Running,high priority
11 Aug 2008 4:31:14 UTC
Topic 193844
(moderation:
I seem to only get wu's that say that on my dual core machine but right now for the 1st time I have one running on "high priority" and the other is just "running" and they are both S5R4a's
(and the only difference as far as the due date is 2 seconds)
Wonder why that is.
Running,high priority
)
Hmmm...
I'm assuming you're talking about your PD?
If so, with the number of tasks it's carrying I'm going to say that BOINC is just being overly cautious here. Mostly likely due to the Task Duration Correction Factor (TDCF) not being fully adjusted to S5R4 yet.
Looking at the average runtimes showing for R4 for it so far, I don't think your host is really in any danger of missing deadlines (especially since it looks like you're only running EAH right now).
I'd just let it go on it's own right now, but keep an eye on it.
HTH,
Alinator
RE: I seem to only get wu's
)
If you are regularly seeing task tagged as running at high priority you might consider lowering your queue setting
Computing Preferences|Maintain enough work for an additional n.nn days
The symptom you describe suggests you are often getting enough work in your queue that BOINC doubts it will finish on time.
Many folks probably got that as a transient state at the transition from S5R3 to S5R4, but if you got it away from that transition you have an issue to address.
Before I finished posting Alinator had something to say--we both agree that it makes a difference whether you have seen the "high priority" state only at the moment of transition to S5R4.
LOL... Agreed, and since
)
LOL...
Agreed, and since two's a quorum in EAH, I guess we can call this case closed for the time being! :-D
Alinator
LOL.......um no I wasn't
)
LOL.......um no I wasn't worried about them ever being late.....never had that problem.
I have had this high priority on my dual core since the previous version and now on the S5R4's.
But on my old P4 2.5 and old AMD 2200 I always have them loaded the same and they never say high priority.
Just wondered why it would say one was high priority and the other wasn't when they were running at the same time on the dual core and due at the exact same time (minus 2 seconds)
And yes.....just running Einstein since LHC is sound asleep most of the time and I only run seti if both are down at the same time
I am only interested in Einstein and LHC .....I leave the rest of the projects for the other million or so people to run.
So no it had nothing to do with the transition.
Just wondering......no big deal.
I am physicist so I tend to have to look at everything and when you have a house of pc's and live in the middle of nowhere it is either my pc's......or watching satellite tv and I have 5 satellite dishes too
RE: LOL.......um no I
)
If you have an always-on internet connection, try setting 'Connect every...' to 1.0 or less, and 'Additional' to the number of days you want to keep in your cache (up to 10).
If your Connect interval is too long, then BOINC gets worried whether it can finish and report before the deadline.
[edit]fix tags[/edit]
Seti Classic Final Total: 11446 WU.
RE: [b][=purple][=16]LOL...
)
Yeah, after looking over the publicly available info about your multi core hosts, it does seem kind of curious that you should see High Priority (aka the operating mode formerly known as EDF) on them when EAH is the only project currently running.
There are a few things which could account for it.
1.) You don't run 24/7 all the time or have a 'variable' uptime schedule. Your time metrics play a role when the Core Client (CC) does the Round Robin simulation to check for potential scheduling jams.
2.) 'Flakey' or inconsistent Benchmark runs. Obviously, having the Benchmarks accurately model the hosts calculational performance would be the best situation across the board. However, consistent will do in most cases, since there aren't that many really popular projects which use straight BM-T scoring anymore.
I've had times where I've gotten a 'bad' benchmark run which caused the CC to start making, shall we say poor, scheduling and work fetch decisions temporarily. Ordinarily this isn't a problem unless you are trying to carry the absolute maximum amount of work with your cache settings. In your case that doesn't appear to be the case.
3.) Once in a while, the CC is known to screw up what it has stored for your time metrics for no apparent reason. This is usually a temporary situation as well, and normally doesn't cause a problem as the CC will straighten it out fairly quick on its own.
That's all that comes to mind right now, maybe some of the other folks have more thoughts on the matter.
HTH,
Alinator
Yes the "no apparent reason"
)
Yes the "no apparent reason" sounds right to me Alinator
The only time my machines stop running is if the power goes out and that has been about 2 years now.
And am running on a dsl connection (finally got to replace the satellite dish isp)
Both are running at "high priority" now and I have 17 more to finish before the 23rd.....they will be done by the 18th.
I only change the Preferences if maybe LHC has some and it is set for 5 days now.
But it was at 1 day for the last few months and I keep them all running with plenty of work to do since they don't do much else (haven't played a video game since the days of pong and back then the only computer I saw was what NASA had)
Ok sunny day so I better let this thing do it's work and maybe go hit some golf balls around.
Thanks guys.
-Samson-
RE: Ok sunny day so I
)
LOL...
Sounds like a plan...
A bad day of golf beats a good day of BOINCing, hands down! :-D
Alinator