVIDIA GPU is enabled when running caffe. There are a few approaches to do this.

http://nvidia.custhelp.com/app/answers/detail/a_id/2615/~/how-do-i-customize-optimus-profiles-and-settings%3F

#3

Hi,
I just have the same problem as you, did you solve it?
can you help me?
I will be really very grateful!!
Hope you reply as soon as possible.Thank u very much.

#4

getting the same error during caffe install

nvcc warning : The ‘compute_20’, ‘sm_20’, and ‘sm_21’ architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning)