Ehsan Totoni
Ehsan Totoni
I think we need deeper performance analysis (with VTune?) to find out what the problem is.
Yes, this is interesting for both ParallelAccelerator and HPAT. We have already looked at ALS to some extent for HPAT but we need sparse matrix support first.
I investigated supporting Windows a while ago. The issue I couldn't solve was that ParallelAccelerator couldn't load the shared library file.
I think we can do the same think we did with GEMM; use MKL or LAPACK if available, otherwise use a naive sequential C code. [dgetrf of LAPACK](http://www.netlib.org/lapack/double) and [MKL](https://software.intel.com/en-us/node/520877)...
> > This inconsistency causes issues for unboxing the values for Bodo JIT > > Can you elaborate on this? The code paths for unboxing each array type is different....
> Thanks for the report. I don't think the API for extension arrays is very clear. Did you try to simply remove the subclass [here](https://github.com/pandas-dev/pandas/blob/main/pandas/core/arrays/string_.py#L220) and see what fails? I...
Looks like when parallel accelerator is used. Updated the issue.
Can we check the reflected flags and allow non-reflected cases?
Having Debug.jl on v0.5 would be great. Gallium crashes all the time.
@dlee992 have you considered [Bodo JIT](https://bodo.ai/community-edition/) for compiling Pandas dataframes in your system? It is based on Numba, and is available on Pip and Conda (community edition is free up...