Hi,

If you\'re not too far along (i.e. not late in phase 3) it may be a good idea to \"Reset Project\", get the BOINC client 4.05, and reattach to us and get the new \"launch\" version (hadsm3_4.03). If you\'re happy with your run you can continue to use the beta client & CPDN app.

Thanks, but I'll stick with the 4.02 client for now, until I know where my bottom trickle is. ;)
--------------------
Jordâ„¢
<img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=2&amp;trans=off">

I installed BOINC v4.05 over v4.02. Works fine without resetting the project (and I'm still running hadsm3_4.02).

[Edit 2004-08-25 16:27 UTC]
Screensaver bug found in this <a href="//climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=41">thread</a> not fixed in BOINC v4.05

Sure it will still work okay, that's because you are only changing the BOINC core client, not the actual application doing the crunching.

So in essence you didn't change anything. ;)

Isn't it possible to download &amp; install the actual crunching client seperately and let that rip on the workunit we still use, Carl? Or was that workunit made specifically for the 4.02/4.03 client?
--------------------
Jordâ„¢
<img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=2&amp;trans=off">

&gt; Isn't it possible to download &amp; install the actual crunching client
&gt; seperately and let that rip on the workunit we still use, Carl? Or was that
&gt; workunit made specifically for the 4.02/4.03 client?

everyone should probably get the 4.05 client since it handles suspends/resumes/quits better than the 4.02/3 IMHO. I have set 402-era workunits as "deprecated" in BOINC, so I believe when you finish a run BOINC will update your CPDN software to the hadsm3_4.03 versions.

BOINC doesn't upgrade app software within a workunit, which is sensible.

I installed 4.05 over 4.03 and it reports 4.05 as a beta as well,
Is this an artifact of installing over 4.03?

&gt; I installed 4.05 over 4.03 and it reports 4.05 as a beta as well,
&gt; Is this an artifact of installing over 4.03?

well I guess that BOINC/SETI consider this beta still, it's probably just hardcoded after the version # in the gui client.

Hi, Carl,

Re: moving models to different machine. On 18 Aug (0746 UTC "Exit code -5" thread), Honza posted that he successfully moved a run to a different machine. Is this still possible with boinc v.4.05 &amp; CPDN v.4.03?

Jim
________________________________________________
Video meliora, proboque; Deteriora sequor
I see the better way, and approve it; I follow the worse
-- Ovid (43BC-17AD)

&gt; Hi, Carl,
&gt;
&gt; Re: moving models to different machine. On 18 Aug (0746 UTC "Exit code -5"
&gt; thread), Honza posted that he successfully moved a run to a different machine.
&gt; Is this still possible with boinc v.4.05 &amp; CPDN v.4.03?
&gt;
&gt; Jim
&gt; ________________________________________________
&gt; Video meliora, proboque; Deteriora sequor
&gt; I see the better way, and approve it; I follow the worse
&gt; -- Ovid (43BC-17AD)
&gt;
BOINC by default is setup so that the server verifies that the computer that was assigned the work is also the one that crunched the work. So unless CPDN has changed this behavour, it is not possible to move a WU from one machine to another.

jm7

Is this version compatible with Windows ME yet?

&gt; Is this version compatible with Windows ME yet?

I think it works with Win ME, I did get ME working on 4.02.

PS -- you will probably crash a run if you try to move it, I dont think you can go from 4.02-&gt;4.03

&gt; &gt; Is this version compatible with Windows ME yet?
&gt;
&gt; I think it works with Win ME, I did get ME working on 4.02.
&gt;
&gt; PS -- you will probably crash a run if you try to move it, I dont think you
&gt; can go from 4.02-&gt;4.03
&gt;
&gt;
&gt;
I'll give it a try :)
Had it working (sort of) with ME before , complete with console window, but I got too many error messages.

&gt; &gt; Hi, Carl,
&gt; &gt;
&gt; &gt; Re: moving models to different machine. On 18 Aug (0746 UTC "Exit code
&gt; -5"
&gt; &gt; thread), Honza posted that he successfully moved a run to a different
&gt; machine.
&gt; &gt; Is this still possible with boinc v.4.05 &amp; CPDN v.4.03?
&gt; &gt;
&gt; &gt; Jim
&gt; &gt; ________________________________________________
&gt; &gt; Video meliora, proboque; Deteriora sequor
&gt; &gt; I see the better way, and approve it; I follow the worse
&gt; &gt; -- Ovid (43BC-17AD)
&gt; &gt;
&gt; BOINC by default is setup so that the server verifies that the computer that
&gt; was assigned the work is also the one that crunched the work. So unless CPDN
&gt; has changed this behavour, it is not possible to move a WU from one machine to
&gt; another.
&gt;
&gt; jm7
&gt;

Hi, John,

Thanks for the reply.

To quote Will Rogers: All I know is what I read in the newspapers. The machine ID was updated in Honza's experience; my question was/is -- does this continue into the new release? (I have a 'classic' THC run finishing in a few days and I want to re-boot my Alpha/Beta M$ test machine to Linux and give its M$ run to another machine [for reasons to do with my home-network configuration].)


Honza's post:

Honza
Joined: Aug 5, 2004
Posts: 45

Posted: 18 Aug 2004 7:46:10 UTC
in response to Message ID 1110.

&gt; &gt; Still how about migration CPDN BOINC model from one computer to another?
&gt;
&gt; This isn't possible with WUs for other BOINC projects, so I suspect the same
&gt; is true with CPDN.
&gt;
Well, if it isn't possible, there should be check before running WU that it is on the same computer (ID) and prevent it from running.

My migration (moving whole BOINC folder to another computer) worked with no glitch so far - no complaining from BOINC, just running benchmarks, Machine description updated. New model is doing good (16%); the old one went almost to Phase 3...
btw, i wan't had such problem with Alpha.

ID: 1165 / Score: 0
________________________________________________
Indeed I tremble for my country when I reflect that God is just.
-- Thomas Jefferson

&gt;
&gt; PS -- you will probably crash a run if you try to move it, I dont think you
&gt; can go from 4.02-&gt;4.03
&gt;

Hi, Carl,

Not what I want to do; I want to move a 4.03 run, lock-stock-and-barrel for two P4 Models -- i.e., the entire folder as Honza did with an earlier version.

________________________________________________
Indeed I tremble for my country when I reflect that God is just.
-- Thomas Jefferson

&gt; To quote Will Rogers: All I know is what I read in the newspapers. The
&gt; machine ID was updated in Honza's experience; my question was/is -- does this
&gt; continue into the new release? (I have a 'classic' THC run finishing in a few
&gt; days and I want to re-boot my Alpha/Beta M$ test machine to Linux and give its
&gt; M$ run to another machine [for reasons to do with my home-network
&gt; configuration].)

I don't think it's a matter of machine ID. It's a matter of returning the WU. With the other BOINC projects, returning a WU on a machine that didn't download that WU will equate to zero credit.

Why Honza's transfer appears to be working, I couldn't say. Unless maybe the trickles are approached differently? With the other BOINC projects, you aren't credited until completion. (and other hosts verify your results) CPDN gives incremental credits based on trickles.

However, as me and UK_Nick just found when finishing a model, BOINC is still trying to do something with credits at the end, because we received double credits. The credits we had from the trickles, and the credits BOINC gave for completing the WU. (when the other projects normally receive them)

So perhaps a surprise is in store for Honza as his model completes? Could be interesting to see what happens.

<a><img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=18"></a>

Heffed wrote:

&gt; I don't think it's a matter of machine ID. It's a matter of returning the WU.
&gt; With the other BOINC projects, returning a WU on a machine that didn't
&gt; download that WU will equate to zero credit.

I fully understand the reason for this in the SETI context, and the solution is a sensible one when dealing with small WUs. But once again the needs of CPDN are different. If we need to protect the project against fraudulent claiming of credit, and I suppose we do, can there be some other trap to catch these attempts? There's a problem about public discussion of the matter on the board, I suppose, because it may help someone to circumvent it, but I can think of other approaches.

There are several reasons for legitimately wanting to transfer a run, and this has happened often enough under classic CPDN.

&gt; Why Honza's transfer appears to be working, I couldn't say. Unless maybe the trickles are approached differently? With the other BOINC projects, you aren't credited until completion. (and other hosts verify your results) CPDN gives incremental credits based on trickles. Note that is has been done under Beta - not sure about public version...


&gt; However, as me and UK_Nick just found when finishing a model, BOINC is still trying to do something with credits at the end, because we received double credits. The credits we had from the trickles, and the credits BOINC gave for completing the WU. (when the other projects normally receive them)
Should be fixed now...

&gt; So perhaps a surprise is in store for Honza as his model completes? Could be interesting to see what happens.
It a WU 5985, so you can monitor yourself; i'm expecting phase 2-&gt;3 transition overnight.
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/trickle.php?resultid=5985

I agree with Andrew H. that transfering WUs over machines is a sensitive thing but still needed under some circumstances.
Note that i have been moving (not copying) whole BOINC folder, not WU alone.

... and sometime soon you're going to have to run a delete query against the database to allow for all of us who picked up phantom models and/or machines (to add to our previous phantoms) due to our not resetting (still haven't worked out/been told what reset actually does).
Cheers,

PeterV.
<IMG SRC="http://boinc.mundayweb.com/cpdn/stats.php?userID=4&amp;trans=off">

&gt; (still haven't worked out/been told what reset actually does).

It removes everything related to the project (including work units in progress) from your system and requests a new set of programs and work from the server. The computer retains the same identity (as opposed to detaching and reattaching, which will allocate a new computer identity).

<a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=3"><img src="http://www.teampicard.net/templates/fisubice/images/phpbb2_logo.jpg"></a>

&gt; ... and sometime soon you're going to have to run a delete query against the
&gt; database to allow for all of us who picked up phantom models and/or machines
&gt; (to add to our previous phantoms) due to our not resetting (still haven't
&gt; worked out/been told what reset actually does).
&gt; Cheers,
&gt;
&gt; PeterV.
&gt; <IMG SRC="http://boinc.mundayweb.com/cpdn/stats.php?userID=4&amp;trans=off">
&gt;

Hi, Peter,

Do you mean 'merge' the phantoms out of existence?

I think 'merge' can make trickles become phantoms: Installing Carl's revised 'um_' created a machine with no work assigned to it -- nonetheless, it received subsequent Trickles. (The top-rated machine no longer received credit.)
Then, installation of 4.05cc/4.03'um' created another machine.
Then I merged them. The merged value seemed the same as the original machine alone, the second image was seemingly lost -- with 4-5 days of crunching on two models. (The third machine image had not yet received a Trickle.)

Curious, eh? (I wonder if it wouldn't work better if the machines were merged backward rather than forward in time...? On the other hand, once things settle down, new phantoms should not be much of a problem.)

Cheers,
Jim

________________________________________________
Indeed I tremble for my country when I reflect that God is just.
-- Thomas Jefferson