After multiple runs of our perftest suite over 80MB-800GB with Spark 2.1 and 2.2, with and without codegen, as well as with and without compression, we found and fixed quite a number of issues. However, there are a couple of remaining performance issues with sparse maxpooling operations, so I would recommend to defer the RC1 for a couple of days until these issues are fixed as well.
Regards, Matthias On Tue, Nov 14, 2017 at 1:31 PM, Krishna Kalyan <[email protected]> wrote: > +1 > > Regards, > Krishna > > On Sun, Nov 12, 2017 at 5:53 AM, Matthias Boehm <[email protected]> wrote: > > > just FYI: I've created SYSTEMML-2011 to track all QA-related tasks and > > issues. The first run of our (old) perftest suite for 80MB-80GB ran fine > > except for a data generation issue on stratstats for the 80GB case. Btw, > I > > think it would be good to run both the old and new perftest for > additional > > validation. > > > > Regards, > > Matthias > > > > On Wed, Nov 8, 2017 at 8:30 AM, Glenn Weidner <[email protected]> > wrote: > > > > > +1 > > > > > > Regards, > > > Glenn > > > > > > [image: Inactive hide details for "Niketan Pansare" ---11/08/2017 > > 06:51:44 > > > AM---+1. Thanks,]"Niketan Pansare" ---11/08/2017 06:51:44 AM---+1. > > Thanks, > > > > > > From: "Niketan Pansare" <[email protected]> > > > To: [email protected] > > > Date: 11/08/2017 06:51 AM > > > Subject: Re: SystemML 1.0 release timeline > > > ------------------------------ > > > > > > > > > > > > > > > +1. > > > > > > Thanks, > > > > > > Niketan. > > > > > > > On Nov 7, 2017, at 10:50 PM, Matthias Boehm <[email protected]> > > > wrote: > > > > > > > > Hi all, > > > > > > > > we made some good progress regarding deep learning support, code > > > > generation, and low-latency scoring - so, I'm looking forward to our > > > > upcoming 1.0 release. Since it's our first stable release, I think it > > > would > > > > be a good idea to allocate some extra time for QA. How about we shoot > > for > > > a > > > > release candidate Dec 1? > > > > > > > > Regards, > > > > Matthias > > > > > > > > > > > > > > >
