| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Mapping Cancer Markers Forum
Thread: Large amounts of Invalid Results |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 15
|
|
| Author |
|
|
KerSamson
Master Cruncher Switzerland Joined: Jan 29, 2007 Post Count: 1684 Status: Offline Project Badges: 
|
After a mainboard and CPU update last year, I did experience a similar situation: no error, just invalid results.
----------------------------------------After loosing several weeks of investigation, I ran dry the machine and detached the machine from WCG. After reinstalling Boinc - without deleting client_state.xml - and reboot, I did re-attach WCG. Since then everything is running fine without any invalid WUs any more. Frankly, I do not know why the WUs computed after a board/CPU change become invalid. In particular, because I did it in the past several time, without any trouble. Now, I know what I have to do in case of hardware update. Good luck, Yves |
||
|
|
ThreadRipper
Veteran Cruncher Sweden Joined: Apr 26, 2007 Post Count: 1324 Status: Offline Project Badges: 
|
I am happy you resoolved it. RAM issue was my guess above as well since I had similar problems. Often it is not incorrectly binned RAM kit itself but rather the Motherboard or Memory controller in the CPU that can't keep up.
----------------------------------------Ryzen 1000- and 2000-series are known to be picky about memory. My RAM was not on the QVL so I expected some trouble. Things to try might be to increase SOC voltage a bit (I wouldn't go above 1.13V or so for 24/7 since memeory controller degradation may occur with higher SOC voltages - as in, 6 months ahead one may suddenly get instability without having changed anything). Anyway, happy to hear you got it working!  Join The International Team: https://www.worldcommunitygrid.org/team/viewTeamInfo.do?teamId=CK9RP1BKX1 AMD TR2990WX @ PBO, 64GB Quad 3200MHz 14-17-17-17-1T, RX6900XT @ Stock AMD 3800X @ PBO AMD 2700X @ 4GHz |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: 
|
Ryzen 1000- and 2000-series are known to be picky about memory. My RAM was not on the QVL so I expected some trouble. Things to try might be to increase SOC voltage a bit (I wouldn't go above 1.13V or so for 24/7 since memeory controller degradation may occur with higher SOC voltages - as in, 6 months ahead one may suddenly get instability without having changed anything). There is a problem with some of the ASRock AM4 motherboards. They don't set the voltage right for some memory modules. I have had to change it (in the BIOS) for several of them. Here is my post on the ASRock forum about it. http://forum.asrock.com/forum_posts.asp?TID=7...b350m-pro4-freezing#71526 [Edit 1 times, last edit by Jim1348 at Mar 7, 2020 6:46:28 PM] |
||
|
|
ThreadRipper
Veteran Cruncher Sweden Joined: Apr 26, 2007 Post Count: 1324 Status: Offline Project Badges:
|
Ryzen 1000- and 2000-series are known to be picky about memory. My RAM was not on the QVL so I expected some trouble. Things to try might be to increase SOC voltage a bit (I wouldn't go above 1.13V or so for 24/7 since memeory controller degradation may occur with higher SOC voltages - as in, 6 months ahead one may suddenly get instability without having changed anything). There is a problem with some of the ASRock AM4 motherboards. They don't set the voltage right for some memory modules. I have had to change it (in the BIOS) for several of them. Here is my post on the ASRock forum about it. http://forum.asrock.com/forum_posts.asp?TID=7...b350m-pro4-freezing#71526 Very interesting! I usually run my RAM at 1.4V or so since I try to lower the timings, and after a while one starts to notice which crashes are RAM related and which a CPU related. But yeah, when the board is acting up, there's a good risk of confusion. Join The International Team: https://www.worldcommunitygrid.org/team/viewTeamInfo.do?teamId=CK9RP1BKX1 AMD TR2990WX @ PBO, 64GB Quad 3200MHz 14-17-17-17-1T, RX6900XT @ Stock AMD 3800X @ PBO AMD 2700X @ 4GHz |
||
|
|
Piri1974
Cruncher Joined: Jun 3, 2018 Post Count: 24 Status: Offline Project Badges: 
|
Maybe you could try running with half the RAM modules?
It is possible that then you will be able to run them at the speed in their XMP profile... I have seen on several computers that filling up all RAM sockets is not the very best idea unless you really have to. Example: If you have 4 RAM sockets, using just 2 of them seems to make it easier for the memory controller to run at the XMP profile speed without issue. Just make sure you use 1 of each RAM channel so that you continue to run in dual channel mode. This solution has solved my RAM speed problem twice already. We always seem to forget that the fastest official DDR4 speed is 2133Mhz if I am not mistaken. And if you use only BOINC, you don't need a lot of RAM anyway. Having more RAM than you need will not shorten your average run times. Cheers; Carl Philip |
||
|
|
|