Hey GPUGRID crunchers,

BOINC v7.2.28 has been released!

This is a recommended upgrade, and includes several scheduling and work fetch fixes that I helped put in :)

Download: http://boinc.berkeley.edu/download.php
Release notes: http://boinc.berkeley.edu/wiki/Release_Notes
Version history: http://boinc.berkeley.edu/trac/wiki/VersionHistory

Enjoy!

WRT the new 'Download BOINC + Virtualbox' option, what version of Virtual Box is included, is it user configurable and is it a C:/ only install or can this be installed on another drive?

I thought VBox 4.2.16 was the bugged version? If so, why is it now available with BOINC 7.2.28? I am under the impression Vbox 4.3.? fixes the bug that has been causing grief for Rom's vboxwrapper and expected to see 4.3.? bundled with 7.2.28. Is 4.2.16 a typo here or am I confused, again.

____________
BOINC <<--- credit whores, pedants, alien hunters

4.2.16 is fine.
4.2.18 and 4.3.0 have bugs that affect T4T and probably RNA World.
4.3.2 has a bug that affects T4T but does not affect RNA World.

See here for details:
http://lhcathome2.cern.ch/test4theory/forum_thread.php?id=1364&postid=15560#15560
(I have done extensive research into all of these, and have filed bugs with Oracle.)

Because 4.3.2 has a known bug, and because the last version to work clean on all projects was 4.2.16, BOINC developers intentionally shipped with 4.2.16.

If you don't want to do any manual intervention, and you want to ensure your tasks complete without problems, stick with 4.2.16 until Oracle resolves Ticket #12291 (which affects 4.3.2)
https://www.virtualbox.org/ticket/12291

If you want to run 4.3.2 while also running T4T, then you should run T4T anonymous, using vboxwrapper 26033 or better, as the executable. See:
http://boinc.berkeley.edu/wiki/Anonymous_platform
http://boinc.berkeley.edu/dl/?C=M;O=D

Good luck!
Jacob

Jacob, thanks for quick clarification :)

____________
BOINC <<--- credit whores, pedants, alien hunters

You're welcome!

It's a bit of a chore to keep track of all these Oracle bugs, especially in the middle of a BOINC release, but we nailed it. The BOINC release itself is in very good shape, and I'm quite proud of our efforts.

Oracle still has 1 more bug to knock out.

Ironically, if T4T wanted to make 4.3.2 work for them, they could. They'd have to upgrade to the very latest vboxwrapper (I helped Rom put in a showstopping fix yesterday). T4T is thinking about doing that over the upcoming weeks. But without Oracle fixing the bug or T4T moving to the new wrapper, though, the user's choices for bug-free operation are to either stick with 4.2.16, or go anonymous platform with T4T and use the latest vboxwrapper manually along with any vbox version including 4.3.2.

- Jacob

Ok im very interested how the new shedulersystem can handle POEM ;)

Didnt see any changes in the GUI, so it seems there are only but important codechanges.
____________
DSKAG Austria Research Team: http://www.research.dskag.at

You're welcome!

It's a bit of a chore to keep track of all these Oracle bugs, especially in the middle of a BOINC release, but we nailed it. The BOINC release itself is in very good shape, and I'm quite proud of our efforts.

@dskagcommunity: Yeah, the UI hasn't really changed, but the task scheduler has a couple fixes for bugs that would leave GPUs idle sometimes, and the work fetch scheduler has fixes (especially for anyone using GPU exclusions) including a better implementation that should further minimize RPC network calls to the projects, while better honoring user resource share preferences (especially when contributing to both CPU and GPU projects at the same time).

@Dagorath: Thanks for the kind words. I'm not as young as you think (33), I just happen to have a huge passion toward making BOINC work better! I used to run distributed.net before I switched to BOINC, I've investigated Folding@Home (when I was trying to find something that would work on my GTS 240 video card; it now runs Albert/Einstein/Seti/SetiBeta). I'll have to take a peek at Volpex -- thanks for mentioning it.