Error querying the number of OpenCL platforms in the system

[krikru]

After following the instructions as closely as possible (I'm using Python3 and I have an Intel GPU, if you can call it that), when I run the mnist example (by running python3 convolutional.py from the terminal) I get the output

Extracting data/train-images-idx3-ubyte.gz
Extracting data/train-labels-idx1-ubyte.gz
Extracting data/t10k-images-idx3-ubyte.gz
Extracting data/t10k-labels-idx1-ubyte.gz
2017-01-26 22:30:09: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
2017-01-26 22:30:09: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
terminate called after throwing an instance of 'cl::sycl::cl_exception'
  what():  Error: [ComputeCpp:RT0408] Error querying the number of OpenCL platforms in the system
Aborted

Why do I get this error (googling for the error message yields zero hits so I don't know where it is generated) and what can I do to fix it?

[krikru]

I solved it. The problem was that I hadn't installed Intel's OpenCL driver and library packages, only the Intel SDK for OpenCL Applications Package (apparently); my mistake was that I assumed that the SDK would include everything you needed to work with OpenCL. Maybe we can write something about this in the instructions.

[amalic]

can someone close this issue?

[JulyJohn]

@krikru
you said you have an Intel GPU, I was wondering whether it means a Core graphics card Integrated in a CPU, like Intel HD530?

[krikru]

@JulyJohn I'm pretty sure the GPU was integrated on the CPU. It wasn't a very expensive laptop and I don't remember that it should have had a dedicated graphics card (which I think I should if it would have had one).