Still have problems with comps running dry (overnight).
If they don't get work within the first requests of the BOINC manager then the requests get delayed more and more - and the boxes are sitting idle. *sniff sniff*

PS. The lame P4 boxes have more chances to have WUs running of course, the faster comps need to be fed more often. ;-)

---

Lovely greetings, Cori

Compared to yesterday my rigs (CPU only) seem to be running well, and fed. Yesterday, with the cache at 20 per CPU, I was in the same position as everyone else - zilch.

Everyone? Zilch? I got a shed load of WUs yesterday, I can tell you. Did anyone else get any? Did everyone get zilch WUs yesterday? Was I the only one who got any WUs yesterday?

If you're going to bitch and complain, at least don't drag me into your 'everyone'.

---

At this moment I see also workunits that are getting q credits. I did not see this before this night. Now it looks like to happen with about 10% of the wu.

I am seeing a rate of about 15% and it only just started with the newer work units....

Compared to yesterday my rigs (CPU only) seem to be running well, and fed. Yesterday, with the cache at 20 per CPU, I was in the same position as everyone else - zilch.

Everyone? Zilch? I got a shed load of WUs yesterday, I can tell you. Did anyone else get any? Did everyone get zilch WUs yesterday? Was I the only one who got any WUs yesterday?

If you're going to bitch and complain, at least don't drag me into your 'everyone'.

You'd have to be someone first...

With the 20 limit, I got more 0 responses than before, but on the other hand I got more WUs when I occasionally got some.

With the limit at 6, I anticipated a lot of 'limit reached' responses, so I lowered my cache from one to half a day. So far, I've seen *none* of the 0 responses, but only a few WUs with each request.

Of course, both my ATI-equipped crunchers are asking for work almost continuously now, but that doesn't seem to matter much.

---

If there are so many problems with getting as many WU's out as is necessary to keep all crunching cores happily working, why do not you increase the crunching length of a WU? If you increase it by a factor of two, then the amount of SQL-requests to the dataserver should be lowered and server responsiveness should increase accordingly...

---

OK shoot me for this!
The baby ATI cruncher grew into a monster with an unsatisfying appetite
A quick calculation shows that a 4850 in a quad can gobble up +/- 1400 work units in an hour!
These monsters multiplied rapidly when a lot of crunchers rushed to the stores to get theirs
Now the server and network can’t keep up feeding these hungry monsters.

Result:
Everyone is starving!

I still run out of work, two CPU apps dried up over the last 3 hours left to their own devices. The small WU queue did improve the situation a lot though.
This is with a 4870- not so good:

02/28/09 06:10:38|Milkyway@home|Sending scheduler request: To fetch work. Requesting 129629 seconds of work, reporting 8 completed tasks
02/28/09 06:10:43|Milkyway@home|Scheduler request completed: got 8 new tasks
02/28/09 06:10:53|Milkyway@home|Sending scheduler request: To fetch work. Requesting 115461 seconds of work, reporting 0 completed tasks
02/28/09 06:10:58|Milkyway@home|Scheduler request completed: got 0 new tasks
02/28/09 06:12:00|Milkyway@home|Sending scheduler request: To fetch work. Requesting 132473 seconds of work, reporting 8 completed tasks
02/28/09 06:12:05|Milkyway@home|Scheduler request completed: got 0 new tasks
02/28/09 06:13:05|Milkyway@home|Sending scheduler request: To fetch work. Requesting 149010 seconds of work, reporting 0 completed tasks
02/28/09 06:13:10|Milkyway@home|Scheduler request completed: got 0 new tasks
02/28/09 06:14:11|Milkyway@home|Sending scheduler request: To fetch work. Requesting 165398 seconds of work, reporting 9 completed tasks
02/28/09 06:14:16|Milkyway@home|Scheduler request completed: got 0 new tasks
02/28/09 06:15:16|Milkyway@home|Sending scheduler request: To fetch work. Requesting 172803 seconds of work, reporting 7 completed tasks
02/28/09 06:15:21|Milkyway@home|Scheduler request completed: got 0 new tasks

OK shoot me for this!
The baby ATI cruncher grew into a monster with an unsatisfying appetite
A quick calculation shows that a 4850 in a quad can gobble up +/- 1400 work units in an hour!
These monsters multiplied rapidly when a lot of crunchers rushed to the stores to get theirs
Now the server and network can’t keep up feeding these hungry monsters.

Result:
Everyone is starving!

Bang! Travis has recently repeated that there are plenty of WUs on the server side. The problem is releasing them to the clients (us), which I understand requires a server memory fix which will be done on Monday, hopefully, the 20/6 WUs being a temporary fix. It has nothing to do do with WU shortage so blaming ATI will do nothing for you.

Go here; http://milkyway.cs.rpi.edu/milkyway/server_status.php

There are 700 WUs for you not being gobbled by ATIs

[edit] it just went up to 800 +

---

Although the server side says there is plenty of WUs to be released so, clearly, the issuing to work to requests are the problem still. Again I am, uncharacteristicly, out of work.

Let us hope the problem, and server cache work waiting, issues can be solved early next week.

::Waving to Anton:: who is still sore!

Sore? Try a few more idiotic statements to entertain us.

---

1 WU today 0 yesterday.

:lolol: :lolol: :lolol: :lolol:

Funny kindergarten here.

---

Lovely greetings, Cori

1 WU today 0 yesterday.

The work can be had, I've been getting some across my entire Pharm, so you have some other reason your not getting any WU's ...

Beats me, ill check out the second PC to see if it has any work.

_________________________
Other one has a ton of work, weird.

[/quote]
Bang! Travis has recently repeated that there are plenty of WUs on the server side. The problem is releasing them to the clients (us), which I understand requires a server memory fix which will be done on Monday, hopefully, the 20/6 WUs being a temporary fix. It has nothing to do do with WU shortage so blaming ATI will do nothing for you.

Go here; http://milkyway.cs.rpi.edu/milkyway/server_status.php

There are 700 WUs for you not being gobbled by ATIs

[edit] it just went up to 800 +[/quote]

Sorry I think you missed

Server status shows 753 ready
I immediately hit update on my dry quad and received 0
I’m not blaming ATI (its ability is amazing!) but this started to escalate when more and more crunchers uses their ATI’s.
I looked randomly at 10 of the top 100 computer’s RAC – they are all ATI equipped.

Other fast PC’s may also contribute to this.
I’m also guilty because I OC’d my Q8200 to 2.9 GHz to pump up my RAC
It’s doing a S82 in 250 seconds
Yesterday I turned down the cache to minimum 0.01 but its scheduler still wants 9000 seconds of work and repeatedly ask for more every 60 seconds even with 76 wu’s ready to start
Doesn’t this put a big strain on the network and server?

Edit
I can also install ATI but with my luck my RAC will go down

---

A BLAST FROM YOUR PAST

Beats me, ill check out the second PC to see if it has any work.

_________________________
Other one has a ton of work, weird.

Just Reset the Project, don't Detach but just reset it if it doesn't have any work & see if you get some. You may not right first but in awhile you might again ...

Beats me, ill check out the second PC to see if it has any work.

_________________________
Other one has a ton of work, weird.

Just Reset the Project, don't Detach but just reset it if it doesn't have any work & see if you get some. You may not right first but in awhile you might again ...

Wait a second...

Resetting the project as a workaround to get more work might be counter productive. The reason is that even if you already have the third party apps to install afterwards, your host is going to have to re-download all the input files it needs that got dumped from the reset. If running stock, it will have to download the app again too.

This is probably not going to do much to help the overall situation. ;-)

Alinator

Hmm ya, one out of two working is good enough.

Beats me, ill check out the second PC to see if it has any work.

_________________________
Other one has a ton of work, weird.

Just Reset the Project, don't Detach but just reset it if it doesn't have any work & see if you get some. You may not right first but in awhile you might again ...

Wait a second...

Resetting the project as a workaround to get more work might be counter productive. The reason is that even if you already have the third party apps to install afterwards, your host is going to have to re-download all the input files it needs that got dumped from the reset. If running stock, it will have to download the app again too.

This is probably not going to do much to help the overall situation. ;-)

Alinator

No, Resetting the Project won't get rid of any Optimized App's, Third Party App's I don't know about but the Optimized ones don't get taken out when Resetting. Yes you'll have to re-download the input files & if you have Dial-Up it could take some time or pose a problem.

I do it all the time but I have Cable so it's no big deal for me to re-download the input files, it only takes a few seconds for me to do that. It works for me but if somebody has a problem with it then don't do it, more work for me then ... :)