The OpenIFS model batches require the option 'Leave non-GPU in memory while suspended' to be enabled under boincmgr -> Disk&Memory. This will prevent the task from frequently restarting and reduce the risk of task failure.

Ok, but that will leave tasks in memory for *all projects*. Let's see if my machine likes that....

It shouldn't really hurt anything. The tasks remain in memory for as long as the tasks are running, so unless you pause/suspend work a lot and try to do other things on the GPU that requires VRAM it shouldn't be an issue.

Ok, but that will leave tasks in memory for *all projects*. Let's see if my machine likes that....

I have been doing that for years, without any trouble. My machine now has 64 GigaBytes of RAM, though.
Right now, I am running no CPDN work, while allowing accumulated stuff from the last coupla weeks to upload.
I have been allowing 5 Oidf -ps tasks to run at a time along with other projects. Once those 5 Oidf-ps tasks were running along with one hadsm4 and they all worked out fine.

$ free -hw
              total        used        free      shared     buffers       cache   available
Mem:           62Gi       3.9Gi        24Gi       124Mi       148Mi        34Gi        57Gi
Swap:          15Gi       465Mi        15Gi

---

Finally CPDN has used the BOINC push notification:

climateprediction.net: Request to volunteers to please enable: 'Leave non-GPU tasks in memory'
The OpenIFS model batches require the option 'Leave non-GPU in memory while suspended' to be enabled under boincmgr -> Disk&Memory. This will prevent the task from frequently restarting and reduce the risk of task failure.
Thu 05 Jan 2023 04:54:57 PM EET more... 

It could've been also used for the missing 32 bit libs as well :) But so far so good

Ok, but that will leave tasks in memory for *all projects*. Let's see if my machine likes that....

It's a good idea to make sure there is sufficient memory anyway, you don't want the OS to start swapping to fulfill memory requirements of the tasks. This is particularly true of the OpenIFS (oifs) tasks, they need ~ 5Gb RAM each to run and will crawl if the system has to start swapping.

Finally CPDN has used the BOINC push notification:

Finally is correct.

It could've been also used for the missing 32 bit libs as well

Couldn't agree more.

I wanted to know if this application will be there for the Windows platform in the future?

Finally CPDN has used the BOINC push notification:

Finally is correct.

It could've been also used for the missing 32 bit libs as well

Couldn't agree more.

I thought it was for the third blurb on the front page back in 2019.
https://www.cpdn.org/cpdnboinc/index.php

It's a good idea to make sure there is sufficient memory anyway, you don't want the OS to start swapping to fulfill memory requirements of the tasks. [the system] will crawl if the system has to start swapping.

Amen!

In the very old days. it was suggested to have twice as much disk space for swapping as there was RAM in the system. But when the last of my 32-bit machines had 8 Gigabytes of RAM, I found that it was pointless because if it started swapping, the system response got so bad that I fixed the swapping problem long before I used up the swap space. So I cut it down to make the swap space equal to the RAM size. And now I have 64 GigaBytes of RAM in the machine and am thinking about doubling it if these Oifs programs start to need it. The machine came with 16 GigaBytes of swap space and I have never used more than 1/4 of it; usually much less. At the moment there are 5 oifs_43r3_ps tasks running, 4 WCG tasks, one Einstein, two MilkyWay all running and only 550 MegaBytes of swap space being used.

Back in those old days, DEC was a major manufacturer of mini-computers. At one point, they had a hardware department designing a head-per-track swapping disk, eliminating the seek time (but retaining the rotational latency time). But another department was designing solid state memory to replace the magnetic core memory they had been using. It turned out that in addition to being faster, it was so much cheaper that instead of swapping you just put more RAM in the machine.

---

 vi /var/lib/boinc-client/global_pref.xml

<leave_apps_in_memory>1</leave_apps_in_memory>

 service boinc-client restart

People who don't speak English but machine languages. This is for Linux/Debian-ish translation.

I hope to finally be able to get some task again? Seems like I`m never able to get workunits.

sorry for ignorance: how do i enable the requested command in windows 10?

---

The only work currently available is for Linux and the message to enable leave tasks in memory is meant for those users who are able to process these Linux tasks. Windows work is rare although it might become more frequent in the future but only via VBox, not native Windows work. It might happen this year but we'll see. For Windows users interested in doing Linux work, check out WSL2 (Windows Subsystem for LInux). It's part of Windows (10 and 11) and is a very light way to virtualize Linux on Windows.

To enable the leave in memory while suspended feature in Windows, go to Options -> Computing preferences -> Disk and memory and look for it in the Memory section.

thanx for your help

---

I can't download task.
My machine Windows 10 has RAM 64GB.

Indeed. I just rejoined after some tie. I see no project data coming at all. Is the climate project still active?

Indeed. I just rejoined after some tie. I see no project data coming at all. Is the climate project still active?

Very much so. About 65,000 workunits went out to Linux machines over the last couple of months, with more to come. I believe CPDN are preparing some Windows work-units though I have no timescale for when they will go out.

I'm running ubuntu linux.

I'm running ubuntu linux.

I am running Red Hat Enterprise Linux release 8.7 (Ootpa).
I do not feel I am getting enough work from Climate-Prediction, but that is just because they are not sending any right now except for some left-overs that failed for the previous users for various reasons.

Most recently, I have completed these (and no failures):

22306007 	12184070 	6 Feb 2023, 3:24:45 UTC 	6 Feb 2023, 18:43:23 UTC 	Completed 	54,452.23 	53,778.93 	0.00 	OpenIFS 43r3 Perturbed Surface v1.05
x86_64-pc-linux-gnu
22305925 	12186447 	5 Feb 2023, 22:23:58 UTC 	6 Feb 2023, 13:58:54 UTC 	Completed 	56,008.96 	55,378.24 	0.00 	OpenIFS 43r3 Perturbed Surface v1.05
x86_64-pc-linux-gnu
22305507 	12173514 	4 Feb 2023, 12:25:23 UTC 	5 Feb 2023, 3:55:04 UTC 	Completed 	55,176.56 	54,089.42 	0.00 	OpenIFS 43r3 Perturbed Surface v1.05
x86_64-pc-linux-gnu
22305076 	12188450 	2 Feb 2023, 0:25:27 UTC 	2 Feb 2023, 16:25:25 UTC 	Completed 	56,709.73 	55,985.71 	2,642.00 	OpenIFS 43r3 Perturbed Surface v1.05
x86_64-pc-linux-gnu
22305044 	12169787 	1 Feb 2023, 18:24:00 UTC 	2 Feb 2023, 11:24:42 UTC 	Completed 	58,063.42 	57,280.70 	2,642.00 	OpenIFS 43r3 Perturbed Surface v1.05
x86_64-pc-linux-gnu
22305045 	12172513 	1 Feb 2023, 18:24:00 UTC 	2 Feb 2023, 11:24:42 UTC 	Completed 	58,140.38 	57,357.18 	2,642.00 	OpenIFS 43r3 Perturbed Surface v1.05
x86_64-pc-linux-gnu

---

Yay! A Climatological run just showed up, 014050100 hq0f 0399 983 12200043 123 oifs_43r3_ps 1