<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><div><blockquote type="cite" class=""><div class=""><font color="#000000" class="">Thanks Eric,</font><br class=""></div></blockquote><div><br class=""></div><div><br class=""></div>These RFCs looks very good. The RFC on throughput barriers would be the highest priority item. My discussions with low latency people suggest that this will be a huge barrier to them migrating to 9. I would suggest that a parallel full is the least important on the list as I would hope we can find strategies to tune them out of systems so that one would only see them in (pseudo-) failure conditions that are often a result of “work queues†backing up and thus are outside the realm of GC tuning.</div><div><br class=""></div><div>I look forward to testing them.</div><div><br class=""></div><div>Kind regards,</div><div>Kirk</div><div><br class=""><blockquote type="cite" class=""><div class=""><div class="gmail_extra"><br class=""><div class="gmail_quote">On Wed, Nov 9, 2016 at 7:26 AM, Erik Helin <span dir="ltr" class=""><<a href="mailto:erik.helin@oracle.com" target="_blank" class="">erik.helin@oracle.com</a>></span> wrote:<br class=""><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Jeremy,<br class="">
<br class="">
Thanks for taking notes. I have created separate email threads on <a href="mailto:hotspot-gc-dev@openjdk.java.net" target="_blank" class="">hotspot-gc-dev@openjdk.java.ne<wbr class="">t</a> for three enhancements to G1 that we think will make G1 be more like CMS for the use cases discussed during the meeting. The ideas are described in:<br class="">
- "RFC: Throughput barriers for G1"<br class="">
<br class="">
<a href="http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/2016-November/019215.html" rel="noreferrer" target="_blank" class="">http://mail.openjdk.java.net/p<wbr class="">ipermail/hotspot-gc-dev/2016-N<wbr class="">ovember/019215.html</a><br class="">
- "RFC: Parallel full collection for G1"<br class="">
<br class="">
<a href="http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/2016-November/019216.html" rel="noreferrer" target="_blank" class="">http://mail.openjdk.java.net/p<wbr class="">ipermail/hotspot-gc-dev/2016-N<wbr class="">ovember/019216.html</a><br class="">
- "RFC: Rebuilding remembered sets during concurrent mark"<br class="">
<br class="">
<a href="http://mail.openjdk.java.net/pipermail/hotspot-gc-dev/2016-November/019217.html" rel="noreferrer" target="_blank" class="">http://mail.openjdk.java.net/p<wbr class="">ipermail/hotspot-gc-dev/2016-N<wbr class="">ovember/019217.html</a><br class="">
<br class="">
Please respond to these ideas in the respective email thread. We would be happy to help out with the design and integration if you and/or any external contributor would like too look into any (or all) of these ideas.<br class="">
<br class="">
Thanks,<br class="">
Erik<span class=""><br class="">
<br class="">
On 10/21/2016 08:37 PM, Jeremy Manson wrote:<br class="">
</span><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">
Hi folks,<br class="">
<br class="">
We had the face to face at JavaOne on 9/20. I took notes, which are<br class="">
linked from the bug:<br class="">
<br class="">
<a href="https://bugs.openjdk.java.net/browse/JDK-8142518" rel="noreferrer" target="_blank" class="">https://bugs.openjdk.java.net/<wbr class="">browse/JDK-8142518</a><br class="">
<<a href="https://bugs.openjdk.java.net/browse/JDK-8142518" rel="noreferrer" target="_blank" class="">https://bugs.openjdk.java.net<wbr class="">/browse/JDK-8142518</a>><br class="">
<br class="">
(Sorry about the delay - anything that happens at a quarterly boundary<br class="">
takes me twice as long because of the Google planning cycle).<br class="">
<br class="">
The current plan for followup includes Erik Helin figuring out what<br class="">
extracting CMS code from Hotspot will actually look like, and Paul Su<br class="">
providing a list of ideas for helping G1 close the performance gap. See<br class="">
meeting notes for more followup actions.<br class="">
<br class="">
We didn't specify a timeline for the next meeting. Paul (cc'd) said he<br class="">
had to go do some due diligence about getting the list of ideas. It<br class="">
would be great if he and / or Erik could talk about when a good time for<br class="">
the next sync would be.<br class="">
<br class="">
Jeremy<br class="">
<br class="">
On Sat, Aug 27, 2016 at 5:42 PM, Jeremy Manson <<a href="mailto:jeremymanson@google.com" target="_blank" class="">jeremymanson@google.com</a><br class=""></span><div class=""><div class="h5">
<mailto:<a href="mailto:jeremymanson@google.com" target="_blank" class="">jeremymanson@google.co<wbr class="">m</a>>> wrote:<br class="">
<br class="">
Thanks, Jon.<br class="">
<br class="">
If we are going to do it, I'd (selfishly) like to see it done<br class="">
earlier in the JDK 10 time frame, so that we can start contributing<br class="">
our patches for it for JDK 10 earlier, rather than later. But JDK<br class="">
10 time frame is fine.<br class="">
<br class="">
Assuming you are asking me about management issues: I just<br class="">
(Thursday) got back from the vacation I mentioned, so I've only just<br class="">
spoken about it with our management. Naturally, no immediate<br class="">
answer, but there's a lot of interest, and we plan to have more<br class="">
conversations in the near term.<br class="">
<br class="">
For a variety of reasons (most of which I've already articulated), I<br class="">
strongly believe we will have to do *something*, and it will mostly<br class="">
be a question of whether this is the right path. Since we have<br class="">
~three years before the issue becomes immediate, that gives us a bit<br class="">
of breathing room to do the right thing. Other possibilities<br class="">
include spending the time until JDK 10 making G1 do what we need, or<br class="">
figuring out another GC entirely (we'd want to evaluate, e.g.,<br class="">
Shenandoah).<br class="">
<br class="">
(Since the G1 possibility is clearly of interest to you folks: We<br class="">
would have to decide a) that it is technically feasible, and b) that<br class="">
you folks were likely to take the patches. In addition to finding a<br class="">
path forward to lowering the cost of the write barrier dramatically<br class="">
and dealing with the memory footprint issues (not sure this is<br class="">
possible), we would also have to reimplement, e.g., our parallel<br class="">
Full GC patch, as well as a number of other changes we've made to<br class="">
CMS locally.)<br class="">
<br class="">
I'd like what we end up doing to be something palatable to the<br class="">
community, since CMS is obviously filling a very important role in<br class="">
the ecosystem that won't get filled if it gets abandoned.<br class="">
<br class="">
Jeremy<br class="">
<br class="">
On Tue, Aug 23, 2016 at 9:23 PM, Jon Masamitsu<br class=""></div></div><div class=""><div class="h5">
<<a href="mailto:jon.masamitsu@oracle.com" target="_blank" class="">jon.masamitsu@oracle.com</a> <mailto:<a href="mailto:jon.masamitsu@oracle.com" target="_blank" class="">jon.masamitsu@oracle.c<wbr class="">om</a>>> wrote:<br class="">
<br class="">
All,<br class="">
<br class="">
I've been working on the examples and we're been discussing<br class="">
(arguing)<br class="">
about them here. They are failing into at least the following<br class="">
classes<br class="">
<br class="">
- Use inheritance to specialize for CMS (specialize card table is<br class="">
envisioned as an example).<br class="">
- Use a GC Interface to request something different for CMS<br class="">
(different write<br class="">
barriers for example).<br class="">
- Compile code under macros<br class="">
Similar to INCLUDE_ALL_GCS<br class="">
Leaves calls to specialized CMS code visible in the shared code<br class="">
Involves some code refactoring<br class="">
- Move more code to the cms directory (ParNew for example)<br class="">
- Custom solutions which might use one or more of the above<br class="">
techniques.<br class="">
- Argument processing I don't know about yet.<br class="">
<br class="">
We're still arguing about the example so they're not here.<br class="">
<br class="">
As I've said Oracle would not drop support of CMS until at least<br class="">
jdk 10 so working on separating out the code would be work<br class="">
for the jdk 10 time frame. Do people have reasons to get that<br class="">
done earlier rather than later?<br class="">
<br class="">
Also is your management's support for spending engineering time<br class="">
on this effort a slam-dunk? Or a qualified maybe? I'm trying to<br class="">
get a feel for how real this CMS project is.<br class="">
<br class="">
Jon<br class="">
<br class="">
<br class="">
<br class="">
On 8/1/2016 1:49 PM, Jeremy Manson wrote:<br class="">
</div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class=""><div class="h5">
For those following along at home, the meeting happened last<br class="">
week. We took notes, which are linked from the bug:<br class="">
<br class="">
<a href="https://bugs.openjdk.java.net/browse/JDK-8142518" rel="noreferrer" target="_blank" class="">https://bugs.openjdk.java.net/<wbr class="">browse/JDK-8142518</a><br class="">
<<a href="https://bugs.openjdk.java.net/browse/JDK-8142518" rel="noreferrer" target="_blank" class="">https://bugs.openjdk.java.net<wbr class="">/browse/JDK-8142518</a>><br class="">
<br class="">
The current plan for followup is to set up a meeting for the<br class="">
end of August, and then a f2f during JavaOne.<br class="">
<br class="">
Jeremy<br class="">
<br class="">
On Thu, Jul 21, 2016 at 9:46 AM, Jeremy Manson<br class=""></div></div><span class="">
<<a href="mailto:jeremymanson@google.com" target="_blank" class="">jeremymanson@google.com</a> <mailto:<a href="mailto:jeremymanson@google.com" target="_blank" class="">jeremymanson@google.co<wbr class="">m</a>>> wrote:<br class="">
<br class="">
I was drafting an agenda. My expectation is that the<br class="">
meeting will go something like this:<br class="">
<br class="">
Introductions<br class="">
<br class="">
Discussion of motivations for JEP 291 (Jon M)<br class="">
<br class="">
Discussions of concerns about JEP 291 (roundtable)<br class="">
<br class="">
Discussion / Brainstorming about potential ways forward,<br class="">
as well as an understanding of what level of commitment<br class="">
organizations are willing to make.<br class="">
<br class="">
Discussion of action items and followup.<br class="">
<br class="">
Jeremy<br class="">
<br class="">
<br class="">
On Thu, Jul 21, 2016 at 7:39 AM, Martijn Verburg<br class="">
<<a href="mailto:martijnverburg@gmail.com" target="_blank" class="">martijnverburg@gmail.com</a><br class=""></span><span class="">
<mailto:<a href="mailto:martijnverburg@gmail.com" target="_blank" class="">martijnverburg@gmail.c<wbr class="">om</a>>> wrote:<br class="">
<br class="">
Hi Jon,<br class="">
<br class="">
For me it's to see if there are enough folks who are<br class="">
willing *and* capable (to be blunt, it will need the<br class="">
backing of large companies) of maintaining CMS as a<br class="">
collector going forwards (it's important to several of<br class="">
our customers). If that's a yes then I'd hope to have<br class="">
a technical discussion around the options of how we<br class="">
could achieve that without causing a major headache<br class="">
for the Oracle GC engineers.<br class="">
<br class="">
<br class="">
Cheers,<br class="">
Martijn<br class="">
<br class="">
On 20 July 2016 at 23:50, Jon Masamitsu<br class="">
<<a href="mailto:jon.masamitsu@oracle.com" target="_blank" class="">jon.masamitsu@oracle.com</a><br class=""></span><span class="">
<mailto:<a href="mailto:jon.masamitsu@oracle.com" target="_blank" class="">jon.masamitsu@oracle.c<wbr class="">om</a>>> wrote:<br class="">
<br class="">
What are peoples expectations for this meeting?<br class="">
<br class="">
Jon<br class="">
<br class="">
On 07/18/2016 10:16 AM, Jeremy Manson wrote:<br class="">
</span><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">
Presumably, everyone will be happy with meeting<br class="">
notes rather than an actual recording.<br class="">
<br class="">
Otherwise, I'll add:<br class="">
<br class=""></span>
<a href="mailto:ysr1729@gmail.com" target="_blank" class="">ysr1729@gmail.com</a> <mailto:<a href="mailto:ysr1729@gmail.com" target="_blank" class="">ysr1729@gmail.com</a>><br class="">
<a href="mailto:mark.reinhold@oracle.com" target="_blank" class="">mark.reinhold@oracle.com</a><br class="">
<mailto:<a href="mailto:mark.reinhold@oracle.com" target="_blank" class="">mark.reinhold@oracle.c<wbr class="">om</a>><span class=""><br class="">
<br class="">
<br class="">
<br class="">
On Mon, Jul 18, 2016 at 8:13 AM,<br class="">
<a href="mailto:kirk.pepperdine@gmail.com" target="_blank" class="">kirk.pepperdine@gmail.com</a><br class=""></span>
<mailto:<a href="mailto:kirk.pepperdine@gmail.com" target="_blank" class="">kirk.pepperdine@gmail.<wbr class="">com</a>><br class="">
<<a href="mailto:kirk.pepperdine@gmail.com" target="_blank" class="">kirk.pepperdine@gmail.com</a><span class=""><br class="">
<mailto:<a href="mailto:kirk.pepperdine@gmail.com" target="_blank" class="">kirk.pepperdine@gmail.<wbr class="">com</a>>> wrote:<br class="">
<br class="">
ok we can record if everyone is open to that.<br class="">
<br class="">
</span><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">
On Jul 18, 2016, at 7:48 AM, Jeremy Manson<br class="">
<<a href="mailto:jeremymanson@google.com" target="_blank" class="">jeremymanson@google.com</a><br class=""></span><span class="">
<mailto:<a href="mailto:jeremymanson@google.com" target="_blank" class="">jeremymanson@google.co<wbr class="">m</a>>> wrote:<br class="">
<br class="">
Richard - I'll bring someone who can take<br class="">
readable notes. :)<br class="">
<br class="">
Jeremy<br class="">
<br class="">
On Sat, Jul 16, 2016 at 5:07 AM,<br class="">
<a href="mailto:kirk.pepperdine@gmail.com" target="_blank" class="">kirk.pepperdine@gmail.com</a><br class=""></span>
<mailto:<a href="mailto:kirk.pepperdine@gmail.com" target="_blank" class="">kirk.pepperdine@gmail.<wbr class="">com</a>><br class="">
<<a href="mailto:kirk.pepperdine@gmail.com" target="_blank" class="">kirk.pepperdine@gmail.com</a><span class=""><br class="">
<mailto:<a href="mailto:kirk.pepperdine@gmail.com" target="_blank" class="">kirk.pepperdine@gmail.<wbr class="">com</a>>> wrote:<br class="">
<br class="">
Hi Richard,<br class="">
<br class="">
I think it’s in everyone’s interest to<br class="">
keep this out in the open.<br class="">
<br class="">
Kind regards,<br class="">
Kirk<br class="">
<br class="">
</span><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">
On Jul 16, 2016, at 1:34 AM, Richard<br class="">
Warburton <<a href="mailto:richard.warburton@gmail.com" target="_blank" class="">richard.warburton@gmail.com</a><br class=""></span>
<mailto:<a href="mailto:richard.warburton@gmail.com" target="_blank" class="">richard.warburton@gmai<wbr class="">l.com</a>>><span class=""><br class="">
wrote:<br class="">
<br class="">
Hi,<br class="">
<br class="">
Sounds as if the consensus is a<br class="">
telephone / video conference in the<br class="">
near term, and then a F2F during /<br class="">
close to JavaOne.<br class="">
<br class="">
People who want to be invited include:<br class="">
<br class="">
<a href="mailto:kirk@kodewerk.com" target="_blank" class="">kirk@kodewerk.com</a><br class=""></span>
<mailto:<a href="mailto:kirk@kodewerk.com" target="_blank" class="">kirk@kodewerk.com</a>><br class="">
<a href="mailto:aph@redhat.com" target="_blank" class="">aph@redhat.com</a> <mailto:<a href="mailto:aph@redhat.com" target="_blank" class="">aph@redhat.com</a>><br class="">
<a href="mailto:volker.simonis@gmail.com" target="_blank" class="">volker.simonis@gmail.com</a><br class="">
<mailto:<a href="mailto:volker.simonis@gmail.com" target="_blank" class="">volker.simonis@gmail.c<wbr class="">om</a>><br class="">
<a href="mailto:jon.masamitsu@oracle.com" target="_blank" class="">jon.masamitsu@oracle.com</a><br class="">
<mailto:<a href="mailto:jon.masamitsu@oracle.com" target="_blank" class="">jon.masamitsu@oracle.c<wbr class="">om</a>><br class="">
<a href="mailto:jeremymanson@google.com" target="_blank" class="">jeremymanson@google.com</a><br class="">
<mailto:<a href="mailto:jeremymanson@google.com" target="_blank" class="">jeremymanson@google.co<wbr class="">m</a>><br class="">
<a href="mailto:martijnverburg@gmail.com" target="_blank" class="">martijnverburg@gmail.com</a><br class="">
<mailto:<a href="mailto:martijnverburg@gmail.com" target="_blank" class="">martijnverburg@gmail.c<wbr class="">om</a>><br class="">
<a href="mailto:jwha@google.com" target="_blank" class="">jwha@google.com</a><br class="">
<mailto:<a href="mailto:jwha@google.com" target="_blank" class="">jwha@google.com</a>><span class=""><br class="">
<br class="">
Did I miss anyone? We can take<br class="">
date / time planning and logistics<br class="">
off-list.<br class="">
<br class="">
<br class="">
Given that this change affects a lot of<br class="">
organisations and people in the wider<br class="">
Java community I think it would be<br class="">
really appreciated that a brief summary<br class="">
of the discussion be published<br class="">
somewhere in public. Maybe this mailing<br class="">
list?<br class="">
<br class="">
regards,<br class="">
<br class="">
Richard Warburton<br class="">
<br class="">
<a href="http://insightfullogic.com/" rel="noreferrer" target="_blank" class="">http://insightfullogic.com</a><br class="">
@RichardWarburto<br class=""></span>
<<a href="http://twitter.com/richardwarburto" rel="noreferrer" target="_blank" class="">http://twitter.com/richardwar<wbr class="">burto</a>><br class="">
</blockquote>
<br class="">
<br class="">
</blockquote>
<br class="">
<br class="">
</blockquote>
<br class="">
<br class="">
<br class="">
<br class="">
</blockquote>
<br class="">
<br class="">
<br class="">
</blockquote>
</blockquote></div><br class=""></div>
</div></blockquote></div><br class=""></body></html>