[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Updates about BLAS64



Le dimanche 16 décembre 2018 à 01:51 +0000, Mo Zhou a écrit :

> Some time ago we discussed about packages for BLAS64 ABI/API.
> Here are some updates about the subsequent works:
> 
> * blas64 and lapack64 symlinks and alternatives have been added to intel-mkl.
> 
>    src:intel-mkl
>      bin:libmkl-rt
>        enhances: libblas.so.3, libblas64.so.3, liblapack.so.3, liblapack64.so.3
>        alternative: libblas64.so.3 -> libmkl_rt.so
>                     liblapack64.so.3 -> libmkl_rt.so
>                     ...
>      bin:libmkl-dev
>        enhances: libblas64.so, liblapack64.so, ...
>        alternative: libblas64.so -> libmkl_rt.so
>                     liblapack64.so -> libmkl_rt.so
> 
> * A new BLAS implementation called BLIS[1] is nearly ready to upload.
>   This is likely the first package that will provide 64bit BLAS API/ABI
>   in main section of our archive.
> 
>   BLIS supports three different threading models, just like openblas.
>   The difference is that, BLIS will be compiled 6 times with different
>   configurations:
> 
>   (openmp, pthread, serial) x (32-bit, 64-bit) = 6 combinations
> 
>   BLIS shared libraries with different threading models and the same
>   index length cannot coexist. Any 32-bit lib can co-exist with any
>   64-bit one.
> 
>    src:blis
>      bin:libblis1 (meta)
>        deps: libblis1-openmp | libblis1-pthread | libblis1-serial
>        provides: libblas.so.3
>      bin:libblis64-1 (meta)
>        deps: libblis64-1-openmp | ...
>        provides: libblas64.so.3
>      ...
> 
>   Note that BLIS doesn't provide LAPACK implementation.
> 
>   @Sebastian: Does this look good to OpenBLAS?

I don't understand what your actual question is. Could you be more
specific?

-- 
⢀⣴⠾⠻⢶⣦⠀  Sébastien Villemot
⣾⠁⢠⠒⠀⣿⡁  Debian Developer
⢿⡄⠘⠷⠚⠋⠀  http://sebastien.villemot.name
⠈⠳⣄⠀⠀⠀⠀  http://www.debian.org

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: