| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Official Messages
Forum: News
Thread: Regarding ARP1 and MCM1 download issues since ARP1's launch on Monday Nov 4th, 2024 |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 159
|
|
| Author |
|
|
geophi
Advanced Cruncher U.S. Joined: Sep 3, 2007 Post Count: 102 Status: Offline Project Badges: 
|
All seems fine this morning, with two more ARP1 WUs probably finishing within the hour and nothing stuck on uploads or downloads this morning. Only strange thing is that two WUs that show the status as "no reply" are still running too which are well beyond their normal processing time, about 4-5x slower than would be usual on those two hosts. And they are resends that errored before, so not sure if there's something like a bad batch that is/was causing headaches too... Ralf @TCBF What generation are those WUs from? Just wondering if they're in the Extreme range. If you recall way back when IBM was managing this, some WUs in the extreme generations would fail/get stuck and they would have to half the time slice for those grid squares for a period to get them through the problem. |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 1950 Status: Offline Project Badges: 
|
All seems fine this morning, with two more ARP1 WUs probably finishing within the hour and nothing stuck on uploads or downloads this morning. Only strange thing is that two WUs that show the status as "no reply" are still running too which are well beyond their normal processing time, about 4-5x slower than would be usual on those two hosts. And they are resends that errored before, so not sure if there's something like a bad batch that is/was causing headaches too... Ralf @TCBF What generation are those WUs from? Just wondering if they're in the Extreme range. If you recall way back when IBM was managing this, some WUs in the extreme generations would fail/get stuck and they would have to half the time slice for those grid squares for a period to get them through the problem. The other is ARP1_0015263_143_2, is running for 25h, with estimated 58h remaining, while only 20% done. The first one is usually running a bit on the slow side, about 40-48h per WU (but does so reliably!), but the second one is really strange, as that is one of my most performing hosts, which returned previous ARP1 WUs in about 12-15h (while being used with InDesign and Photoshop), but that host isn't really used much this week at all... The first one is a rather old Xeon based Windows server, only lightly/moderately used still (probably going to be decommissioned when moving early next year), and don't ever recall any such issues back when ARP1 was still running during IBM's days (or the time it was running ARP1 under Krembil 2 years ago). As I mentioned, this machine is a very reliable machine... Ralf  [Edit 1 times, last edit by TPCBF at Nov 21, 2024 8:23:06 PM] |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 1950 Status: Offline Project Badges:
|
so how come that I could not download any MCM for a whole day now? How do you seriously expect someone to be able to give you an answer to that?What should be definitely clear is that this IS NOT a general WCG system issue, as there are plenty of people who have stated so far that the problems have disappeared and/or is now completely gone. And to make it clear, in case you didn't understand savas posts, there are no new ARP1 WUs send out for now, until the dust settles and all pending WUs have been returned with a valid quorum. But MCM1 WUs are flowing better than they have for sure the last three weeks since ARP1 was restarted, and I seem to get around the 1,000 WUs/day mark for the last 4 days now, across about 20 different hosts... Ralf |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7660 Status: Offline Project Badges:
|
I'm definitely seeing an improvement here. Downloads were first to improve ... Unfortunately, I do NOT see any improvement yet. I have been trying to get MCM tasks since yesterday, not a single one was downloaded  Definitely also seeing an improvement. We don't really have many files still waiting and MCM1 seems to be flowing pretty well. so how come that I could not download any MCM for a whole day now? Post about 50 lines or so from the beginning of your log and we may be able to determine what the hold up is for you. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 952 Status: Offline Project Badges: 
|
All seems fine this morning, with two more ARP1 WUs probably finishing within the hour and nothing stuck on uploads or downloads this morning. Only strange thing is that two WUs that show the status as "no reply" are still running too which are well beyond their normal processing time, about 4-5x slower than would be usual on those two hosts. And they are resends that errored before, so not sure if there's something like a bad batch that is/was causing headaches too... Ralf @TCBF What generation are those WUs from? Just wondering if they're in the Extreme range. If you recall way back when IBM was managing this, some WUs in the extreme generations would fail/get stuck and they would have to half the time slice for those grid squares for a period to get them through the problem. Cheers - Al. |
||
|
|
erich56
Senior Cruncher Austria Joined: Feb 24, 2007 Post Count: 295 Status: Offline Project Badges: 
|
I'm definitely seeing an improvement here. Downloads were first to improve ... Unfortunately, I do NOT see any improvement yet. I have been trying to get MCM tasks since yesterday, not a single one was downloaded Definitely also seeing an improvement. We don't really have many files still waiting and MCM1 seems to be flowing pretty well. so how come that I could not download any MCM for a whole day now? Post about 50 lines or so from the beginning of your log and we may be able to determine what the hold up is for you. Cheers this is what is shown in the event log, everytime the same text: 22.11.2024 08:51:30 | World Community Grid | Sending scheduler request: To fetch work. 22.11.2024 08:51:30 | World Community Grid | Requesting new tasks for NVIDIA GPU 22.11.2024 08:51:33 | World Community Grid | Scheduler request completed: got 0 new tasks 22.11.2024 08:51:33 | World Community Grid | No tasks sent 22.11.2024 08:51:33 | World Community Grid | No tasks are available for Mapping Cancer Markers 22.11.2024 08:51:33 | World Community Grid | No tasks are available for Smash Childhood Cancer 22.11.2024 08:51:33 | World Community Grid | Tasks for CPU are available, but your preferences are set to not accept them 22.11.2024 08:51:33 | World Community Grid | Tasks for AMD/ATI GPU are available, but your preferences are set to not accept them 22.11.2024 08:51:33 | World Community Grid | Tasks for Intel GPU are available, but your preferences are set to not accept them 22.11.2024 08:51:33 | World Community Grid | Project requested delay of 121 seconds I think no work for Smash Childhood Cancer is available at the moment, but why don't I get any MCM ??? |
||
|
|
maeax
Advanced Cruncher Joined: May 2, 2007 Post Count: 142 Status: Offline Project Badges:
|
22.11.2024 08:51:33 | World Community Grid | Scheduler request completed: got 0 new tasks
----------------------------------------22.11.2024 08:51:33 | World Community Grid | Tasks for CPU are available, but your preferences are set to not accept them Do you have GPU active? Only CPU-Tasks atm.
AMD Ryzen Threadripper PRO 3995WX 64-Cores/ AMD Radeon (TM) Pro W6600. OS Win11pro
|
||
|
|
Link64
Advanced Cruncher Joined: Feb 19, 2021 Post Count: 129 Status: Offline Project Badges: 
|
Another possibility is that an initial task for that workunit was released to a 32-bit system -- it that happens, all the wingmen have to follow suit. The 32-bit version of the ARP1 application seems to take between 30% and 50% longer to run when I get one... Also the 32-bit version of MCM is significantly slower, at least on my Ryzen 5700G. To avoid running 32-bit applications, use <no_alt_platform>1</no_alt_platform> in cc_config.xml. Keep in mind, that this is for all BOINC projects, so projects, which do not have 64-bit applications like for example Einstein's FGRP5 (only Windows), won't run as long as this setting is enabled. [Edit 1 times, last edit by Link64 at Nov 22, 2024 10:17:35 AM] |
||
|
|
Link64
Advanced Cruncher Joined: Feb 19, 2021 Post Count: 129 Status: Offline Project Badges:
|
22.11.2024 08:51:33 | World Community Grid | Tasks for CPU are available, but your preferences are set to not accept them This is the issue, you allow for WCG only GPU tasks and all available tasks are CPU only. There you need to change it in the profile used by your computer: https://www.worldcommunitygrid.org/ms/device/viewProfiles.do |
||
|
|
erich56
Senior Cruncher Austria Joined: Feb 24, 2007 Post Count: 295 Status: Offline Project Badges:
|
22.11.2024 08:51:33 | World Community Grid | Tasks for CPU are available, but your preferences are set to not accept them This is the issue, you allow for WCG only GPU tasks and all available tasks are CPU only. There you need to change it in the profile used by your computer: https://www.worldcommunitygrid.org/ms/device/viewProfiles.dothanks a lot for the hint - this was exactly it  |
||
|
|
|