zhengruifeng commented on pull request #28229:
URL: https://github.com/apache/spark/pull/28229#issuecomment-747989608
@xwu99 Sorry for the late reply.
We had just make SVC/LiR/LoR/AFT using blocks instead of instances in 3.1.0,
but in a new adaptive way to blockify instances
(https
zhengruifeng commented on pull request #28229:
URL: https://github.com/apache/spark/pull/28229#issuecomment-636473457
@xwu99 Thanks for your work. The speedup is promising.
Since this issue (blockify+gemv/gemm) need more discussion with other
committers, so I am working on retest th
zhengruifeng commented on pull request #28229:
URL: https://github.com/apache/spark/pull/28229#issuecomment-624991855
@xwu99 I think you can also refer to those two PRs, since some utils were
added.
This is an automated mess
zhengruifeng commented on pull request #28229:
URL: https://github.com/apache/spark/pull/28229#issuecomment-624989291
@xwu99 There was a
[ticket](https://issues.apache.org/jira/browse/SPARK-30641) for this.
Now I had merged high-level BLAS supports for
[LinearSVC](https://github.com/ap
zhengruifeng commented on pull request #28229:
URL: https://github.com/apache/spark/pull/28229#issuecomment-620990814
@xwu99 My previous works include:
LinearSVC: https://github.com/apache/spark/pull/27360
LogisticRegression: https://github.com/apache/spark/pull/27374
LinearRegress
zhengruifeng commented on pull request #28229:
URL: https://github.com/apache/spark/pull/28229#issuecomment-620975624
> I saw your PR was merged, I will rebase.
I had some reverted PRs on using high-level BLAS in LoR/LiR/SVC/GMM, they
were reverted because of performance regression
