3
To: Sean Owen<mailto:sro...@gmail.com>
Cc: Ludovic Henry<mailto:luhe...@microsoft.com>;
dev@spark.apache.org<mailto:dev@spark.apache.org>; Bernhard
Urban-Forster<mailto:beu...@microsoft.com>
Subject: Re: Usage of JDK Vector API in ML/MLLib
Regarding selective compilation
ntations on my end.
>
>
>
> Thank you,
>
> Ludovic
>
>
>
> [1] https://github.com/luhenry/blas
>
>
>
> *From: *Erik Krogen
> *Sent: *Tuesday, 15 December 2020 17:33
> *To: *Sean Owen
> *Cc: *Ludovic Henry ; dev@spark.apache.org; Bernhard
> Urban-
<mailto:luhe...@microsoft.com>;
dev@spark.apache.org<mailto:dev@spark.apache.org>; Bernhard
Urban-Forster<mailto:beu...@microsoft.com>
Subject: Re: Usage of JDK Vector API in ML/MLLib
Regarding selective compilation, you can hide sources behind a Maven profile
such as `-Pvector
Regarding selective compilation, you can hide sources behind a Maven
profile such as `-Pvectorized`. Check out what we do to switch between the
`hive-1.2` and `hive-2.3` profiles where different source directories are
grabbed at compile-time (the hive-1.2 profile was recently removed so you
might h
Yes it's intriguing, though as you say not readily available in the wild
yet.
I would also expect native BLAS to outperform f2j also, so yeah that's the
interesting question, whether this is a win over native code or not.
I suppose the upside is eventually, we may expect this API to be available
in
Hello,
I’ve, over the past few days, looked into using the new Vector API [1] to
accelerate some BLAS operations straight from Java. You can find a gist at [2]
containing most of the changes in
mllib-local/src/main/scala/org/apache/spark/ml/linalg/BLAS.scala.
To measure performance, I’ve a
