 |
|
|
61)
Message boards :
Number crunching :
Computation Errors
(Message 65666)
Posted 19 Jul 2022 by  geophi
Post: A few weeks ago I decided to test and max out my Ryzen 5900X (12C/24T) with 50GB RAM dedicated to WSL2 Ubuntu 22.04. Ran 24 HadAM4 N144s at the same time and they all finished without errors. The CPU has 64MB of L3 cache so about 2.6MB per task available on average. They all got done in about 20 days so about 1.2 tasks per day average, not a bad throughput I thought. I'm assuming you are talking about the 13 month HADAM4 N144 tasks. Running 5 at a time on my 5600X, each task takes about 4 days, so in 20 days it would finish about 25. I really think that you should test this with no use of the SMT threads, running 12 at a time. My guess is that total model throughput would be considerably higher than what happened running 24 at a time. Now I realize that the comparison of my PC with yours is not apples to apples as you are running these in a VM, with the associated performance penalty, and my 5600X is running these natively in Linux. Also, it was running at 4.4 to 4.5 GHz and I'm sure yours is throttling more running that many. But it's been a long time since running a significant number of models above the total number of cores resulted in more total model throughput. Perhaps with something like hadcm3s (if it were again to be released for Linux), using some of the SMT threads would increase throughput, but I doubt the HADAM4 N144 models would see much, if any, by running more tasks than cores. |
|
62)
Message boards :
Number crunching :
HadSM4 Error when completed and Uploading
(Message 65654)
Posted 18 Jul 2022 by geophi
Post: @Jean-David The install of his Fedora 36 did not include the libnsl file that is apparently needed. This results in upload failures (for some reason). He has installed this file now and the 6th and final file did upload correctly. However, since the other five monthly zip files did not go up (before he installed libnsl), boinc marked the results as errors. @Conan Looking at the stderr on your task webpages, the 6th zip must have been uploaded successfully, but the other 5 monthly zips weren't. So boinc marked the result as an error. Now that libnsl is installed, you shouldn't have any more errors of this type. |
|
63)
Message boards :
Number crunching :
HadSM4 Error when completed and Uploading
(Message 65649)
Posted 17 Jul 2022 by geophi
Post: This library is provided for backwards compatibility only; applications should use libnsl2 instead to gain IPv6 support. @Dave That was the error I got a couple years back when I installed Fedora 32? to try to help troubleshoot a problem a user was having with that distribution. That was when I sent you the instructions for updating the post on 32bit libraries to include Fedora and this libnsl in the 32bit library instruction post. @Conan I believe that error crops up in upload transfers, so if that error resulted in one or more uploads from a task not making it to the servers, the task will likely be marked as an error in the database. |
|
64)
Message boards :
Number crunching :
Task completed, but not all trickles acknowledged yet. Normal?
(Message 65610)
Posted 2 Jul 2022 by geophi
Post: Not over worried as I have got the credit for the work. Presumably in about 12months time it will come round again for reissue if the task set hasn't been pulled. The only way it can be reissued faster would be if you set no new work for cpdn in boinc manager let your cpdn tasks finish detach from cpdn, and reattach That will set the task state to abandoned, and it will be reissued immediately. |
|
65)
Message boards :
Number crunching :
Task completed, but not all trickles acknowledged yet. Normal?
(Message 65607)
Posted 30 Jun 2022 by geophi
Post: Looking at the timing of the uploads that would seem to be a reasonable supposition. Will it rectify itself? Unfortunately, it didn't for me. |
|
66)
Message boards :
Number crunching :
Task completed, but not all trickles acknowledged yet. Normal?
(Message 65603)
Posted 29 Jun 2022 by geophi
Post: Similar problem but in reverse. All trickles loaded, out and restart zips sent but task showing as not complete. Also showing full credit! I've infrequently had this happen, but not for a long time. I noted it happening when a task is reported during a particularly busy server time, like when the trickles are being counted for the weekly credit run. At least that's my guess as a possible factor in this issue. |
|
67)
Message boards :
Cafe CPDN :
World Community Grid mostly down for 2 months while transitioning
(Message 65586)
Posted 22 Jun 2022 by geophi
Post: The website and forums are back up at WCG. BOINC on the servers has not been restarted yet, but should be "soon". So no tasks to download at this time. We'll see what "soon" means. A few days? Longer? |
|
68)
Message boards :
Number crunching :
New work Discussion
(Message 65511)
Posted 8 Jun 2022 by geophi
Post: When did Boinc stop being 32bit? I ran CPDN on Windows ok a couple of years back, I had whatever the latest Boinc was then. I thought people over at Boinc were putting off going to 64bit because of their own problems. They haven't updated the 32bit windows Boinc for 3.5+ years and the 32bit linux version for 7.5+ years. I guess they think if you are running older operating systems, you can download older versions of boinc and make do. The problem came about last fall due to the certificates being installed with older version of Windows boinc expiring. They came out with a new 64 bit version with a newer certificates list, but no new 32 bit version. They do have a certificates file to download at the top of this page: https://boinc.berkeley.edu/download_all.php if one is still running the older Windows versions. |
|
69)
Message boards :
Number crunching :
New Work Announcements
(Message 65500)
Posted 7 Jun 2022 by geophi
Post: Batch 931 is available for Linux PCs. 2450 Work Units of hadam4h N216 models, running 5 model months. |
|
70)
Message boards :
Number crunching :
New work Discussion
(Message 65481)
Posted 3 Jun 2022 by geophi
Post: We tested some WAH2 tasks for ANZ back around April 1st. They were successful, and the main investigator from New Zealand was going to look at the output and get back to Sarah if things looked okay to send out main site batches. No communications since then. I will query if this is still the plan. |
|
71)
Message boards :
Number crunching :
New work Discussion
(Message 65459)
Posted 19 May 2022 by geophi
Post: I'm getting the "Model crashed: INANCLA: Error opening file " error in stderr as well, but the tasks finish successfully and upload all zips and trickles. I brought that error up on the dev site and they said it wasn't an issue and everything was returned as expected. |
|
72)
Message boards :
Number crunching :
New work Discussion
(Message 65453)
Posted 18 May 2022 by geophi
Post: Dave, do you also have an Intel PC? If so, have you tried setting up a macOS VM on it? For me, I couldn't get it to work on Ryzen 5900X but did on i7-4790 (both Windows 10). Andrey, Both SolarSonyk and I have had success on AMD Ryzens with Mac guests on a Linux host. Whatever problem you are having with VirtualBox in Windows on AMD, isn't translating to KVM on a Linux host. |
|
73)
Questions and Answers :
Windows :
macOS Mojave installation on Windows 10 with VirtualBox
(Message 65393)
Posted 3 May 2022 by geophi
Post: Just released...656 HADCM3S work units in batch 930 for Intel Macs. |
|
74)
Questions and Answers :
Unix/Linux :
Running 32-bit MacOS Tasks on Linux with KVM
(Message 65392)
Posted 3 May 2022 by geophi
Post: Just released...656 HADCM3S work units in batch 930 for Intel Macs. |
|
75)
Message boards :
Number crunching :
New work Discussion
(Message 65391)
Posted 3 May 2022 by geophi
Post: Just released...656 HADCM3S work units in batch 930 for Intel Macs. |
|
76)
Message boards :
Number crunching :
New Work Announcements
(Message 65390)
Posted 3 May 2022 by geophi
Post: Just released...656 HADCM3S work units in batch 930 for Intel Macs. |
|
77)
Message boards :
Cafe CPDN :
Climate model code is so outdated, MIT starts from scratch
(Message 65369)
Posted 14 Apr 2022 by geophi
Post: Typical Register article. I'd be interested in hearing what a long list of respected climate scientists think of this article, and the claims in the proposal. That was what I saw too. Many models around the world are using higher resolution than they stated. It will be interesting to see if anything in forecast improvement will come out of this effort. Perhaps so, but the claims, like in many research proposals, are likely exaggerated or overly optimistic. |
|
78)
Message boards :
Cafe CPDN :
Climate model code is so outdated, MIT starts from scratch
(Message 65366)
Posted 14 Apr 2022 by geophi
Post: Typical Register article. I'd be interested in hearing what a long list of respected climate scientists think of this article, and the claims in the proposal. |
|
79)
Message boards :
Number crunching :
No work for Windows OR Linux?!
(Message 65344)
Posted 12 Apr 2022 by geophi
Post: I did say over on the BOINC boards that the 32bit libraries needed to be installed though not having them doesn't stop you downloading work, it just makes it all crash. Dave, I've had both happen. It won't download work, or it immediately crashes if 32 bit libraries are not installed. May depend on the distribution or other things that are installed before boinc and cpdn. |
|
80)
Message boards :
Number crunching :
New Work Announcements
(Message 65329)
Posted 6 Apr 2022 by geophi
Post: Batch 929 just released. Another 2568 work units of HADAM4 N144 model tasks for Linux. |
©2024 climateprediction.net