Windows S5R2/S5R3 App 4.40 available for Beta Test

rbpeake
rbpeake
Joined: 18 Jan 05
Posts: 266
Credit: 967827065
RAC: 1210978

RE: Bernd, who currently

Message 72906 in response to message 72905

Quote:

Bernd, who currently has poor Internet connectivity, informed us that some last minute improvements were included in the S5R3 codebase. This means that, contrary to previous announcements, the current set of beta apps will NOT be able to crunch S5R3 results.

There are still a few days worth of S5R2 results in the server pipeline, so there's no immediate hurry. But you should be prepared to de-install any beta app you might still have, including the most recent one, in the next few days so that your computer(s) will automatically download the new S5R3 app once the new run begins. It's safest (and the most traffic friendly way) to let the whole currently loaded pipeline of results finish and only then remove the app_info.xml file (no matter how long that takes). If you perform the change earlier, any crunching and un-crunched results in your pipeline may be rejected, but you will get new ones after the currently available official app is downloaded automatically.

CU

H-BE


Thanks very much for the "heads-up"! I may just uninstall the Beta after the current work queue is finished, perhaps that means I will go back to 4.38 for a few days, but so be it! :)

anders n
anders n
Joined: 29 Aug 05
Posts: 123
Credit: 1656300
RAC: 0

This app works fine even with

This app works fine even with a slow computer like this one. :)

Anders n

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 2139
Credit: 2752659155
RAC: 1473243

Just got an interesting

Just got an interesting series of messages on a 4.40 Beta host:

Einstein@Home 22/09/2007 13:39:13 Requesting 386 seconds of new work, and reporting 1 completed tasks
Einstein@Home 22/09/2007 13:39:18 Scheduler RPC succeeded [server version 601]
Einstein@Home 22/09/2007 13:39:18 Message from server: No work sent
Einstein@Home 22/09/2007 13:39:18 Message from server: To get more Einstein@Home work, finish current work, stop BOINC, remove app_info.xml file, and restart.
Einstein@Home 22/09/2007 13:39:18 Deferring communication for 4 hr 0 min 0 sec

Nice. Server version 601: clear and informative message: and we're getting ever closer to S5R3. Good news all round.

I was just slightly surprised to get the message so early, when there are still over 33,000 S5R2 results "unsent".

Also, I see that there is a new v4.01 Windows application (and similarly for the other platforms) on the application list page - dated today! Does anyone know what will happen if I delete the Beta app_info on the machine, as recommended? It's a quad-core, which I've "lent" to Einstein full-time to help finish off the run. It would be a shame to trash the Beta work on the other three cores, just to keep the fourth core running....

Edit - the message is so clear, I completely skipped the finish current work phrase! Looks like I will be allowing SETI back onto the machine slightly earlier than planned, to soak up those spare cycles as the cores finish (at staggered intervals, of course).

DanNeely
DanNeely
Joined: 4 Sep 05
Posts: 1364
Credit: 3562358667
RAC: 1580

RE: Nice. Server version

Message 72909 in response to message 72908

Quote:

Nice. Server version 601: clear and informative message: and we're getting ever closer to S5R3. Good news all round.

I was just slightly surprised to get the message so early, when there are still over 33,000 S5R2 results "unsent".

I got a new S5R2 WU at 13:14, so you seem to've just missed the last. I'd assume the 30k remaining are the initial s5r3.

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 686042851
RAC: 592401

RE: RE: Nice. Server

Message 72910 in response to message 72909

Quote:
Quote:

Nice. Server version 601: clear and informative message: and we're getting ever closer to S5R3. Good news all round.

I was just slightly surprised to get the message so early, when there are still over 33,000 S5R2 results "unsent".

I got a new S5R2 WU at 13:14, so you seem to've just missed the last. I'd assume the 30k remaining are the initial s5r3.

No, those 30k units are still S5R2. I guess the scheduler is just trying to minimize downloads for participants, so it wants to give tasks (either S5R2 or S5R3) to hosts that already have the required datapacks (which are compatible for both runs, while the apps are not). The 4.x apps are apps for S5R3.

CU

H-BE

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 2139
Credit: 2752659155
RAC: 1473243

RE: No, those 30k units are

Message 72911 in response to message 72910

Quote:
No, those 30k units are still S5R2. I guess the scheduler is just trying to minimize downloads for participants, so it wants to give tasks (either S5R2 or S5R3) to hosts that already have the required datapacks (which are compatible for both runs, while the apps are not). The 4.x apps are apps for S5R3.


I don't think that's the explanation. I agree that the 30k are S5R2 WUs (we haven't seen any sign of an S5R3 work generator yet, and the number 'unsent' has been going down steadily and consistently since Bernd opened the floodgates and released the last of the S5R2 run a few days ago).

I have a number of datapaks on the host in question, and the schedulers have been happily sending new ones as required - that machine is currently working on 1 @ 533.65, 3 @ 533.70, and has a 527.25 waiting 'ready to run'. Today's new behaviour is that the scheduler won't issue any of the available work to an S5R2 application, but wants me to let it download the S5R3 app first.

archae86
archae86
Joined: 6 Dec 05
Posts: 3145
Credit: 7023304931
RAC: 1827262

RE: (we haven't seen any

Message 72912 in response to message 72911

Quote:

(we haven't seen any sign of an S5R3 work generator yet, and the number 'unsent' has been going down steadily and consistently since Bernd opened the floodgates and released the last of the S5R2 run a few days ago).


Ah... after this was posted, Bernd has posted on issue of S5R3 results

archae86
archae86
Joined: 6 Dec 05
Posts: 3145
Credit: 7023304931
RAC: 1827262

RE: and the number

Message 72913 in response to message 72911

Quote:
and the number 'unsent' has been going down steadily and consistently since Bernd opened the floodgates and released the last of the S5R2 run a few days ago).


That may have changed as well. Just now I did a page refresh, and the "unsent" number crept up slightly.

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 2139
Credit: 2752659155
RAC: 1473243

RE: RE: and the number

Message 72914 in response to message 72913

Quote:
Quote:
and the number 'unsent' has been going down steadily and consistently since Bernd opened the floodgates and released the last of the S5R2 run a few days ago).

That may have changed as well. Just now I did a page refresh, and the "unsent" number crept up slightly.


Could just have been some deadlines expiring, of course - I could do with some of my pendings getting a kick......

But on the whole, and coupled with Bernd's post that you linked, I suspect that there's a 'hidden' S5R3 work generator running somewhere: they just haven't yet (reasonably enough) rearranged the server status page to make room for separate S5R3 stats.

archae86
archae86
Joined: 6 Dec 05
Posts: 3145
Credit: 7023304931
RAC: 1827262

Ah, yet more progress as S5R3

Ah, yet more progress as S5R3 slips slowly into view. Now the status page has entries for S5R3 generator, validator, and assimilator, and says all three are running at the moment.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.