climateprediction.net home page
72 days for wah2_sam25?

72 days for wah2_sam25?

Message boards : Number crunching : 72 days for wah2_sam25?
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · Next

AuthorMessage
Jim1348

Send message
Joined: 15 Jan 06
Posts: 372
Credit: 15,130,852
RAC: 56
Message 59045 - Posted: 20 Nov 2018, 15:34:46 UTC
Last modified: 20 Nov 2018, 15:36:16 UTC

I am running two wah2_sam25's that after 1 day are only about 1.36% complete on my i7-4790 (Win7 64-bit) machine.
The estimated completion time is about 72 days. All my other work units are reasonable enough.

Can this be correct? They have each errored out on other machines, all running Win10 after a short period of time. Note that they are running in separate BOINC instances, so have different machine numbers. But only six cores total are in use for BOINC, with another supporting a GPU on Folding, and another "free".
https://www.cpdn.org/cpdnboinc/result.php?resultid=21383066
https://www.cpdn.org/cpdnboinc/result.php?resultid=21382971

I was assuming that these estimates would correct themselves, but they have not done so thus far.
ID: 59045 · Report as offensive     Reply Quote
Profile geophi
Volunteer moderator

Send message
Joined: 7 Aug 04
Posts: 1821
Credit: 36,393,036
RAC: 10
Message 59046 - Posted: 20 Nov 2018, 16:29:35 UTC - in response to Message 59045.  

I have 4 running on my Ryzen and the manual estimates are for 25 to 35 days. BOINC estimated 45 to 60 days when the tasks began.
ID: 59046 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 15 Jan 06
Posts: 372
Credit: 15,130,852
RAC: 56
Message 59047 - Posted: 20 Nov 2018, 17:03:45 UTC - in response to Message 59046.  
Last modified: 20 Nov 2018, 17:07:40 UTC

Calculating it manually only makes things worse. The longest running one for me is now at 28.5 hours, and 1.442% complete. That works out to 82 days. I expect it is slowing down and is going to get stuck. I will keep it around a while longer.

EDIT: But the other one is now at 17.3 hours and 1.595% complete, which works out to 45 days. Who knows what is happening.
ID: 59047 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 6952
Credit: 20,843,205
RAC: 0
Message 59048 - Posted: 20 Nov 2018, 17:29:54 UTC

Here is one of mine.
This was on Linux Mint, running Wine, running a Windows version of BOINC.

Your problem is probably due to you having fiddled with your computer by putting 2 BOINCs on it.
ID: 59048 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 15 Jan 06
Posts: 372
Credit: 15,130,852
RAC: 56
Message 59049 - Posted: 20 Nov 2018, 17:40:10 UTC - in response to Message 59048.  
Last modified: 20 Nov 2018, 17:50:18 UTC

Your problem is probably due to you having fiddled with your computer by putting 2 BOINCs on it.

Maybe. But the wah2_global are estimating 12 to 14 days on either instance.

ALSO: I have another machine, an i7-3770 also running Win7 64-bit with only a single BOINC instance. It has several wah2_global currently running (after 3 days) with a total time of around 12 days, as manually calculated. (BOINC estimates 6 days left, or 9 days total). But it also has two wah2_sam25 that have not started yet, but have an estimated time of 73 days 18 hours. There is something different about them.
ID: 59049 · Report as offensive     Reply Quote
Profile JIM

Send message
Joined: 31 Dec 07
Posts: 1086
Credit: 19,554,876
RAC: 805
Message 59050 - Posted: 20 Nov 2018, 18:14:03 UTC

A wah2_sam25, 85 month Wu’s on one of my machines is predicted to take 102days 8 hours! I have 3 of these. That’s a lot of crunching. It’s almost like the old days when Wu’s took 6 – 8 month to complete.

I-5 2.6 GHz processor with 4 gb of ram running Win7. My oldest, slowest machine.
ID: 59050 · Report as offensive     Reply Quote
mngn

Send message
Joined: 13 Jul 18
Posts: 26
Credit: 10,557,907
RAC: 239
Message 59051 - Posted: 20 Nov 2018, 18:41:47 UTC - in response to Message 59050.  

On my slow computers the 85 month tasks (batches 762, 763, 764, 765) will take 80 to 160 days.
ID: 59051 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 6952
Credit: 20,843,205
RAC: 0
Message 59052 - Posted: 20 Nov 2018, 19:09:47 UTC

The sam25's (South American), and the "Global" models are 2 totally different things.
And the Global's have their own problems, but way off near the end.

So much so, that in our talks with the project people, it's felt that they should be killed off, and that research re-thought.
A couple of us would like to see them continue, as some may make it. IF we can get some storage space :(

SO ...
If anyone has any Global models, batches 766-770, that have only just started, or are still waiting, then I'd suggest that they be Aborted now, and save the electricity for the New Look Globals, Coming Soon to a Download Server near you. :)
ID: 59052 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 15 Jan 06
Posts: 372
Credit: 15,130,852
RAC: 56
Message 59053 - Posted: 20 Nov 2018, 19:31:48 UTC - in response to Message 59052.  

SO ...
If anyone has any Global models, batches 766-770, that have only just started, or are still waiting, then I'd suggest that they be Aborted now, and save the electricity for the New Look Globals, Coming Soon to a Download Server near you. :)

I was hoping to hear something like that. It is getting near Thanksgiving here in the U.S., so I can get out my axe.
ID: 59053 · Report as offensive     Reply Quote
bernard_ivo

Send message
Joined: 18 Jul 13
Posts: 342
Credit: 9,943,922
RAC: 50
Message 59054 - Posted: 20 Nov 2018, 20:58:37 UTC - in response to Message 59052.  

I've got two 766s at 11% in 2 days... so I guess the end should come rather planned and soon than in few days right?
ID: 59054 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 15 Jan 06
Posts: 372
Credit: 15,130,852
RAC: 56
Message 59056 - Posted: 20 Nov 2018, 21:09:54 UTC - in response to Message 59054.  

I've got two 766s at 11% in 2 days...

2/0.11 = 18 days. That is not too bad, all things considered.

I would hope that the scientists would let us know about unusual work units (size, length, etc.) so we will know that we don't have a machine problem of some sort or a bad work unit.
Otherwise, they might get aborted unnecessarily.
ID: 59056 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 6952
Credit: 20,843,205
RAC: 0
Message 59057 - Posted: 20 Nov 2018, 21:46:14 UTC - in response to Message 59054.  

I've got two 766s at 11% in 2 days... so I guess the end should come rather planned and soon than in few days right?


Yes, you may as well dump them.

*****************

On the bright side, for me: I've just uploaded a set of zip 144 and the restarts, (which are a little over 40 Megs), so one set of zippity dodahs to go. Which should be in about an hour and a half from now.

Now for the 2nd computer.
ID: 59057 · Report as offensive     Reply Quote
Profile JIM

Send message
Joined: 31 Dec 07
Posts: 1086
Credit: 19,554,876
RAC: 805
Message 59058 - Posted: 21 Nov 2018, 0:17:38 UTC
Last modified: 21 Nov 2018, 0:39:20 UTC

My problem is that I have 4 globals on 2 different machines that are more than 11 days in and have 19 days to go. Abort or not, that is the question.
ID: 59058 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 6952
Credit: 20,843,205
RAC: 0
Message 59059 - Posted: 21 Nov 2018, 0:51:11 UTC

I'd be inclined to hang in there.

My 3 on this computer finished OK, and are currently uploading.

This will be one of them in an hour or so, after the uploads and Reporting.
About an hour under 16 days, but that's with no other projects.
ID: 59059 · Report as offensive     Reply Quote
Profile JIM

Send message
Joined: 31 Dec 07
Posts: 1086
Credit: 19,554,876
RAC: 805
Message 59060 - Posted: 21 Nov 2018, 2:56:40 UTC - in response to Message 59059.  

I think I will stick with them and hope.
ID: 59060 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 6952
Credit: 20,843,205
RAC: 0
Message 59061 - Posted: 21 Nov 2018, 3:44:21 UTC

And now there's a new problem:

21/11/2018 12:39:49 PM | climateprediction.net | Server can't open database

Hopefully that's the last of the problems.
Except for repeating the whole cycle again in a few terabytes.
ID: 59061 · Report as offensive     Reply Quote
Profile JIM

Send message
Joined: 31 Dec 07
Posts: 1086
Credit: 19,554,876
RAC: 805
Message 59067 - Posted: 22 Nov 2018, 0:04:33 UTC - in response to Message 59053.  

[quote]SO ...
If anyone has any Global models, batches 766-770, that have only just started, or are still waiting, then I'd suggest that they be Aborted now, and save the electricity for the New Look Globals, Coming Soon to a Download Server near you. :)


I see that the global models from batches 766 – 770 are still in the hopper. I found that I had downloaded 4 more of them overnight. They were all _0 so they were first timers. They went to model heaven.
ID: 59067 · Report as offensive     Reply Quote
bernard_ivo

Send message
Joined: 18 Jul 13
Posts: 342
Credit: 9,943,922
RAC: 50
Message 59080 - Posted: 24 Nov 2018, 11:46:16 UTC - in response to Message 59067.  
Last modified: 24 Nov 2018, 12:00:57 UTC

I got one 766 and it is at 30% after six days, created on 4 Nov and this is its 2nd attempt. I kind of want to give it a chance, though Les suggested to abort them.

and I have two sam25s at 10% on 10th day. Expected to finish in 100 days ;) I've reduced the load to 2 full cores so let's see if I get some gain on the sam25s
ID: 59080 · Report as offensive     Reply Quote
bernard_ivo

Send message
Joined: 18 Jul 13
Posts: 342
Credit: 9,943,922
RAC: 50
Message 59127 - Posted: 3 Dec 2018, 19:57:54 UTC - in response to Message 59080.  

and I have two sam25s at 10% on 10th day. Expected to finish in 100 days ;) I've reduced the load to 2 full cores so let's see if I get some gain on the sam25s


No real difference btw 2 or 4 cores on my i5-2520M sec/Ts almost the same. And I still keep one global from 766 now at 61% after 14 days I will just leave it until it crashes as expected.
ID: 59127 · Report as offensive     Reply Quote
Eirik Redd

Send message
Joined: 31 Aug 04
Posts: 364
Credit: 114,089,091
RAC: 1,509
Message 59133 - Posted: 6 Dec 2018, 7:07:41 UTC - in response to Message 59127.  

Yeah -- those long-running sam models look like taking 6-12 weeks on a typical machine running 24/7 per core but less problems than some other long-runners -
Eh? I've got a few, will let them run until they die (don't think likely)

and I have two sam25s at 10% on 10th day. Expected to finish in 100 days ;) I've reduced the load to 2 full cores so let's see if I get some gain on the sam25s


No real difference btw 2 or 4 cores on my i5-2520M sec/Ts almost the same. And I still keep one global from 766 now at 61% after 14 days I will just leave it until it crashes as expected.

ID: 59133 · Report as offensive     Reply Quote
1 · 2 · 3 · Next

Message boards : Number crunching : 72 days for wah2_sam25?

©2019 climateprediction.net