Hardware question

Igor_G

New Member
May 14, 2020
1
0
1
I am going to start a course in machine learning soon
I have two cards of nvidia gtx titan purchased in a far away 2013
(pretty old, but all new games still run quite well, sorry for a poor assessment way:))
We are suggested to work on 2 platforms for our choice: matlab (deep learning toolbox), or python Keras(based on tensorflow as far as I understood).
-------
my PC specs are:
CPU: Intel (R) CoretM i9-9900k @ 3.6 gherz
ram 32 GB

---------
Any ideas about are those 2 titans going to be more useful than just using CPU , and if yes, can I use them on python and on matlab or they work only with one of them?
 
Last edited:

Agremlin

New Member
May 3, 2020
11
2
3
They appear to be Kepler architecture GeForce 700 series - Wikipedia. Will be pretty slow but should be able to install drivers (AWS still support K80's for machine learning). Will still be a lot faster than cpu's. Should work fine with either matlab (though i have no experience with it) or keras/tensorflow. The only minor hiccup will be installing cuda drivers
 

L.P.

New Member
Feb 17, 2019
3
0
1
As @Agremlin has already mentioned, your GPUs will still be faster than the 9900k.

There is only one major difficulty. You need to find a way to get the ML-frameworks, which support your older version of CUDA. Keras can work with Tensorflow, Theano or CNTK, you just need to configure the backend. If you know your way around Bazel you can build your own Tensorflow package optimized for your CUDA version and CPU, which would give you the best performance.