Welcome to MilkyWay@home

Posts by Septimus

21) Message boards : Number crunching : Validation Pending too many tasks (Message 74620)
Posted 2 Nov 2022 by Septimus
Post:
Thanks for your comments. I am not sure why that WU is being bounced around, it used 41,000 CPU seconds, I ran it on 8 CPUs in around 5,600 seconds. I have had much bigger ones validated.


Just got validated at last !
22) Message boards : Number crunching : Validation Pending too many tasks (Message 74619)
Posted 2 Nov 2022 by Septimus
Post:
Thanks for your comments. I am not sure why that WU is being bounced around, it used 41,000 CPU seconds, I ran it on 8 CPUs in around 5,600 seconds. I have had much bigger ones validated.
23) Message boards : Number crunching : Validation Pending too many tasks (Message 74615)
Posted 1 Nov 2022 by Septimus
Post:
I am beginning to wonder if some long run Nbody WU’s will ever get validated. I have one that I completed, throughout its life back in September it has been Aborted twice, failed to start on time 3 times and is now back “in progress “. Have to wonder how many more there are like this.
24) Message boards : Number crunching : Daily graphs of server_status (Message 74611)
Posted 31 Oct 2022 by Septimus
Post:
I am just winding down on 2 Nbody where they say they 4 hours left. Actual run time bears no resemblance to estimated run time.
25) Message boards : Number crunching : Daily graphs of server_status (Message 74593)
Posted 29 Oct 2022 by Septimus
Post:
My last long Nbody 30 days and counting, yet to be validated.
26) Message boards : Number crunching : Daily graphs of server_status (Message 74585)
Posted 28 Oct 2022 by Septimus
Post:
Seems very reluctant to drop below 5 Million waiting for validation.
27) Message boards : Number crunching : Daily graphs of server_status (Message 74564)
Posted 25 Oct 2022 by Septimus
Post:
Personally I am still of the view that WU generation should stop, and run the backlog down. The memory was the same before the Nbody apparent issue. Separation run times are largely the same , Nbody have increased by at least 20 times. I think what could happen is more memory and a faster processor could make things worse until the application problem is located.
28) Message boards : News : Server Issues (Message 74562)
Posted 25 Oct 2022 by Septimus
Post:
Main components are down, shown on server status as Not Running.
29) Message boards : Number crunching : Daily graphs of server_status (Message 74552)
Posted 24 Oct 2022 by Septimus
Post:
...
Tom already said the problem is not enough memory in the current Server but the IT people are in charge of moving the stuff over to the new Server they already have ready and waiting

Hmmm, waiting for what?


Presumably the long running Nbody WU’s have exacerbated the problem ?
30) Message boards : Number crunching : Daily graphs of server_status (Message 74548)
Posted 24 Oct 2022 by Septimus
Post:
Maybe it’s time to stop producing new WU’s and get the Q down to a manageable level, it clearly is having adverse effects at present.
31) Message boards : Number crunching : Daily graphs of server_status (Message 74527)
Posted 21 Oct 2022 by Septimus
Post:
Brilliant….maybe someone will explain at least what changes were made to Nbody on or around 9th September.


I presume some explaination is here https://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=4924&postid=74351#74351


Thanks for that.
32) Message boards : Number crunching : Daily graphs of server_status (Message 74525)
Posted 21 Oct 2022 by Septimus
Post:
Brilliant….maybe someone will explain at least what changes were made to Nbody on or around 9th September.
33) Message boards : Number crunching : Validation Pending too many tasks (Message 74520)
Posted 21 Oct 2022 by Septimus
Post:
It would be nice to know the breakdown of the waiting for validation backlog between the applications. My guess is that Nbody is growing exponentially since the changes in profile from September. My small separation backlog has at last got to zero after 3 weeks. Whether the inconclusive Nbody’s will ever validate I don’t know. Have moved onto to other projects in the interim.
34) Message boards : Number crunching : Validation Pending too many tasks (Message 74519)
Posted 21 Oct 2022 by Septimus
Post:
It would be nice to know the breakdown of the waiting for validation backlog between the applications. My guess is that Nbody is growing exponentially since the changes in profile from September. My small separation backlog has at last got to zero after 3 weeks. Whether the inconclusive Nbody’s will ever validate I don’t know. Have moved onto to other projects in the interim.
35) Message boards : Number crunching : Daily graphs of server_status (Message 74513)
Posted 20 Oct 2022 by Septimus
Post:
Thanks very useful…..
36) Message boards : Number crunching : Daily graphs of server_status (Message 74512)
Posted 20 Oct 2022 by Septimus
Post:
Thanks very useful…..
37) Message boards : Number crunching : Validation Pending too many tasks (Message 74504)
Posted 20 Oct 2022 by Septimus
Post:
As far as I can tell the current problem with validation did not start until early September when the profile of Nbody jobs changed from a few minutes to an unspecified number of hours, even days, ever since then the total waiting for validation has escalated. All the Nbody jobs I did have been aborted at least twice before I got them. Whether they will ever validate I don’t know. The coincidence between increasing Nbody run times and the validation backlog surely needs checking out ?
38) Message boards : Number crunching : Daily graphs of server_status (Message 74501)
Posted 19 Oct 2022 by Septimus
Post:
Thanks again Kiska would it be possible to make that the daily graph please.
39) Message boards : Number crunching : Daily graphs of server_status (Message 74500)
Posted 19 Oct 2022 by Septimus
Post:
Thanks for that Kiska most informative. There seems to be a good correlation to me , especially latterly.
40) Message boards : Number crunching : Daily graphs of server_status (Message 74496)
Posted 19 Oct 2022 by Septimus
Post:
Is it possible to track 2 items on a graph ..would be interesting to see Nbody average run time and total waiting for validation on the same graph. My own, probably misguided view is things have not been the same since we got these very long Nbody Simulation WU’s.

I have messed around with the graphs and going back 90 days there seems to be a massive spike in Nbody average run time that coincides with the growth of waiting for validation.Average Nbody run time is around 15 times higher than it used to be. I may be misguided in thinking the two are linked.


Previous 20 · Next 20

©2024 Astroinformatics Group