Page 8 of 8
Re: Low PPD on some WUs
Posted: Mon Apr 18, 2016 5:38 pm
by Bryman
Well, only had time to run it for 40min before I left (guess I could of left it going) but anyway... no errors or anything, ran 160 times
Re: Low PPD on some WUs
Posted: Mon Apr 18, 2016 9:02 pm
by Bryman
Okay... this Intelburntest thing isn't working very well
I have my processor at 4.6ghz and it's been 15 minutes and it hasn't had an error... folding@home would of errored almost instantly at that frequency
Although the temperature is getting up to 95°C even with fans turned up.. voltage is about 1.3
Re: Low PPD on some WUs
Posted: Mon Apr 18, 2016 9:11 pm
by 7im
Re: Low PPD on some WUs
Posted: Mon Apr 18, 2016 9:56 pm
by Bryman
Even ran at 4.7ghz... maybe I just didn't have enough voltage when I ran folding@home or something
Also it's wrong when it says it gets the cpu hotter than prime 95
The Gromacs cpu tester got the temp up to 93 and prime 95 got it up to 104
This was at 1.375 volts 4.7ghz
Not sure exactly how much voltage I need for 4.7ghz... I think it's about 1.35 to be stable.. maybe even a little more
4.5ghz only requires 1.225...
Ran at 4.7ghz for over 30 minutes, ran with no issues... I'm thinking the voltage wasn't high enough when I tried 4.6 with folding@home
Re: Low PPD on some WUs
Posted: Mon Apr 18, 2016 11:20 pm
by Joe_H
How much floating point calculations do you expect to see used by Prime 95? I believe it is mostly doing integer calculations. That is why it is not comparable to folding, the F@H CPU cores use mostly floating point calculations done in a different section of the arithmetic unit in the CPU cores.
Re: Low PPD on some WUs
Posted: Mon Apr 18, 2016 11:32 pm
by Bryman
Yes, I didn't say prime 95 put more stress on the CPU, I simply said that it made the processor hotter
Gromacs got the core temp up to 93, while Prime 95 got it up to 104
But 4.7ghz seemed to run fine for 30 minutes, I know that's not nearly long enough to tell if it's stable or not but it'll at least catch anything big
Re: Low PPD on some WUs
Posted: Tue Apr 19, 2016 12:09 am
by 7im
To rephrase what Joe_H said, if your overclock settings pass the StressCPU test, your folding client is much less likely to fail than if you go by whatever settings are passible in Prime 95. Noting that Prime 95 may have improved its ability to heat up a chip (since StressCPU was posted in 2007 and hasn't changed), but it's not quite an apples to apples comparison. The statement about Prime95 was accurate at the time. Use whatever works best for you.
Re: Low PPD on some WUs
Posted: Tue Apr 19, 2016 4:47 am
by Bryman
Said the cpu WU crashed again.. only had my processor clocked to 4.2ghz
I did, however, make my memory timings tighter... not sure if that's what caused it
It's stable in prime 95 and linpack, but they don't do crap... I ran prime 95 for about an hour total (about 10 minutes 6 times) and my computer was still crashing about once every 30 minutes... didn't matter if I was running prime 95, folding, gaming, or browsing the web... it would just randomly crash on me about every 30 minutes
Basically my computer just kept shutting off on me... sometimes it would freeze for couple seconds and then shut off and sometimes it would just turn off... One time when I was playing a game the game closed then a split second later the computer froze and then a couple seconds later the computer shut off
I think I had it at 10-10-10 and that was what was causing the crashes... when I set it to 10-11-10 it seemed to stop crashing
Right now I have it set to 2133mhz 10-11-10-18 with 140 refresh cycle and 1 command rate at 1.47 volts, pretty sure the voltage is high enough.. I got errors at 1.42 but I ran it for about 10 minutes at 1.43 and got no errors... so 1.47 should be plenty
I think it's stable down to about 14 maybe even 13 on the last number but no point in going that low... computer booted at 12 but didn't boot at 11
04:30:02:WARNING:WU03:FS00:FahCore returned an unknown error code which probably indicates that it crashed
04:30:02:WARNING:WU03:FS00:FahCore returned: UNKNOWN_ENUM (-1073741783 = 0xc0000029)
It started back at a checkpoint, so I'll see if it crashes again
Well, it seems to have made it a few percent past when it crashed and is still going, so dunno
So... 10-11-10-13 and 15 seconds into prime 95 with blend and computer shut off
10-11-10-14 and with blend it ran fine for 10 minutes, and then I set the ram to 10000mb instead of the default 1600 and computer shut off within 30 seconds
10-11-10-15 12000mb computer froze after 2 minutes
10-11-10-16 12000mb fine for 10+ minutes
So... I think I'll do 10-11-10-18... should be pretty stable
But my ram should of been stable, I think.. so not sure why the CPU WU crashed
Re: Low PPD on some WUs
Posted: Tue Apr 19, 2016 6:25 pm
by bruce
Bryman wrote:Basically my computer just kept shutting off on me... sometimes it would freeze for couple seconds and then shut off and sometimes it would just turn off... One time when I was playing a game the game closed then a split second later the computer froze and then a couple seconds later the computer shut off
have you checked Microsoft's Event Viewer? When something like that happens, sometimes Windows is able to write an informative message in it's error log.
Re: Low PPD on some WUs
Posted: Tue Apr 19, 2016 7:21 pm
by foldy
If you change your RAM frequency/timings you need to check with a tool like "windows memory diagnostic" if RAM is stable.
Re: Low PPD on some WUs
Posted: Tue Apr 19, 2016 9:05 pm
by Bryman
bruce wrote:Bryman wrote:Basically my computer just kept shutting off on me... sometimes it would freeze for couple seconds and then shut off and sometimes it would just turn off... One time when I was playing a game the game closed then a split second later the computer froze and then a couple seconds later the computer shut off
have you checked Microsoft's Event Viewer? When something like that happens, sometimes Windows is able to write an informative message in it's error log.
Yes, it wasn't saying anything useful
"The operating system started at system time 2016-04-19T09:19:19.610798500Z."
"File System Filter 'FileInfo' (6.1, 2009-07-13T18:34:25.000000000Z) has successfully loaded and registered with Filter Manager."
"Crash dump initialization failed!"
"The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly."
"Intel(R) Management Engine Interface driver has started successfully."
"Broadcom NetLink (TM) Gigabit Ethernet: Driver initialized successfully."
And then 8 Kernel-Processor-Power events
"Microsoft (R) Windows (R) 6.01. 7601 Service Pack 1 Multiprocessor Free."
"The previous system shutdown at 4:18:20 AM on 4/19/2016 was unexpected."
"The system uptime is 6 seconds."
"The Event log service was started."
"One or more of the Plug and Play service's subsystems has changed state."
"PlugPlay install subsystem enabled: 'true'
PlugPlay caching subsystem enabled: 'true'"
"The Plug and Play service entered the running state."
"The Power service entered the running state."
"Broadcom NetLink (TM) Gigabit Ethernet: Network controller configured for 1Gb full-duplex link."
"File System Filter 'luafv' (6.1, 2009-07-13T18:26:13.000000000Z) has successfully loaded and registered with Filter Manager."
And about 15 Service Control Manager events
And.... pretty much what happened when the computer booted, in that order
foldy wrote:If you change your RAM frequency/timings you need to check with a tool like "windows memory diagnostic" if RAM is stable.
Yeah, I need to run memtest86 again
Re: Low PPD on some WUs
Posted: Wed Apr 20, 2016 12:41 am
by Bryman
Dang... I thought it was stable... restarted after 18 hours :/
No idea what thing I changed to not make it stable, though :/