LHC@home

New version 4.45 now shown as recommended version

> New version 4.45 now shown as recommended version
>
recommended by whom? Seti guys? LHC staff? Anyone out there?

---

> > New version 4.45 now shown as recommended version
> >
> recommended by whom? Seti guys? LHC staff? Anyone out there?
>

LHC and CPDN still have 4.19 on their download page. All the other projects link to the Boinc download page which lists 4.45 as the current recommended release.

It's been running a couple days on some user's systems in the Seti forums.

A glance at their threads don't show too many complaints. It's a bit early to declare it 'stable'.

> It's been running a couple days on some user's systems in the Seti forums.
>
> A glance at their threads don't show too many complaints. It's a bit early to
> declare it 'stable'.
>
True but problems with the 4.25 and 4.43 versions made rushing it a bit desireable.

---

BOINC WIKI

BOINCing since 2002/12/8

> > > New version 4.45 now shown as recommended version
> > >
> > recommended by whom? Seti guys? LHC staff? Anyone out there?
> >
>
> LHC and CPDN still have 4.19 on their download page. All the other projects
> link to the Boinc download page which lists 4.45 as the current recommended
> release.

I think that the devs here at LHC should put a notice like this:

"Recommended version for running the project solo. For running multiple projects download from BOINC DOWNLOAD PAGE."

---

Professor Desty Nova
Researching Karma the Hard Way

>
> True but problems with the 4.25 and 4.43 versions made rushing it a bit
> desireable.
>
Seems to me that not rushing it is what's needed with all the recent hiccups.

---

Gaspode the UnDressed
http://www.littlevale.co.uk

My suggestion about which BOINC client version to use still remains: Use the version which is lowest, determined by what's recommended by the project(s) you're attached to. LHC says 4.19 (for now) - I haven't heard of any BOINC 4.19/LHC related problems in months - so that's what I use. On my other two machines which are only attached to P@H & CPDN are still runing 4.25 and as Alex was saying, it's a good idea to watch the forums for problems related to the "recommended" version. The last few "recommended" versions seemed to have a very "beta" feel to them, which had been reflected by threads in various forums across all boinc-enabled projects.

---

@ Mr. "constructor of the new client 4.45" John McLeod VII,
How can I get new work when I report work?
On "multi-projects" the client get only work from the "last done" project.
To get work from the other projects, I must "suspend" all other projects.
Is that the way to work with this silly client?
This client is more to prevent future participants to join the BOINC-projects!!!

greetz littleBouncer

I agree lb,although I like the options of 4.45 , I have had a host sitting there with no work requesting no work because (I beleive) my resource share for LHC was too high even though I had no LHC work. That is silly and makes you wonder why they don't still call this a development version. It would be nice for once to have Boinc work as advertised and not have to manually check hosts all the time. At least 4.45 is not in constant panic mode doing edm like 43 and 44 were and honors resource settings

---

I have had a host sitting there with no work requesting no work because (I beleive) my resource share for LHC was too high even though I had no LHC work.
==========

I find that quite normal for the Client, I have the same problem because I usually run LHC at a 99% - 100% Resource share setting. If I run out of LHC WU's I have to back LHC off in order to get more work from some of the other Projects ...

I've even received messages from the other Projects that there was work available but because blah blah blah (basically the other projects are telling me I'm not running a high enough Resource share for that Project for them to send me any work) ... Then as soon as I back off LHC I get some work from the other Projects ...

> I have had a host sitting there with no work requesting no work because (I
> beleive) my resource share for LHC was too high even though I had no LHC work.

There are 24 hours in a day. So, just a suggestion, if your resource share amongst all projects totaled 24, then for every number you assign a project it gets an hour of time.

e.g. (my resource shares):

LHC = 16
CPDN= 2
P@H = 4
E@H = 2
Total 24

So LHC gets 16 hours, Climateprediction.net gets 2 hours a day, etc.

---

> > I have had a host sitting there with no work requesting no work because
> (I
> > beleive) my resource share for LHC was too high even though I had no LHC
> work.
>
> There are 24 hours in a day. So, just a suggestion, if your resource share
> amongst all projects totaled 24, then for every number you assign a project it
> gets an hour of time.
>
> e.g. (my resource shares):
>
> LHC = 16
> CPDN= 2
> P@H = 4
> E@H = 2
> Total 24
>
> So LHC gets 16 hours, Climateprediction.net gets 2 hours a day, etc.
>
>
with the clients using the rules of the debt and earliest deadlines it would rather put the most time to P@H, due they have shorter deadlines than the other projects.

with 4.45 it looks a little better, earlier releases had the situation, put in your case, they rather get new work from P@H instead of downing the rest of the queue.

The new downloaded P@H work, depending of it's new deadtime, have been done until the deadline of the other projects hit the rules.

several bugfixing have been checked in, might they need a bid time to test and provide them;)
___________________________________

The debt level also play it's game...
even dev people gave the advice to manualy edit the file client_state.xml
and reset the debt (new short_term_debt) and long_term_debt

<long_term_debt>0.000000</long_term_debt>

important is the 0.000000, because this is more than only 0.0 ;)

sometime this value went to the "wrong way" or has invalid entries with ##
inside.

_____________________________________

I'm not sure, if a boinc day has 24 Hours

> The debt level also play it's game...
> even dev people gave the advice to manualy edit the file client_state.xml
> and reset the debt (new short_term_debt) and long_term_debt
>
> 0.000000
>
> important is the 0.000000, because this is more than only 0.0 ;)
>
> sometime this value went to the "wrong way" or has invalid entries with ##
> inside.
-----
With 'the "wrong way" do you mean negativ numbers? as for example:
[long_term_debt]-17028.941350[/long_term_debt] (LHC Alpha)
[long_term_debt]-75936.719322[/long_term_debt] (EAH)
[long_term_debt]-66431.815251[/long_term_debt] (PAH)
[long_term_debt]-22604.844718[/long_term_debt] (SAH)

4 projects have negative numbers (LHC Alpha, Einstein, PAH and SAH)
the other 2 projects (LHC(Prod) and CPDN ) have positives numbers.

greetz littleBouncer

ja die! setze beide auf 0.000 zurück..

> ja die! (=4 negative) setze beide??? auf 0.000 zurück..
>
Also ich habe 4 negative und 2 positive, wenn Du schreibst 'ja die!' meinst Du damit die 4 welche ich aufgelistet habe? was dann mit 'setze beide' , welches ja meine einzigen positiven wären, zu einem Widerspruch führt.

Eigentlich würde ich schliessen: die 4 negativen Werte sind 'falsch' und sollten mit 0.000000 ersetzt werden, ist das richtig?

Danke littleBouncer

BTW: @ all ,sorry to write in german, it was to complicate to translate, and it was nothing of importance (or evidence)...

hast die Richtigen erwischt

each project has 2 xml tags for keeping the values of debt


Setze bei allen Deinen 4 Projekten die beiden debt Einträge auf je 0.00000, dann beginnt der scheduler wieder neu mit dem "zählen".
Mehr negativ =je schlechter das Projekt berücksichtigt wird , je positiver, je besser.

With the time, in theorie, those values should, yes should turn back to the other side (-+) and the balance will be established. But not now :(

Die Werte sind nicht falsch, sie sind das Ergebnis des neuen scheduler, der offensichtlich noch nicht richtig "tickt".

good luck!

Ok Now I have tried to increase my time to connect from 3 to 4 days instead of changing resources.... on my p4 single thread I go into deadline mode with predictor... says I won't finish b4 next connect time( Don"t have this host set up for posting so no logs sorry) Deadline is 6-21 whats wrong with this scenario? Again a rushed development version made mainstream.... I'm considering setting all back to 4.19 (even with Burp) until they address these issues edit.. guess I need to try 4.19 to see if it is worth giving up all the extra choices 4.45 delivers.Maybe 4.25 is next choice.... this is getting to be tooooo much work with 4.45 1st too much work now not enough...I bet devs gettin an earfull hehe edit

---

Increasing the connect time, as you've seen, causes BOINC to have kittens regarding deadline modes since it doesn't like to have wu's in it's cache that have a deadline that is more than 2 times the connect time. So since Predictor has deadlines of 10,000minutes, 6.94 days, BOINC doesn't want you to have a connect to network preference set to more than 3.47 days. And in reality since, as soon as you fill your cache with work a second later it is within 3.47 days of the deadline BOINC goes directly into deadline/panic mode. So if you want BOINC to work with anything resembling 4.19 operation you need to ensure you do not set your connect to network preference at more than 80 to 90% of half the shortest deadline. Sucks the big one IMHO.

Why on earth the devs thought that this was a good idea I cannot comprehend. I can understand a setting that puts BOINC into deadline mode when you get to within say 2 days of a deadline (for the shorter wu completion projects), but not when a wu gets to within half of the deadline. This just severely limits cache levels.

Live long and crunch.

---

Paul
(S@H1 8888)
BOINC/SAH BETA

From what has been discussed over at SETI, it appears that the 2X connect time is built in to allow for modem users. So far the DEVs have not settled on a way to make it friendly for always connected folks. This has been point brought up many times to the DEVs to fix and make a lot of people rest easy with their wanting higher cache levels.

---