climateprediction.net home page
What ever it was modeling it did it fast.

What ever it was modeling it did it fast.

Message boards : Number crunching : What ever it was modeling it did it fast.
Message board moderation

To post messages, you must log in.

AuthorMessage
lazlo_vii

Send message
Joined: 11 Dec 19
Posts: 64
Credit: 1,971,833
RAC: 5,265
Message 61776 - Posted: 22 Dec 2019, 19:43:28 UTC

https://www.cpdn.org/result.php?resultid=21827977

I am not sure that work unit will have valid results. As the Myth Busters used to say: "Failure is always an option!"
ID: 61776 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7230
Credit: 23,154,247
RAC: 5,416
Message 61777 - Posted: 22 Dec 2019, 19:54:09 UTC - in response to Message 61776.  

Nothing wrong with that one.
The computer could do with more memory though.
ID: 61777 · Report as offensive     Reply Quote
lazlo_vii

Send message
Joined: 11 Dec 19
Posts: 64
Credit: 1,971,833
RAC: 5,265
Message 61778 - Posted: 22 Dec 2019, 20:18:58 UTC - in response to Message 61777.  

Nothing wrong with that one.
The computer could do with more memory though.


That container was set up with a profile that gave it 50% of the system's total RAM. I later changed it to have access to all 32GB. That ways a few days ago and I am not sure why the specs haven't updated on the web site.
ID: 61778 · Report as offensive     Reply Quote
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 2722
Credit: 3,327,142
RAC: 1,283
Message 61793 - Posted: 24 Dec 2019, 11:02:08 UTC - in response to Message 61778.  

Nothing wrong with that one.
The computer could do with more memory though.


That container was set up with a profile that gave it 50% of the system's total RAM. I later changed it to have access to all 32GB. That ways a few days ago and I am not sure why the specs haven't updated on the web site.


As a result of requests from moderators, there have been a few changes to information on the website recently. Really to cope with all the Linux task types (including openifs which will appear at some point in the future) machines need to have 4GB/core in use. My desktop if running N216 tasks, I am limiting to one out of two cores but I think that is more a cache memory than a RAM issue. The openifs tasks in testing used over 5GB RAM/task but that was peak. several tasks at once OK on 4GB/core because they tend not to peak at once.
ID: 61793 · Report as offensive     Reply Quote
lazlo_vii

Send message
Joined: 11 Dec 19
Posts: 64
Credit: 1,971,833
RAC: 5,265
Message 61795 - Posted: 24 Dec 2019, 11:31:39 UTC - in response to Message 61793.  
Last modified: 24 Dec 2019, 11:36:01 UTC

Nothing wrong with that one.
The computer could do with more memory though.


That container was set up with a profile that gave it 50% of the system's total RAM. I later changed it to have access to all 32GB. That ways a few days ago and I am not sure why the specs haven't updated on the web site.


As a result of requests from moderators, there have been a few changes to information on the website recently. Really to cope with all the Linux task types (including openifs which will appear at some point in the future) machines need to have 4GB/core in use. My desktop if running N216 tasks, I am limiting to one out of two cores but I think that is more a cache memory than a RAM issue. The openifs tasks in testing used over 5GB RAM/task but that was peak. several tasks at once OK on 4GB/core because they tend not to peak at once.


In a "worst case" scenario four or more threads peaking at the same time could cause my server to start swapping. It does more than just crunch. It is has 4 other full time "low CPU, low RAM, low I/O" tasks and one other full time "low CPU, variable RAM, variable I/O" task. So that makes me wonder how long these peaks might last and how much I/O they might need to sustain full speed calculation. I don't expect any answers at this point. I am just "typing out loud" as it were. The new tasks will be released on time, when ever that may be. For now there are still over 2000 N144 models and (at least from what I can guess) less than 100 of us working on them 24/7.
ID: 61795 · Report as offensive     Reply Quote
wolfman1360

Send message
Joined: 18 Feb 17
Posts: 71
Credit: 4,941,141
RAC: 4,855
Message 61808 - Posted: 25 Dec 2019, 6:55:44 UTC

Is there any plan to port openifs over to Windows as well, or will this be a Linux exclusive?
Seems there sadly are not many Linux crunchers out there, or at least connected to this project in comparison to others.
ID: 61808 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7230
Credit: 23,154,247
RAC: 5,416
Message 61809 - Posted: 25 Dec 2019, 7:04:47 UTC - in response to Message 61808.  
Last modified: 25 Dec 2019, 7:05:13 UTC

There's been no mention of OpenIFS since the big run back in June.

Don't worry too much about the lack of Linux users.
There's been very little work for us for most of the year, so when the current lot are gone, it may be well into next year before we get more.
ID: 61809 · Report as offensive     Reply Quote
Jean-David Beyer

Send message
Joined: 5 Aug 04
Posts: 381
Credit: 3,690,501
RAC: 324
Message 61810 - Posted: 25 Dec 2019, 13:16:35 UTC - in response to Message 61808.  

Seems there sadly are not many Linux crunchers out there, or at least connected to this project in comparison to others.


I have been signed up for CPDN for ages, running only Linux. If memory serves, I went about a whole year without getting any CPDN work.
Presently, I am running one hadam4h and two hadam4 work units. Also one WCG work unit.
My normal intention is to run 50% CPDN, 25% WCG, and 12.5% each of Seti@home and rosetta@home. I enable these last two only when I have idle processor cores and cannot get either CPDN or WCG work units. When the latest hadam4 and hadam4h Linux units started coming out, I was running 100% CPDN for a while.

I do not know if you would count me as a Linux cruncher during the year when I crunched no CPDN work units, because I did not get any.
ID: 61810 · Report as offensive     Reply Quote
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 2722
Credit: 3,327,142
RAC: 1,283
Message 61813 - Posted: 25 Dec 2019, 14:38:09 UTC

Is there any plan to port openifs over to Windows as well, or will this be a Linux exclusive?


From what I remember reading on the openifs site, it is for Linux and Mac only so, major work to try and port it to Windows. A major advantage over the Met Office code is that it is open source. It is also 64bit so while it might exclude a few computers, it will not suffer from the missing libs issues.
ID: 61813 · Report as offensive     Reply Quote

Message boards : Number crunching : What ever it was modeling it did it fast.

©2020 climateprediction.net