> Das ist alles :-(
>

Da hast Du wahrscheinlich leider Recht. Schade, denn das Boinc Teil hat eigentlich ordentlich Potential. Ich habe nur echt keine Lust mir wieder so ein Mega-Schei*-Teil wie BOINC 4.25 (Hab' ich mal probiert) auf den Rechner zu ziehen. Passt in keinen einzigen MSStyle und ist schlicht 'unbedienbar'.

[english]
Your'e probably right. What a pity, cause the client has a big potential. But i'm tired to install such buggy Bonkers like 4.25 (I had already installed). Screws all MSStyles and is simply 'unusable'.
[/english]

---

Aloha, Uli

> Ok, so in short please...
> What is the real benefit in upgrading from 4.19 now? ;?
>

The benchmarks are more accurate.

If you are on dail-up or some other itermittant internet connection the 'disable BOINC communication' setting is remembered.

Various other new features that may be a reason to upgrade for some people.

---

BOINC WIKI

BOINCing since 2002/12/8

> Ok, so in short please...
> What is the real benefit in upgrading from 4.19 now? ;?
>
Well, the Ugly Big blue box on the left is a little narrower and now has a white background and more attractive buttons.

does this do it for you?

lol

Here is proof that v4.35 does know how to switch between projects

4/30/05 6:32:03 AM||Starting BOINC client version 4.35 for windows_intelx86
4/30/05 4:30:57 PM|BURP|Requesting 105408.00 seconds of work
4/30/05 4:30:57 PM|BURP|Sending request to scheduler: http://burp.boinc.dk/burp_cgi/cgi
4/30/05 4:30:58 PM|BURP|Scheduler RPC to http://burp.boinc.dk/burp_cgi/cgi succeeded
4/30/05 4:30:58 PM|BURP|No work from project
4/30/05 4:30:58 PM|BURP|Deferring communication with project for 1 hours, 49 minutes, and 17 seconds
4/30/05 5:26:26 PM||request_reschedule_cpus: process exited
4/30/05 5:26:26 PM|Einstein@Home|Computation for result H1_1467.4__1467.7_0.1_T23_Run2_0 finished
4/30/05 5:26:26 PM||schedule_cpus: must schedule
4/30/05 5:26:26 PM|SETI@home|Starting result 24ja05ab.16105.5632.209664.10_3 using setiathome version 4.09
4/30/05 5:26:27 PM|Einstein@Home|Started upload of H1_1467.4__1467.7_0.1_T23_Run2_0_0
4/30/05 5:26:33 PM|Einstein@Home|Finished upload of H1_1467.4__1467.7_0.1_T23_Run2_0_0
4/30/05 5:26:33 PM|Einstein@Home|Throughput 33790 bytes/sec
4/30/05 5:28:40 PM||request_reschedule_cpus: process exited
4/30/05 5:28:40 PM|SETI@home|Computation for result 24ja05ab.16105.5632.209664.10_3 finished
4/30/05 5:28:40 PM||schedule_cpus: must schedule
4/30/05 5:28:40 PM|SETI@home|Starting result 08ja05aa.5574.27120.622156.79_3 using setiathome version 4.09
4/30/05 5:28:41 PM|SETI@home|Started upload of 24ja05ab.16105.5632.209664.10_3_0
4/30/05 5:28:45 PM|SETI@home|Finished upload of 24ja05ab.16105.5632.209664.10_3_0
4/30/05 5:28:45 PM|SETI@home|Throughput 18461 bytes/sec
4/30/05 8:24:02 PM||request_reschedule_cpus: process exited
4/30/05 8:24:02 PM|SETI@home|Computation for result 08ja05aa.5574.27120.622156.79_3 finished
4/30/05 8:24:02 PM||schedule_cpus: must schedule
4/30/05 8:24:02 PM|Einstein@Home|Starting result H1_0118.9__0119.1_0.1_T11_Run2_2 using einstein version 4.79
4/30/05 8:24:04 PM|SETI@home|Started upload of 08ja05aa.5574.27120.622156.79_3_0
4/30/05 8:24:09 PM|SETI@home|Finished upload of 08ja05aa.5574.27120.622156.79_3_0
4/30/05 8:24:09 PM|SETI@home|Throughput 4022 bytes/sec
4/30/05 10:52:04 PM|BURP|No work from project
4/30/05 10:52:04 PM|BURP|Deferring communication with project for 45 minutes and 6 seconds
4/30/05 11:37:10 PM|BURP|Requesting 105408.00 seconds of work
4/30/05 11:37:10 PM|BURP|Sending request to scheduler: http://burp.boinc.dk/burp_cgi/cgi
4/30/05 11:37:13 PM|BURP|Scheduler RPC to http://burp.boinc.dk/burp_cgi/cgi succeeded
4/30/05 11:37:13 PM|BURP|No work from project
4/30/05 11:37:13 PM|BURP|Deferring communication with project for 2 hours, 4 minutes, and 3 seconds
4/30/05 11:49:03 PM||schedule_cpus: time 12300.070000
4/30/05 11:49:03 PM|climateprediction.net|Restarting result 2ad6_200127829_0 using hadsm3 version 4.04
4/30/05 11:49:03 PM|Einstein@Home|Pausing result H1_0118.9__0119.1_0.1_T11_Run2_2 (removed from memory)
4/30/05 11:49:03 PM||request_reschedule_cpus: process exited
4/30/05 11:49:03 PM||schedule_cpus: must schedule

---

98SE XP2500+ @ 2.1 GHz Boinc v5.8.8

And God said"Let there be light."But then the program crashed because he was trying to access the 'light' property of a NULL universe pointer.

<OT>

Mind if I steal the name "Boinczilla"
[polite cough]

You're welcome.

</OT>

---

I have no problems with 4.35 and it appears to love P4 HTs One CPU on PP@H, one on SETI@H and then they switch and both hit LHC@H. Couple this with BOINCVIEW 1.04 and I am a happy camper.....

---

Semper Eadem
So long Paul, it has been a hell of a ride.

Park your ego's, fire up the computers, Science YES, Credits No.

> You're welcome.

Thanks NA. ;-)

---

Rob

Is there anywhere I can find a list of what all the new messages mean?
Like:
schedual_cpus: must schedual
schedual_cpus: time 3600.15468
request_reschedual_cpus: project op
request_reschedual_cpus: process exited
request_reschedual_cpus: files downloaded
etc, I'm sure there are some I've missed

I've only just started using 4.35 seems to be working ok so far, just trying to understand what its doing, lol.

---

Cheers

Gav



Nothing is 'fool-proof', someone will always invent a better fool!

No problems for two days on 4.35..........but...........STATS tab

yellow on light grey not good for my old eyes.....how about a change or make it configurable please?

---

Hi

I read a few posts about the 4.35 here and at einstein.
What makes me a bit confused is what sense makes the deadline mode when a unit is crunched in time but will not reported shortly enough so it didn´t reach the deadline too.
Its happend to one of my machines every week on einstein.
The result is finnished but uploaded a few minutes too late.
For example my brother has connection time only once a week from 7 houres to 8 houres PM on sunday.
He downloaded his unit at 7.10 houres the unit is finnished in time but uploaded and reported next week at 7.12 houres he won´t get credits for this.
I think a bit more tolerance with the deadline would make more sense.
The 4.35 is the first one i will not use.

greetz Mike

---

With each crime and every kindness we birth our future.

> No problems for two days on 4.35..........but...........STATS tab
>
> yellow on light grey not good for my old eyes.....how about a change or make
> it configurable please?
>
Seconded!

I cannot think of a worse colour choice ('cept white-on-white) - configurable would be best, but anything would be better.

---

I've installed one of the BOINC v4.3x series a while ago, and it was slow switching between tabs. Has this been fixed in v4.35 yet? I'm still using v4.19, and I have no reason to switch yet.

They are mostly self explanatory, but I will add some details.

> Is there anywhere I can find a list of what all the new messages mean?
> Like:
> schedual_cpus: must schedual

Something has happened that caused a call to the CPU scheduler to refigure what to work on.

> schedual_cpus: time 3600.15468

The time preference has passed so it is time to recalculate.

> request_reschedual_cpus: project op

A project operation has occured that may cause the CPU to be rescheduled. ie. update, attach, detach, suspend ect.

> request_reschedual_cpus: process exited

A workunit has finished or been preempted. Either causing or caused by a CPU rescheduling.

> request_reschedual_cpus: files downloaded

All files for a workunit have finished downloading. Time to recalculate.

> etc, I'm sure there are some I've missed
>
> I've only just started using 4.35 seems to be working ok so far, just trying
> to understand what its doing, lol.
>
Work fetch policy: max projects exceeded

More projects have work on the client than the current settings allow (default=5). No new work will be downloaded.

work fetch policy: host (nearly) overcommited

The host will be close to or miss one or more deadlines. No new work will be downloaded.

work fetch policy: allow work requests

Normal operations. The client is in good shape and work can be downloaded from new projects.

Schedule CPUs: earliest deadline

One or more workunits is likely to miss the deadline. The CPU scheduler will run the earliest deadline first and no new work will be downloaded.

Schedule CPU: highest debt

No workunits are currently in deadline trouble. They will switch ever hour (default). This is the normal mode that is the same as the previous clients.

I think that's all of them. Hope that helps. Paul feel free to steal this although it might be an idea to wait a few versions to see if any of them change.

---

BOINC WIKI

BOINCing since 2002/12/8

> Well, the Ugly Big blue box on the left is a little narrower and now has a
> white background and more attractive buttons.
>
> does this do it for you?

Only if you can get rid of the blue box & right click to select functions as with 4.19.

---

Grant
Darwin NT

> I think that's all of them. Hope that helps. Paul feel free to steal this
> although it might be an idea to wait a few versions to see if any of them
> change.

John,

Thanks! :)

At the rate I seem to be getting things done, I may not have a problem with waiting ... :(

I seem to be sloooowwwly getting back to work ... :(

---

Query: Does resource sharing still play an important part when running multiple projects.

Also, another question ;)

Say I download a CPDN WU which has a deadline in 12 months, will it see any CPU time in the near future? Or will BOINC wait until it "might" miss the deadline and then run ONLY CPDN?

I must say, the general consesus it that this new feature of BOINC is not exactly winning everyone over....perhaps changing it back?

Also, how will 4.35 handle F@H when it comes online? As far as I know, all WUs will not actually have deadlines (TG's or Timeless gromacs)?

---

>
> Say I download a CPDN WU which has a deadline in 12 months, will it see any
> CPU time in the near future? Or will BOINC wait until it "might" miss the
> deadline and then run ONLY CPDN?

Depends on resource-share, cache-setting and what over projects you're running, but normally you'll switch between the projects each hour just like in earlier clients.

Short-deadline-wu will often be crunched in deadline-mode, but afterwards you'll not ask for more work from this project till the other projects has catched-up based on their resource-share.

So you'll maybe crunch Predictor@home and Einstein@home on day 1 and day 2, before you starts switching between LHC, Seti@home, CPDN each hour on day 3-5 till you ask for more predictor/einstein-wu on day 6 and so on.

Well, the switching isn't exactly like this, but with equal resource-share and 5 projects you'll over a month crunch roughly 6 days on each project.

>
> I must say, the general consesus it that this new feature of BOINC is not
> exactly winning everyone over....perhaps changing it back?

Not likely they'll change it back to the old schedling, since this doesn't work when you've joined very many projects. The new scheduling isn't perfect, but it's much better at handling someone attaching to 20+ projects on one computer.

>
> Also, how will 4.35 handle F@H when it comes online? As far as I know, all WUs
> will not actually have deadlines (TG's or Timeless gromacs)?
>

Folding@home is initially planning on only supplying timeless wu.

For me 4.35 is just working fine, except S@H. Since 1 day or so it can´t get trough and gives me those messages:

01.05.2005 15:02:38|SETI@home|Requesting 259200.00 seconds of work
01.05.2005 15:02:40|SETI@home|make_scheduler_request: -108
01.05.2005 15:02:40|SETI@home|scheduler init_op_project to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi failed, error -108
01.05.2005 15:02:40|SETI@home|init_op_project failed, error -108
01.05.2005 15:02:40|SETI@home|Deferring communication with project for 59 seconds

Can anybody tell me whats wrong with it and what this error -108 means?

Greetz from Germany

edit: CPDN also have problems, but to me it seems to be a CPDN-prob.

I'm still patiently waiting for 4.35 to work the way it has been described as to supposed to be working. So far, yesterday it finished the last of my Einsteins and has not downloaded any more Einstein. It switched yesterday to PPAH and has been crunching them since. I have uploaded and reported many of those, however, It also downloaded more PPAH Wus. WUs that have a deadline which is earlier than my Seti and LHC WUs. There's one more PPAH to go until my existing PPAH WUs (ones I downloaded prior to installing 4.35) have been crunched. It'll be interesting to see if it starts crunching LHC and Seti before it starts in on the Newly downloaded PPAH WUs that have a shorter deadline.

tony

> I'm still patiently waiting for 4.35 to work the way it has been described as
> to supposed to be working. So far, yesterday it finished the last of my
> Einsteins and has not downloaded any more Einstein. It switched yesterday to
> PPAH and has been crunching them since. I have uploaded and reported many of
> those, however, It also downloaded more PPAH Wus. WUs that have a deadline
> which is earlier than my Seti and LHC WUs. There's one more PPAH to go until
> my existing PPAH WUs (ones I downloaded prior to installing 4.35) have been
> crunched. It'll be interesting to see if it starts crunching LHC and Seti
> before it starts in on the Newly downloaded PPAH WUs that have a shorter
> deadline.
>
> tony
>

It will very likely crunch the new batch of predictor@home-wu also, but when finished the long-term-debt is most likely negative for predictor@home so switches to the other projects.