It's quite risky to switch right before the show. Xiao-Feng, what workloads you tried with gcv5 except specs and eclipse?
+ I'm working on "lazy resolution" task in both JITs now and going to submit the patch this week for review. I think its commit should be delayed for a few weeks until JavaOne is finished. On 4/18/07, Xiao-Feng Li <[EMAIL PROTECTED]> wrote:
On 4/18/07, Mikhail Loenko <[EMAIL PROTECTED]> wrote: > Do you think we should switch before "end of month"? Yes, that's my suggestion. > It's certainly a risk, but what is the value in the switch? The risk is minimal since GCv5 is rather stable, and to the least we have command line option to switch back; but the value is substantial since people can have an advanced, scalable, modular, flexible, high performance GC, which I think both runtime researchers and users would like to try, based on my interactions with Harmony users. To demo Harmony, GC is one component that we'd like to have a good story to tell. GCv5 can tell a good story since it has subsumed almore all the recent advances in GC area (for stop-the-world GC), and has a variable of innovations. Importantly, GCv5 can differentiate multi-core platforms with its scalable parallelisms. :-) Thanks, xiaofeng > Thanks, > Mikhail > > 2007/4/18, Xiao-Feng Li <[EMAIL PROTECTED]>: > > GCv5 might be one "major" that we want to put as default GC in DRLVM. > > It still has some issues pending, but overall I think the stability is > > good enough for a switch next week. > > > > Since GC is designed with good modularity, we can simply choose which > > GC implementation to use in command line with > > '-XX:vm.dlls=the_gc_module.dll(so)". This is neat that helps the > > switch a lot: If GCv5 has some problem running a workload, we can > > specify -XX:vm.dlls=gc_cc.dll in command line. > > > > So far the known bugs in GCv5 are not with some workloads, but related > > with certain test cases for finalizer and VM threading. And I think > > they are going to be resolved before next week. > > > > Thanks, > > xiaofeng > > > > On 4/16/07, Tim Ellison <[EMAIL PROTECTED]> wrote: > > > Just a reminder, as discussed in various threads, we shall aim to > > > produce a solid build for Windows and Linux x86 (at least) at the end of > > > next week; so that we have something to demo at ApacheCon and JavaOne > > > that is a true reflection of our current capabilities. > > > > > > Of course, the Milestone will be simply a snapshot, carrying our usual > > > caveats. The idea is that with conference talks taking place we may > > > expect a few people to download a build and try it around that time, so > > > being in the middle of a major restructuring would potentially do us an > > > injustice. > > > > > > Most commits still seem to be on-going bug fixing, so that's all > > > goodness. If you are planning on anything 'major' please ensure there > > > is enough time to get it stable, or please wait until after the > > > milestone build. Similarly, if there is anything that is currently > > > 'broken' that you think really needs fixing for that stability, please > > > shout here on the list. > > > > > > There are still two weeks to go, I think the paranoia about not causing > > > regressions will really kick-in next week :-) > > > > > > Regards, > > > Tim > > > > > > > > > -- > > http://xiao-feng.blogspot.com > > > -- http://xiao-feng.blogspot.com
-- Mikhail Fursov
