January 22, 2020

[Note] Checking GPU use for Theano and Pytorch

Just installed Theano and Pytorch using Anaconda with GPU support.  Here is how to check if it's using GPU.

THEANO

test_gpu.py:

# http://deeplearning.net/software/theano/tutorial/using_gpu.html#testing-theano-with-gpu

from theano import function, config, shared, tensor
import numpy
import time

vlen = 10 * 30 * 768  # 10 x #cores x # threads per core
iters = 1000

rng = numpy.random.RandomState(22)
x = shared(numpy.asarray(rng.rand(vlen), config.floatX))
f = function([], tensor.exp(x))
print(f.maker.fgraph.toposort())
t0 = time.time()
for i in range(iters):
    r = f()
t1 = time.time()
print("Looping %d times took %f seconds" % (iters, t1 - t0))
print("Result is %s" % (r,))
if numpy.any([isinstance(x.op, tensor.Elemwise) and
              ('Gpu' not in type(x.op).__name__)
              for x in f.maker.fgraph.toposort()]):
    print('Used the cpu')
else:
    print('Used the gpu')





PYTORCH

import torch
torch.cuda.current_device()
torch.cuda.device(0)
torch.cuda.device_count()
torch.cuda.get_device_name(0)
torch.cuda.is_available()





REFERENCE
To test later, https://blog.dominodatalab.com/gpu-computing-and-deep-learning/



No comments: