CODEWITHSUNDEEP
pythontensorflowgpu
alyaa mahdi
alyaa mahdi

Reputation: 29

Cuda compute capability 3.0. The minimum required Cuda capability is 3.7

I have python 3.5, gpu: Quadro K1000M, CUDA 9.0 and I want toinstall tensorflow on gpu. The installation was done successfully, but when I check it, it gives me the warning and my python code works with CPU version in spite of I didn't install CPU version. The warning is:

     Cuda compute capability 3.0. The minimum required Cuda capability is 3.7

When I check it in python as follow:

    sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
    sess.close()

It return me 

    no known devices

Can anyone please help me what I do? I'm new in python, any hint may be useful. Thanks

Upvotes: 0

Views: 2316

Answers (1)

GPhilo
GPhilo

Reputation: 19153

Cuda compute capability 3.0. The minimum required Cuda capability is 3.7

Your GPU is too old, the bare minimum according to the list of GPU models on Wikipedia's CUDA page is a Tesla K80. If you want touse TF on the GPU you'll need to upgrade the GPU first.

Upvotes: 1

Related Questions