Arghhhh.......lost another -bigadv :(

amdgamer

Supreme [H]ardness
Joined
Oct 27, 2004
Messages
4,880
My folding rig was running just fine 45 minutes ago, but when I walked by the computer, it was hardfrozen again. It was stuck with the screensaver and was hardlocked. I have to be doing something wrong here, but bigadv #2 was so awesome, and now it got frozen here in the middle of bigadv #3 as you can see. This time, it was unrecoverable and had to discard the entire WU.

Is there are program out there that will allow me to access the bios settings from windows instead of rebooting the computer? I'd love to copy and paste so you guys could have a look since I must be doing something wrong. Temps were in the low 70's around the time of hte freeze.


[11:07:39] Completed 150000 out of 250000 steps (60%)
[11:41:10] Completed 152500 out of 250000 steps (61%)
[12:14:39] Completed 155000 out of 250000 steps (62%)
[12:48:07] Completed 157500 out of 250000 steps (63%)


--- Opening Log file [April 28 17:15:09 UTC]


# Linux SMP Console Edition ###################################################
###############################################################################

Folding@Home Client Version 6.29

http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: /usr/local/fah
Executable: ./fah6
Arguments: -bigadv -smp 8 -verbosity 9

[17:15:09] - Ask before connecting: No
[17:15:09] - User name: amdgamer (Team 33)
[17:15:09] - User ID: 7B378240567E464C
[17:15:09] - Machine ID: 3
[17:15:09]
[17:15:09] Loaded queue successfully.
[17:15:09]
[17:15:09] - Autosending finished units... [April 28 17:15:09 UTC]
[17:15:09] Trying to send all finished work units
[17:15:09] + No unsent completed units remaining.
[17:15:09] - Autosend completed
[17:15:09] + Processing work unit
[17:15:09] Core required: FahCore_a2.exe
[17:15:09] Core found.
[17:15:09] Working on queue slot 03 [April 28 17:15:09 UTC]
[17:15:09] + Working ...
[17:15:09] - Calling './mpiexec -np 8 -host 127.0.0.1 ./FahCore_a2.exe -dir work/ -nice 19 -suffix 03 -checkpoint 15 -verbose -lifeline 989 -version 629'

[17:15:09]
[17:15:09] *------------------------------*
[17:15:09] Folding@Home Gromacs SMP Core
[17:15:09] Version 2.10 (Sun Aug 30 03:43:28 CEST 2009)
[17:15:09]
[17:15:09] Preparing to commence simulation
[17:15:10] - Ensuring status. Please wait.
[17:15:19] - Looking at optimizations...
[17:15:20] - Working with standard loops on this execution.
[17:15:20] - Files status OK
[17:15:28] - Expanded 30326439 -> 159726549 (decompressed 101.8 percent)
[17:15:30] Called DecompressByteArray: compressed_data_size=30326439 data_size=159726549, decompressed_data_size=159726549 diff=0
[17:15:31] - Digital signature verified
[17:15:31]
[17:15:31] Project: 2681 (Run 5, Clone 19, Gen 97)
[17:15:31]
[17:15:31] Entering M.D.
[17:15:37] Using Gromacs checkpoints
[17:15:56] Resuming from checkpoint
[17:15:56] fcSaveRestoreState: I/O failed dir=0, var=0000000040D52240, varsize=54000
[17:16:04] CoreStatus = FF (255)
[17:16:04] Sending work to server
[17:16:04] Project: 2681 (Run 5, Clone 19, Gen 97)
[17:16:04] - Error: Could not get length of results file work/wuresults_03.dat
[17:16:04] - Error: Could not read unit 03 file. Removing from queue.
[17:16:04] Trying to send all finished work units
[17:16:04] + No unsent completed units remaining.
[17:16:04] - Preparing to get new work unit...
[17:16:04] Cleaning up work directory
[17:16:04] + Attempting to get work packet
[17:16:04] Passkey found
[17:16:04] - Will indicate memory of 7889 MB
[17:16:04] - Connecting to assignment server
[17:16:04] Connecting to http://assign.stanford.edu:8080/
[17:16:05] Posted data.
[17:16:05] Initial: 43AB; - Successful: assigned to (171.67.108.22).
[17:16:05] + News From Folding@Home: Welcome to Folding@Home
[17:16:05] Loaded queue successfully.
[17:16:05] Connecting to http://171.67.108.22:8080/
[17:16:42] Posted data.
[17:16:42] Initial: 0000; - Receiving payload (expected size: 30326951)
[17:17:33] - Downloaded at ~580 kB/s
[17:17:33] - Averaged speed for that direction ~538 kB/s
[17:17:33] + Received work.
[17:17:33] Trying to send all finished work units
[17:17:33] + No unsent completed units remaining.
[17:17:33] + Closed connections
[17:17:38]
[17:17:38] + Processing work unit
[17:17:38] Core required: FahCore_a2.exe
[17:17:38] Core found.
[17:17:38] Working on queue slot 04 [April 28 17:17:38 UTC]
[17:17:38] + Working ...
[17:17:38] - Calling './mpiexec -np 8 -host 127.0.0.1 ./FahCore_a2.exe -dir work/ -nice 19 -suffix 04 -checkpoint 15 -verbose -lifeline 989 -version 629'

[17:17:39]
[17:17:39] *------------------------------*
[17:17:39] Folding@Home Gromacs SMP Core
[17:17:39] Version 2.10 (Sun Aug 30 03:43:28 CEST 2009)
[17:17:39]
[17:17:39] Preparing to commence simulation
[17:17:39] - Ensuring status. Please wait.
[17:17:48] - Looking at optimizations...
[17:17:48] - Working with standard loops on this execution.
[17:17:48] - Files status OK
[17:17:53] - Expanded 30326439 -> 159726549 (decompressed 101.8 percent)
[17:17:53] Called DecompressByteArray: compressed_data_size=30326439 data_size=159726549, decompressed_data_size=159726549 diff=0
[17:17:54] - Digital signature verified
[17:17:54]
[17:17:54] Project: 2681 (Run 5, Clone 19, Gen 97)
[17:17:54]
[17:17:55] Entering M.D.
[17:18:18] Completed 0 out of 250000 steps (0%)
 
Sorry to see this, amd. There was a utility almost a decade ago that was able to access a board's BIOS from Windows, but it wasn't always stable and didn't support every board. I don't recall the name now or know what happened to it.

What do you think may be the problem? Are you running any other apps on this system besides folding? All my -bigadv systems are dedicated and stripped of every unneeded process including Windows' stuff.
 
Sorry to see this, amd. There was a utility almost a decade ago that was able to access a board's BIOS from Windows, but it wasn't always stable and didn't support every board. I don't recall the name now or know what happened to it.

What do you think may be the problem? Are you running any other apps on this system besides folding? All my -bigadv systems are dedicated and stripped of every unneeded process including Windows' stuff.

This is a dedicated folding rig, so it is not running anything other than the screensaver, and Catalyst 10.2 drivers for the cheapie video card. Is there a guide out there that you would recommend? I have never thought about disabling other stuff that Windows 7 automatically loads up that are not even needed for folding.

I've gone ahead and downclocked to 3.50ghz from 3.57 now and have kept voltages safe. It looks like it started up another -bigadv, but I am going to quit this one and just start doing SMP A3's for now just to be safe. I managed to do quite a few A3's in windows just fine before I started doing the -bigadv stuff.
 
What's the temps for the chipset?

Most of the time when it's an oc cpu/ram problem you'll get a BSOD...

Temps for the cpu were in the low 70's. Normally they are in the upper 70's but I left the windows open last night.

I don't know about the chipset, but let me check as I think speedfan allows you to read chipset temps as well.
 
I gave up on the bigadv myself. i can run smp all day long @4.0 but have to go to 3.6 for bigadv myself. the point difference between the two at those speeds insnt enough to make it worth while.
 
why do you have a screen saver running if it's dedicated?? Also bump the overclock down some, maybe 100 mhz you've almost hit your sweet spot if it's freezing up at 60%. Did you try using a backup script?

I got two bigadv boxens running at 3.6ghz and 3.8ghz without a hiccup
 
I doubt it's a vcore problem, you're already giving it a lot.

My guess would be the chipset?
Else you could try lowering the speed your ram and loose the timings.
 
if your O/C was rock solid for A3, consider running bigadv smp7. mine ran fine at that setting for over a week with no problems.

i'm in the process of trying to "stabilize" my machine for bigadv smp8 myself. TBH i'm skeptical i'll find it (@ 3.8ghz).

APOLLO gave me some good advice in another thread. occasionally suspend your WM and backup the folder. that way if your system locks up and you lose the WU, you can just restore the backup and resume from there.
 
why do you have a screen saver running if it's dedicated?? Also bump the overclock down some, maybe 100 mhz you've almost hit your sweet spot if it's freezing up at 60%. Did you try using a backup script?

I got two bigadv boxens running at 3.6ghz and 3.8ghz without a hiccup

I run the screensaver mainly because I leave the CRT monitor running. That old CRT monitor is nearly 2 decades old and it doesn't seem to like being turned on and off constantly as I walk by to check progress. My first hardlock was on my very first bigadv and happened at 83%, and I wasn't running a screensaver at that time. Bigadv #2 went the full way without a single problem, and now -bigadv #3 hardlocks at 63%. This time, the WU had to be discarded as the client couldn't recover it.

What is a backup script? I honestly am not familiar at all with now Linux or VM's work as I just point and click. Once Linux loads up in the VM, I just type ./fah6 since I had already set it up manually at the console. I am using the eVGA image file that I got from the eVGA forums.


I doubt it's a vcore problem, you're already giving it a lot.

My guess would be the chipset?
Else you could try lowering the speed your ram and loose the timings.

Yeah, i've been comparing notes based on 930 overclocks people have here and at other forums, and I am definately feeding the cpu a lot of juice based on the overclock I am running on it. I will try to load speed fan on here and monitor chipset temps, although i'm only running a bus speed of 167 now to achieve 3.5ghz. Because of my recent downclocks, my ram is already running at a low speed and now is at around 1330 or something like that with CL8 timings. Max speed for the ram is 1600mhz @ CL8.
 
This is a dedicated folding rig, so it is not running anything other than the screensaver, and Catalyst 10.2 drivers for the cheapie video card. Is there a guide out there that you would recommend? I have never thought about disabling other stuff that Windows 7 automatically loads up that are not even needed for folding.
I don't have a guide that I can recommend off the top of my head, unfortunately, but there are many sites where you can find info on stopping unneeded services. There's a lot of services in Win 7 and Vista that are not needed to have running. Disable anything that isn't necessary. That's what I do and it usually takes some time with Win 7 to fully optimize it.

I've gone ahead and downclocked to 3.50ghz from 3.57 now and have kept voltages safe. It looks like it started up another -bigadv, but I am going to quit this one and just start doing SMP A3's for now just to be safe. I managed to do quite a few A3's in windows just fine before I started doing the -bigadv stuff.
OK, try running the system with SMP2 for a few days. Maybe by then we will get to the bottom of the problem.
 
As an aside, how are RAM module temperatures generally in i7 systems? Do they run hot or are they more moderate? I have no experience with i7 and I'm used to scorching server memory that requires active cooling.
 
I don't have a guide that I can recommend off the top of my head, unfortunately, but there are many sites where you can find info on stopping unneeded services. There's a lot of services in Win 7 and Vista that are not needed to have running. Disable anything that isn't necessary. That's what I do and it usually takes some time with Win 7 to fully optimize it.

OK, try running the system with SMP2 for a few days. Maybe by then we will get to the bottom of the problem.

I sometimes think that I should have just gone ahead and broken the piggy bank open and gotten the Gulftown instead of the 930. At least that makes life easier considering that you only need to fuss with the cpu multiplier on the extreme. :)

However for the cost of that, I could have spent a little more and gone with a Magny Cours.
 
Maybe it's just me, but I don't remember 1.26 being A LOT of vcore. I'd easily bump that to 1.3 especially with your temps in the low 70s. i7's go up to 1.375, so 1.3 shouldn't be a big deal at all.
 
All of mine hover around the 1.3V Vcore mark, and pushing 1.4V Vtt.
 
Is there are program out there that will allow me to access the bios settings from windows instead of rebooting the computer? I'd love to copy and paste so you guys could have a look since I must be doing something wrong. Temps were in the low 70's around the time of hte freeze.

Gigabyte has software that should let you change some settings, I don't recall how much though, since I didn't use it.

It's called Easytune. I'm not sure how good it is either, but its worth a try.
 
Maybe it's just me, but I don't remember 1.26 being A LOT of vcore. I'd easily bump that to 1.3 especially with your temps in the low 70s. i7's go up to 1.375, so 1.3 shouldn't be a big deal at all.

he probably meant 1.2685 vcore is plenty for only 3.57ghz. you only needed to bump the vcore way up when you were approaching 4+ ghz

anyone else think 1.56 vdimm is too low for 12GB 8C Corsair RAM?
what vdimm are you guys all running?
 
he probably meant 1.2685 vcore is plenty for only 3.57ghz. you only needed to bump the vcore way up when you were approaching 4+ ghz

anyone else think 1.56 vdimm is too low for 12GB 8C Corsair RAM?
what vdimm are you guys all running?

My memory is only running at around 1330mhz right now, which is why I took the Vdimm down. It is rated for 1.45 volts at 1333mhz, and 1.65 volts at the max speed of 1600mhz. Right now, I have it now running at 3.50 ghz(167x21) with the same vcore of 1.2685 which should be way more than enough for that speed. I remember that when I tried to take my cpu past 3.67ghz, I needed gobs of voltage.
 
Personally I don't believe there is any harm in running your vDimm at 1.65 even if you are not at max speed. Try bumping to 1.65 and seeing if that improves stability.

With i7, there are a couple of other voltages that you can tweak to increase stability. I found a really good one on the eVGA website that explains all of them in great detail (look in the MoBo subforum for just about any OC'ing guide). I'd look for you, but am supposed to be working :D

Also, if this is a dedicated machine, try running with just 6GB of RAM. You stress systems more running with all banks filled and you really only need 6GB for Big Adv.
 
he probably meant 1.2685 vcore is plenty for only 3.57ghz. you only needed to bump the vcore way up when you were approaching 4+ ghz

anyone else think 1.56 vdimm is too low for 12GB 8C Corsair RAM?
what vdimm are you guys all running?

My son's runs at 1.65. It was one of the things I just setup from the beginning and left alone...even though the memory is running at less than what it's rated for.
 
GB bios volts for vdimm goes in increments of .02v. Vdimm on both of mine is 1.66 and it's completely fine.

OP did you stress test and what did you use?
I suggest using LinX or IBT and testing first your fsb (qpi/vtt), once that's stable for an hour (lower your multi to 12 to eliminate the cpu from the equation), test your ram for at least 1 pass in memtest. Once that passes raise your multi back to 20 or 21 (if you have turbo mode enabled) and use LinX or IBT again for 20 passes using all the memory. (should take about 3-4 hours) If it completes 20 passes, you are 99% stable. I've still had a bigadv crash after a 20 pass LinX run, 1 bump in voltage after that and it is bigadv stable for me.

Your oc is not stable if it's crashing like that. Could be the vtt needs more juice, could be the ram either speed, vdimm or just bad ram...or it could be your vcore. I run 1.295 vtt through both of mine running a 180bclk, vcore is different on each though. My primary (FT02) needs more vcore than my dedicated boxen does...I want to say my FT02 is running around 1.31something and my boxen is running 1.29something.
 
Last edited:
I originally ran it several hours at 3.67ghz using Prime95 blend and it had no problems. It actually ran quite a few windows SMP A3 WU's at that speed without a single problem. However on the very first bigadv I tried, it hardlocked at 83%. I kept the voltages the same and downclocked to 3.57ghz at the time, and lowered Vdimm to 1.56. It recovered bigadv #1 and finished it just fine, and then finished bigadv #2 just fine until is just now hardlocked on me at 63% through bigadv #3. This time it was unrecoverable and the WU was lost. I have kept the voltages the same, and have taken the bus speed down to 167 for an overclock of 3.50ghz now.

When trying to find an overclock, 3.67ghz was initially the wall I found as anything beyond that required a massive increase in voltage to POST, which immediately shot my temps up to danger territory. Now I am finding that 3.57ghz wasn't even stable.

I also cannot discount my PSU possibly being the culprit because this unit is nearly 6 years old, and is an old 550watt Antec Blue that is on its 3rd system build. However since this is running just a cpu, I doubt that it is pushing the PSU very hard though.

GB bios volts for vdimm goes in increments of .02v. Vdimm on both of mine is 1.66 and it's completely fine.

OP did you stress test and what did you use?
I suggest using LinX or IBT and testing first your fsb (qpi/vtt), once that's stable for an hour (lower your multi to 12 to eliminate the cpu from the equation), test your ram for at least 1 pass in memtest. Once that passes raise your multi back to 20 or 21 (if you have turbo mode enabled) and use LinX or IBT again for 20 passes using all the memory. (should take about 3-4 hours) If it completes 20 passes, you are 99% stable. I've still had a bigadv crash after a 20 pass LinX run, 1 bump in voltage after that and it is bigadv stable for me.

Your oc is not stable if it's crashing like that. Could be the vtt needs more juice, could be the ram either speed, vdimm or just bad ram...or it could be your vcore. I run 1.295 vtt through both of mine running a 180bclk, vcore is different on each though. My primary (FT02) needs more vcore than my dedicated boxen does...I want to say my FT02 is running around 1.31something and my boxen is running 1.29something.
 
Prime 95 does not even come close to stressing the system like bigadv does.
 
When trying to find an overclock, 3.67ghz was initially the wall I found as anything beyond that required a massive increase in voltage to POST, which immediately shot my temps up to danger territory. Now I am finding that 3.57ghz wasn't even stable.
Something is definitely not right. I seriously doubt a 930 cannot manage a stable 3.5GHz if everything was installed correctly and the components are good..

I also cannot discount my PSU possibly being the culprit because this unit is nearly 6 years old, and is an old 550watt Antec Blue that is on its 3rd system build. However since this is running just a cpu, I doubt that it is pushing the PSU very hard though.
Do you have another PSU you can use to try? Even though the system is not drawing too much power, if the PSU is on its last legs it won't matter. It could reduce stability and cause issues.
 
Something is definitely not right. I seriously doubt a 930 cannot manage a stable 3.5GHz if everything was installed correctly and the components are good..

Do you have another PSU you can use to try? Even though the system is not drawing too much power, if the PSU is on its last legs it won't matter. It could reduce stability and cause issues.


the 930's are hit and miss compared to the 920's.. some hit 4ghz.. some wont even break 3.5 ghz.. some hit walls at 3.8ghz.. and some can even do 4.4ghz.. they really arent consistent chips thats for sure..

as far as the psu goes i doubt its really the problem..

amdgamer when you hit 3.8ghz what voltage were you using? since everyones definition of gobs of voltage is different.. for me its anything over +.20 and for some its .05v..

also on a second note.. have you tried raising your NB voltage? even though you are running it at below rated speeds.. at 12GB of ram your still going to put extra load on the NB.. i would try that especially since the bigadv WU's use a ton of ram.. also i recommend using intel burn test over prime95.. far better stress tester and will put much higher loads on the cpu then even the bigadv WU's would..
 
Something is definitely not right. I seriously doubt a 930 cannot manage a stable 3.5GHz if everything was installed correctly and the components are good..

Do you have another PSU you can use to try? Even though the system is not drawing too much power, if the PSU is on its last legs it won't matter. It could reduce stability and cause issues.

I don't have another PSU to try right now, but was planning on picking one up with a better case in roughly 4-6 weeks time when my entertainment budget has been recharged. I was looking at possibly a Corsair 550 watt system as they seem to have the best reputation/$$ ratio.

the 930's are hit and miss compared to the 920's.. some hit 4ghz.. some wont even break 3.5 ghz.. some hit walls at 3.8ghz.. and some can even do 4.4ghz.. they really arent consistent chips thats for sure..

as far as the psu goes i doubt its really the problem..

amdgamer when you hit 3.8ghz what voltage were you using? since everyones definition of gobs of voltage is different.. for me its anything over +.20 and for some its .05v..

also on a second note.. have you tried raising your NB voltage? even though you are running it at below rated speeds.. at 12GB of ram your still going to put extra load on the NB.. i would try that especially since the bigadv WU's use a ton of ram.. also i recommend using intel burn test over prime95.. far better stress tester and will put much higher loads on the cpu then even the bigadv WU's would..

When I tried to get my 930 to POST at 3.8ghz, it required a vcore of 1.35 volts which shot my temperatures up way too high. By contrast, it ran Prime95 stably at 3.67ghz at 1.2685 volts. Of course I found while it ran A3 WU's just fine, it couldn't do a bigadv. I believe I have my NB voltage set on auto, as I thought that I would only need to mess with the Vdimm for the memory. The only voltages I messed with were vcore, vdimm, and qti/vtt volt. I left IOH volt on auto as well since it isn't like I am running any video.
 
1.35v sounds about right for 3.8ghz.. though id try raising the NB voltage just to see if it improves anything.. you might be able to drop the cpu voltage a little by doing that.. but at the same time it would add heat by raising the NB.. what kind of temps were you seeing with the voltage at 1.35v?

and yes id recommend the corsair 550VX.. awesome PSU though a tad loud the fan spins fast enough that it acts as an extra 120mm fan which is what i liked about it versus my thermaltake TP 750w which is quite as heck but the fan barely moves any air..
 
1.35v sounds about right for 3.8ghz.. though id try raising the NB voltage just to see if it improves anything.. you might be able to drop the cpu voltage a little by doing that.. but at the same time it would add heat by raising the NB.. what kind of temps were you seeing with the voltage at 1.35v?

and yes id recommend the corsair 550VX.. awesome PSU though a tad loud the fan spins fast enough that it acts as an extra 120mm fan which is what i liked about it versus my thermaltake TP 750w which is quite as heck but the fan barely moves any air..

I was seeing upper 80's in temps at 3.8ghz. It was actually Prim95 stable although I shut it down after 45 minutes because I got scared. Keep in mind i'm used to the temps you see on the Phenoms :)

Is the QTI/Vtt volt control for the north bridge? I think I have that set at 1.275 as I was always told to make sure it is equal to vcore or slightly higher than vcore. Vcore is at 1.2685 right now.

I have further downclocked my system to 3.46ghz with 165x21. I won't have time to fiddle with my folding rig as i'm going to be hardcore busy till Friday. I just want to make sure the system is stable so that I don't have to worry about it, even if I am running too much voltage for this speed. I am trying another -bigadv at these new settings. I am running the same voltages as I was running at 3.50ghz as I was running at 3.57ghz as I was running at 3.67ghz. In other words, we'll find out whether my 930 is a dud, or whether there is another problem at work here. If I walk by my computer and find it crashed again, i'll just go straight to A3 WU's. I'm mostly curious to see how well it will finish a bigadv at 3.46ghz since I have been curious to see what the minimum speed one could run and still make the preferred max-bonus deadline.

As for PSU's, noise won't bother me as this computer is in a room that is used for storage in the house. I had to clear a path to the desk so I could access my folding rig. I've got books, papers, newspapers, and stuff piled everywhere. What doesn't fit here is in storage.
 
Prime95 is not a good stresser (anymore) it seems.
I've had Prime95 pass an oc to 12hrs before quitting it, then bigadv crashes. I tried testing that oc with LinX, it failed.

again try LinX or IBT and test your oc with one of those. You will definitely find out if it's stable or not, and it takes a lot less time too. Again - 20 "all" memory passes in LinX, your oc is 99.9% stable. Bigadv may crash it still, but that's it. Bump the voltage 1 notch from there and it's almost a guaranteed oc.

and yes id recommend the corsair 550VX.. awesome PSU though a tad loud the fan spins fast enough that it acts as an extra 120mm fan

Maybe your system is loading the psu a lot more than mine, but I have 3 Corsair psus, had a 4th - never heard the fan spin up on any of them. I know the initial batches of TX650 and TX750 had fan issues, but I've never had those issues myself. HX850, TX650, VX550 now...used to have an HX520 as well. Never heard the fan on any of them. :)

edit Ok I'm going to say "I can relate" now, fkng my main now is just getting real b!tchy about being bigadv stable. [aargh] Luckily it hasn't flushed a WU yet but I'm sure I'll drop some points off having to stress test the next 6 hours. [sigh] :mad:
 
Last edited:
Thanks as i'll definately use LinX when I get back to fiddling with this processor this weekend. Until this forum, I never knew that existed and thought that Prime95 was more or less the only thing to use to stability test an overclock. I also never knew that Intel had a burn test. So far, it is running at the current downclock of 3.46ghz, but my TPF's are suffering as i'm now upto 36 minute TPF's on a 2681 bigadv WU versus the 34 minute TPF's I was at. If it crashes again, then I know I have problems elsewhere. Messing with it will have to wait till this weekend i'm afraid. What wierd hobbies I get myself into :)

Prime95 is not a good stresser (anymore) it seems.
I've had Prime95 pass an oc to 12hrs before quitting it, then bigadv crashes. I tried testing that oc with LinX, it failed.

again try LinX or IBT and test your oc with one of those. You will definitely find out if it's stable or not, and it takes a lot less time too. Again - 20 "all" memory passes in LinX, your oc is 99.9% stable. Bigadv may crash it still, but that's it. Bump the voltage 1 notch from there and it's almost a guaranteed oc.



Maybe your system is loading the psu a lot more than mine, but I have 3 Corsair psus, had a 4th - never heard the fan spin up on any of them. I know the initial batches of TX650 and TX750 had fan issues, but I've never had those issues myself. HX850, TX650, VX550 now...used to have an HX520 as well. Never heard the fan on any of them. :)
 
Seems a bit low to me, but also dependent on the memory voltage spec.
I usually run 1.65v in bios, my memory is spec'ed at 1.65v.
 
LOL, I just had another crazy scare.I thought my computer was hardlocked again as I was passing by and my monitor was black. It turns out that this nearly 2 decade old monitor decided to kick the bucket and burnt out.

Good thing I had another monitor laying around :)
 
If you keep lowering the bus and it's still doing that, it's definiatly a bad OC, and the temps, all I can think about with regards to that is that your heatsink wasn't mounted well and airflow

I'm mobile so maybe this was covered and I'm not seeing it, but can you take some pics and show us what the area around te CPU is like?
Posted via [H] Mobile Device
 
If you keep lowering the bus and it's still doing that, it's definiatly a bad OC, and the temps, all I can think about with regards to that is that your heatsink wasn't mounted well and airflow

I'm mobile so maybe this was covered and I'm not seeing it, but can you take some pics and show us what the area around te CPU is like?
Posted via [H] Mobile Device

I think that it is the case that is the problem. If I take the side panel off, temps go down by about 6 degrees C. Once I find some time, i'm going to try applying a better thermal compound and reseating since it doesn't hurt to use some better stuff than what came included in the box. I'm planning on switching to a Corsair 550 watt PSU plus a Coolermaster HAF932 case in about 6 weeks time after I recharge my entertainment budget. This case just can't handle an overclocked i7, as it bakes in there. I can try to take some more pics if that will help, but there is so little room to move around in the room and I need to make some more room.

Use this guide when you tinker this weekend.
Best i7 oc guide I've found by far. :)

Thanks. I wish things were as easy as overclocking my Phenom black edition. I am only holding my clocks back in my gaming rig since i'm still using the stock HSF.
 
I think that it is the case that is the problem. If I take the side panel off, temps go down by about 6 degrees C. Once I find some time, i'm going to try applying a better thermal compound and reseating since it doesn't hurt to use some better stuff than what came included in the box. I'm planning on switching to a Corsair 550 watt PSU plus a Coolermaster HAF932 case in about 6 weeks time after I recharge my entertainment budget. This case just can't handle an overclocked i7, as it bakes in there. I can try to take some more pics if that will help, but there is so little room to move around in the room and I need to make some more room.
What you can do in the interim is install an extra fan on the other side of the HS assuming there is space and you have a spare fan. It can make a big difference to stability. I have done this on every system with an elevated tower-type HSF. Even though I run open cases, adding a second fan still helps, especially in the summer.
 
Once I find some time, i'm going to try applying a better thermal compound and reseating since it doesn't hurt to use some better stuff than what came included in the box.

I have the exact same setup for Mobo and processor, and I believe RAM as well (CM3X2G1600C9 according to CPU-Z), though I have an H50 cooling the processor and have no idea what you have. Regardless, your temps are WAY too high if you have a decent cooler. I would suggest reseating with a good compound as well, definitely the right next step. Beyond that, below are the settings I am using for my PC... I just started -bigadv today, but since I have achieved much higher stable OC's on my system than you have this should be helpful to you.

I have put stuff that you might have missed and that really needs to be changed in Bold and Red. If you have the same RAM as me I would recommend setting EVERYTHING except the BLCK to what I have here. If you have different RAM then use whatever the XMP-1600 setting in CPU-Z shows for the corresponding settings.

Anyway, once you do that turn your OC back up to at least 3.67 and check for stability. If you are good then run a few -bigadv. If you can do that for a week or so and all is good then either turn your OC up a little or drop your Vcore, QPI, and RAM by one step each and try again for another week. Keep doing so until you fail, then turn them back up to the last successful settings. Then turn Loadline Calibration down to level 1, test again. If that works, turn it down to normal and test again. If at any time this fails turn Vcore up a notch and try again, keep this up until stable. You are better off with Loadline normal for long-term PC use, but for initial testing of everything else it is very nice to have.

GIGABYTE GA-X58A-UD3
MB Intelligent Tweaker(M.I.T.)
Advanced CPU Features {Press Enter]
CPU Clock Ratio............................: 21
CPU Frequency .....(Ghz)...................:3.9
Intel(R) Turbo Boost Tech..................:[Disabled]
CPU Cores Enabled..........................:[All]
CPU Multi Threading........................:[Enabled]
CPU Enhanced Halt (C1E)....................:[Disabled]
C3/C6/C7 State Support.....................:[Disabled]
CPU Thermal Monitor........................:[Enabled]
CPU EIST Function..........................:[Enabled]
Bi-Directional PROCHOT.....................:[Enabled]
Virtualization Technology..................:[Enabled]

Uncore & QPI Features [Press Enter]
QPI Clock Ratio*...........................:36
QPI Link Speed* ......................(Ghz):6.696
Uncore Clock Ratio.........................:x17
Uncore Frequency......................(Mhz):3,162
Isochronous Support........................:[Enabled]
Advanced Clock Control [Press Enter]
>>>>>Standard Clock Control
Base Clock (BCLK) Control*.................:[Enabled]
BCLK Frequency (Mhz)*......................:186
PCI Exxpress Frequency (Mhz)...............:103
>>>>>Standard Clock Control
CPU Clock Drive (mV).......................:700
PCI Express Clock Drive (mV)..............:700
CPU Clock Skew (ps)........................:0
IOH Clock Skew (ps)........................:0
Advanced Dram Features [Press Enter}
Performance Enhance........................:[Standard]
Extreme Memory Profile (X.M.P.)............:[Disabled]
System Memory Multiplier (SPD).............:x8
Memory Frequency (Mhz).....................:1488
DRAM Timing Selectable (sPD) ..............:[Quick]
Channel A Timing Settings Sub Menu [Press Enter]
>>>>>Channel A Standard Timing Control
Cas Latency Time.............................:9
tRCD.........................................:9
tRP..........................................:9
tRAS.........................................:24
>>>>>Channel A Advanced Timing Control
tRC..........................................:41
tRRD ........................................:[Auto]
tWTR ........................................:[Auto]
tWR..........................................: [Auto]
tWTP.........................................:[Auto]
tWL..........................................:[Auto]
tRFC.........................................:[Auto]
tRTP.........................................:[Auto]
tFAW.........................................: [Auto]
Command Rate (CMD) .....................:2
>>>>>Channel A Misc Timing Control
B2B CAS Delay................................:[Auto]
Round Trip Latency...........................:[Auto]
Channel A Turnaround Setting [Press Enter]
>>>>> Channel A Reads Followed By Reads
Different DIMMs..............................: [Auto]
Different Ranks..............................: [Auto]
On The Same Rank.............................: [Auto]
>>>>> Channel A Writes Followed By Writes
Different DIMMs..............................: [Auto]
Different Ranks..............................: [Auto]
On The Same Rank.............................:[Auto]

Advanced Voltage Control [Press Enter]
>>>>CPU Voltages
LoadLine Calibration.........................:[Level 2]
CPU Vcore....................................:1.3125v
QPI/VTT Voltage..............................:1.335v
CPU PLL......................................: 1.80v
>>>>>MCH/ICH Voltrage
PCIE.........................................: 1.50v
QPI PLL......................................: 1.10v
IOH Core.....................................:1.10v
ICH I/O......................................: 1.50v
ICH Core.....................................:1.10v
>>>>>Dram
DRAM Voltage.................................:]1.64v
DRAM Termination.............................: [Auto]
Ch-A Data Vref...............................:[Auto]
Ch-B Data Vref...............................:[Auto]
Ch-C Data VRef...............................:[Auto]
Ch-A Address Vref............................:[Auto]
Ch-B Address Vref............................:[Auto]
Ch-C Address VRef............................: [Auto]
 
Last edited:
Back
Top