Ehsan Totoni
Ehsan Totoni
HPAT is based on Numba (our ParallelAccelerator in Numba) but is focused on distributed-memory analytics. Yes, we ported HPAT and ParallelAccelerator from Julia.
HPAT is focused on distributed analytics but leverages Numba to large extent. Intel compiler is not used in Numba (LLVM is the backend). We are working on various aspects of...
We have discussed it a couple of times. Doesn't sound like there is a lot of benefit in doing that.
HPAT needs a specific version of Numba which is available in the channel. I think you are using the regular Numba. You can try `conda install numba -c ehsantn` or...
Sure, I'll see how we can mitigate this. This issue will be resolved since standard Numba will support HPAT sometime soon.
Could you try Python 3.6?
could you paste the output of `conda list`?
`boost` is missing and Numba is also old (`0.39` is needed). Why are you not using our conda install command? `conda create -n HPAT -c ehsantn -c numba -c anaconda...
hpat 0.22 in the channel is built with default Anaconda's Boost, which is 1.65 and is built with GCC 7.2. Anything other than this can fail. If you want to...
1. Yes. 2. HPAT already uses parallel HDF5 (collective I/O). The HDF5 version installed through the channel with HPAT is parallel. Also see: https://github.com/IntelLabs/hpat/blob/master/hpat/_io.cpp#L111 3. Similar to reads - collective...