Umar Arshad
Umar Arshad
I was able to reproduce this issue and it seems to be a problem with the way we cache kernels. The kernel cache is there to avoid compiling the OpenCL...
Use -DUSE_CPU_MKL=ON and -DUSE_OPENCL_MKL=ON when running CMake
This is expected. That warning can safely be ignored according to the warning documentation when using the standard containers. https://docs.microsoft.com/en-us/cpp/error-messages/compiler-warnings/compiler-warning-level-2-c4275?view=vs-2019 > C4275 can be ignored in Visual C++ if you...
We should add a pragma for msvc to the header so that this warning doesn't show up in user code. https://docs.microsoft.com/en-us/cpp/preprocessor/warning?view=vs-2019
The f16 datatype is currently in the master branch. We haven't made a release with 16-bit floating point as of now. You will have to build ArrayFire to use it.
Yes, Updated the list
Hey Alex, Thanks for the issue. I can reproduce this on my end. I need to look into this further to figure out what is going on. Will keep you...
This can be done easily using indexing, and the where function: ```c++ #include using namespace af; int main() { array a = randu(10).T(); array bits = randu(10, b8).T(); af_print(a); af_print(bits);...
I am not sure if I understand. We are performing a scan operation to generate these results. Can you show me how you are benchmarking the CPU and GPU operations?
Here is what I see on my system: ```cpp #include #include using namespace af; template void my_timer(FN fun) { for(int i = 0; i < 10; i++) { auto start...