Install Cuda Toolkit for GeForce 304.135 on Ubuntu 14.04

Hi,

I have been trying to install Cuda Driver & Cuda Toolkit on Ubuntu 14.04. I have downgraded from 16.04 as I had so many more bugs and problems.

I have tried to install from .run file, but it still says no kernel found for driver 304.29

I have downloaded cuda tool kit debian for 5 , 5.5, 6.5 and installed as follows:

$ sudo dpkg -i cuda-xxxxx.deb
$ sudo apt update
$ sudo apt upgrade
$ sudo apt install cuda 

I doubt the last line of install cuda to be causing some discrepancies, but it seems to be my only choice.

Then I get Infinite Loenter code heregin Loop which I solve by :

$ sudo apt install nvidia-304

I login and run sudo make clean & sudo make or device query sample and I make sure I have the PATH with CUDA_HOME

$ nvidia-smi 

Sun Apr 1 05:34:24 2018
±-----------------------------------------------------+
| NVIDIA-SMI 4.304… Driver Version: 304.135 |
|-------------------------------±---------------------±---------------------+
| GPU Name | Bus-Id Disp. | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce 8800 GT | 0000:02:00.0 N/A | N/A |
| 30% 56C N/A N/A / N/A | 56% 284MB / 511MB | N/A Default |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Compute processes: GPU Memory |
| GPU PID Process name Usage |
|=============================================================================|
| 0 Not Supported |
±----------------------------------------------------------------------------+

$ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2015 NVIDIA Corporation
Built on Mon_Feb_16_22:59:02_CST_2015
Cuda compilation tools, release 7.0, V7.0.27




$ ./deviceQuery 
./deviceQuery Starting...

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

cudaGetDeviceCount returned 35
-> CUDA driver version is insufficient for CUDA runtime version
Result = FAIL

I have been in this error for ages, please help

This is a very old GPU.

Leave your 304.135 driver in-place. Install CUDA 5.0 toolkit

That should work correctly - run the deviceQuery app from CUDA 5.0.

cross posting:

[url]https://stackoverflow.com/questions/49601430/install-cuda-toolkit-for-geforce-304-135-on-ubuntu-14-04[/url]

Thank you, I have decided to purchase a new GPU card, could you please advise according to my hardware specs:

Mac Pro Server Model 2008
RAM 22 Gbyte
Processor 3.2 Quadcore

More can be found here:

I run on Ubuntu

I need to use Cuda 8 & CudNN 7 & TensorFlow & PyTorch with GPU

Which GPU Nvidia Geforce is compatible with Ubuntu 16.04, Cuda 8.0 and MacPro Apple Server Early 2008 Mac Pro (Early 2008) - Technical Specifications ? for machine learning with Python 3.5, TensorFlow 1.5 & PyTorch v0.3.0 ?