....and until WCG gives it an all clear to use, I would not touch it with a ten foot barge pole!

I put it up on one machine and it works fine so you can put away your barge poles. However, I found nothing new that benefits me as a cruncher (odd for a full number release) and that would justify the time it takes to upgrade the rest of the farm. Sometimes it seems BOINC releases a new version because ... they can.

I do actually like the way the new BOINC client version handles requests for new work (there's a thread from SekeRob somewhere explaining exactly how it works, if anyone's interested). And that's also the reason why I'm using 7.0.25. So, no need for barge poles here, either

I'm actually using 6.12.34 since I got back into BOINC stuff at the end of January 2012 here. It's running smooth as a baby's (well, you know what goes here). hehe

Saw some [internal] reasoning by knreed why he had not activated <dont_use_dcf> which effectively tosses the DCF that hosts compute of their own and use the server 700 algorithms to assign work volumes based on performance and cache size. So happens that WCG was way ahead of the main testbed pack, SETI, where DR. Anderson usually tests his software. They seemed to have upgraded to server 700 just last night and did have the <dont_use_dcf> on, and every 7.0.25 client was starting to process in high priority, due a bug, fixed in 7.0.28. Anyway, Rom wrote back to say:

This amounts to "oops" forgot to add 7.0.28 to the formal testing cycler, but soon it will be promoted [if if passes], and that broken 7.0.25 can be put to bed.

Oh, and would you have incurred this, and work fetching stopped, because of this High Priority processing, and all crunching resources going to SETI [now that was a stealth move if it were planned, not of course], you know why early adoption of client 7 has/had it's unwanted consequences.;>)

--//--