adding AVXCopyRelu to amd64 - i386 is still missing
Fixing dot product OpenCL test.
New type TNeuralFloatPtr.
Pretty minor optimizations.
Super minor convolution backprop optimization.
Small convolution backprop optimization.
Speeding up CPU and OpenCL convolution.
Optimizing OpenCL and interleaved execution.
Speeding up fully connected layers.
Minor full connect optimizations.