I get random program crashes Windows 7 and I don't think it's RAM

[soccerballtux]

I think it's not the RAM because I ran 2 Extended passes of Windows Memory Diagnostic, and 1 full Memtest.

I think it's not the CPU because I can run 20 pass IntelBurnTest.

Could it be my SSD? But I don't have swap enabled, and I have 32GB of RAM, so when a program is just sitting there in the background and crashes (Eclipse)...that doesn't seem like it would be SSD related.

Could it be something in the motherboard? This motherboard+CPU just seems flakey. For example after the Asus BIOS un-throttles the CPU due to too-hotm VRMs, it consistently hard-locks.

It's like it's completely stable for a long time and then BOOM it fails.
I can't Prime95 test for very long because the VRMs get too hot

the following are memory addresses that "could not be read" according to windows--
0x260b6b84 (first time it crashed)

0x8bffffc4 (second time this happened)


---

I just realized I can try disabling C6 state on the CPU and setting to manual voltage, not offset. Trying that now...


---

Now I'm having an issue where every time I navigate to a certain folder in Windows Explorer, Windows Explorer crashes.

Are SSDs super finicky to overclocking?

 

[Cerb]

The SSD could have some bad data on it, but it reads more like your board and/or PSU aren't happy with your overclocking, and anything else is a side effect of that, most likely. Unless, it's a Crucial V4, or older Sandforce-based drive (prior to SF-2281).

Going to and from very different power states is far more, "risky," in terms of causing causing stability problems, than any sustained load.

 

[razel]

You're overclocking and you wonder why it's crashing randomly? If so, I'd throttle back the overclocks a bit. Testing individual components stability separately will not guarantee a stable system. As with any overclock, when you reach the max overclock you really don't want to keep it there if you don't need to all the time. It's a lesson I learned back in the day.

In the meantime, get back to stock and run a SFC command line to ensure your basic Windows files are up to snuff.

 

[soccerballtux]

I think chkdisk solved the Explorer crash issue.

Solid PSU with great overclocking results before (4ghz Ph2-965).
OCZ Arc 100 240GB. How are those?

I need to write a script for my IBT loading (to keep VRMs within thermal boundaries) but I do not feel like this is a usual "you're overclocking" issue as these issues were happening before I did any overclocking.

The part that makes me think it's mobo-BIOS related is at full-turbo speed 4.3ghz and 5% load just playing in Eclipse, I get crashes. In particular, I don't trust the LLC.

 

[VirtualLarry]

Solid PSU with great overclocking results before (4ghz Ph2-965).

The part that makes me think it's mobo-BIOS related is at full-turbo speed 4.3ghz and 5% load just playing in Eclipse, I get crashes. In particular, I don't trust the LLC.

You've got a Phenom II @ 4.3Ghz, and you don't think your overclocking is a problem? LOL!

They weren't meant to run faster than 4.

 

[denis280]

You've got a Phenom II @ 4.3Ghz, and you don't think your overclocking is a problem? LOL!

They weren't meant to run faster than 4.

+1 they don't like it

 

[bbhaag]

I don't think his current cpu is a 965. He meant the psu he is using oc'ed his 965 to 4.3 in the past.

Anyway, have you tried dialing back your oc and seeing if that helps?

 

[soccerballtux]

You've got a Phenom II @ 4.3Ghz, and you don't think your overclocking is a problem? LOL!

They weren't meant to run faster than 4.

+1 they don't like it

I know what you guys are talking about, but this is a FX piledriver at 4.3ghz, not a Ph2. The Ph2 was 4ghz.

 

[soccerballtux]

I don't think his current cpu is a 965. He meant the psu he is using oc'ed his 965 to 4.3 in the past.

Anyway, have you tried dialing back your oc and seeing if that helps?

yes. forum ppl are stupid

that's where I was to begin with, but considering I've never had OC'ing related FS corruptions in the past I didn't think I needed to worry about the chkdisks then.

Since I've
-chkdisk'd
-cleared my RAM drive for Chrome which locked up on the Dominos pizza tracker
-turned off C6
-switched from offset voltage to manual voltage control
-added another 0.02v in case the LLC is lagging too much
I've been fine and made it through a 3 hour Prime95 4-thread test

will post back later

in the past, dialing back the OC didn't help.

 

[VirtualLarry]

Could it be something in the motherboard? This motherboard+CPU just seems flakey. For example after the Asus BIOS un-throttles the CPU due to too-hotm VRMs, it consistently hard-locks.

It's like it's completely stable for a long time and then BOOM it fails.
I can't Prime95 test for very long because the VRMs get too hot

I know what you guys are talking about, but this is a FX piledriver at 4.3ghz, not a Ph2. The Ph2 was 4ghz.

Then your motherboard sucks. Get a new one.

 

[schmuckley]

I know what they say n all but..
Download ultradefrag n run it..see if fixes.

I'm feeling it will

Otherwise..moar cpu/nb voltage
..and nb voltage

fx doesn't like high ht speed
or high cpu/nb speed either

actually ht link speed can be 200mhz higher than k10
but cpu/nb = much lower than k10 stuff

 

[BonzaiDuck]

Set all the overclocking items back to stock.

Do yourself a "BIOS inventory" of motherboard features and determine if there's anything enabled that you don't use and don't want to use.

I know nothing of AMD processors, but the way you describe it, these crashes occur infrequently and under near-idle situations?

If you can confirm that the crashes still occur at all-stock settings, you can then give yourself the right to "diss" other posters, but you must eliminate your clock settings as a cause.

 

[razel]

Good that you went back to stock OC. chkdsk is not the same as SFC. SFC checks that your core Windows files are ok. I hope you get the point in all this. Get your hardware back to stability, then get your software OS to stability. Why continue working off software (that's probably corrupt) running on a clearly unstable system?

 

[soccerballtux]

something new I hadn't tried-- I re-flashed the BIOS update that I flashed when I got it.

Between that and increasing voltage to 1.32v I seem to be doing OK.
I'm torn as to which solved it. I would turn it off for 10 seconds, boot up, and set the voltage to 1.4v, boot into windows, and have the exact same problems.

The LLC or voltage seems more stable now. Watching the voltage in CPU-z it doesn't bounce around +/- 0.02v like it used to. I don't really want to try dropping the voltage to see if that reintroduces stuff TBH.

The warranty is 3 years so I'll keep it where it is for now until I get a better feel for its stability.

 

[soccerballtux]

Set all the overclocking items back to stock.

Do yourself a "BIOS inventory" of motherboard features and determine if there's anything enabled that you don't use and don't want to use.

I know nothing of AMD processors, but the way you describe it, these crashes occur infrequently and under near-idle situations?

If you can confirm that the crashes still occur at all-stock settings, you can then give yourself the right to "diss" other posters, but you must eliminate your clock settings as a cause.

yesterday it was occuring within 5 minutes, then within 30 seconds of bootup. That was when I called Asus and started an RMA, then thought of trying a re-flash of the BIOS as I've seen low-level flashes that look like they went OK, not be OK.

I still question if it's going to do it again as Chrome has started "Aww, snap! (Reload?)"ing again

 

[BonzaiDuck]

something new I hadn't tried-- I re-flashed the BIOS update that I flashed when I got it.

Between that and increasing voltage to 1.32v I seem to be doing OK.
I'm torn as to which solved it. I would turn it off for 10 seconds, boot up, and set the voltage to 1.4v, boot into windows, and have the exact same problems.

The LLC or voltage seems more stable now. Watching the voltage in CPU-z it doesn't bounce around +/- 0.02v like it used to. I don't really want to try dropping the voltage to see if that reintroduces stuff TBH.

The warranty is 3 years so I'll keep it where it is for now until I get a better feel for its stability.

Heh. I'll choose answering this among your posts to answer the one following. I can't say anything about your hardware, so won't make assumptions that you made any mistakes with flashing your BIOS -- or anything else for that matter.

Booting the machine, using it on and off in various ways, and letting it slip into sleep and then hibernation -- waking it up again etc. should prove over several days.

The main thing you said that got my attention: "I'm torn as to which solved it."

I know what you mean! I know what you mean!

 

[soccerballtux]

AMDs have just gotten flakier and flakier it seems. In the Athlon XP days you could still use windows you'd just get errors, and Prime95 would fail before the whole PC would lock up.

 

[BonzaiDuck]

AMDs have just gotten flakier and flakier it seems. In the Athlon XP days you could still use windows you'd just get errors, and Prime95 would fail before the whole PC would lock up.

On my Intel system, the unexpected "resets" or BSODs (both) would occur on average every 11 days to a month. The machine was running 24/7. Sleep was not enabled for reasons related to an HTPC function and my desire to leave my TV on all night. I had a set of planned upgrades for the machine, and I started making the change-outs earlier as part of the troubleshooting. It also prompted me to clean up the reds and yellows in my event logs -- for this and every machine in the house.

It seemed to go away when I disabled the onboard Asmedia USB3 and auxiliary Marvell controller, but to this day I can't be sure. I had also reconfigured my RAM through the "XMP" profiles to yield the identical manual settings. I also discovered I was using a UPS (CyberPower) that was incompatible with Active PFC PSUs. I'd seen similar symptoms in forum posts other than Anand for each and every one of these items. Excluding OC settings as a cause was the easiest thing to do initially.

"Out there" in ordinary life, citizens want to ascribe every problem to a single cause. But in the analysis of causation, there can be multiple causes acting together, hiding one another, acting independently to give the same symptom.

So like I said: "I know what you mean!" All I can say for certain: I don't have any such problem with my OC'd systems anymore -- generally -- not with ANY of my systems. Anyone can certainly imagine, though, the troublesome effort in troubleshooting a problem that occurs so infrequently as mine did.

 

[soccerballtux]

that never happened to me and I am an AMD fanboy.

They failed like clockwork outside the thermal and voltage spec (1.48v and 73.2F)