New CPU = runs rather hot and BSODs

[Edward Yee]

Here's the motherboard, and here's the CPU (mine is revision RB-C3).

There was some movement when I installed the CPU cooler (may have inadvertently spread out the thermal paste), and soon after I installed the CPU, my computer suffered its first BSOD in a long time. My first response (oddly enough) was to inspect the thermal compound, which seemed awfully thin to me -- before. I scraped as much off as possible from the CPU surface (and the "green" circuit board -- some thermal paste seemed to have gotten onto the top side of it) and the cooler before reapplying some thermal paste I happened to have laying around at home, then put it back in. After a BIOS update (from F1 to F5, the minimum for the CPU), it correctly displayed as a 3.4 GHz quad-core... though oddly my Windows Experience Index went down. However, while using a custom power plan based off of the High Performance plan, my system BSOD'ed several times, usually seeming to be driver-related (i.e. "atikmdag.sys," most recently "tcpip.sys") or page file-related, irrespective of whether I was logged into the standard user account or an administrator account, what I was doing, or even the power plan (it BSOD'd once when I was using custom settings). As such, my system is unacceptably less stable when I would use over a hundred Internet browsing tabs on only a 2.6 GHz dual-core CPU...

Also, it seems to run awfully hot for a CPU -- in my most recent experience, when I went into BIOS during a 'cold-start' boot, my CPU temperature went from 30 C to nearly 40 C in only a few minutes, despite its fan running at 3500+ RPM, and the "system temperature" rising by similar amounts in the same time period. As I type, even with a Power Saver or Balanced power plan, according to Core Temp 0.99.8 my CPU temperature tends towards the mid-to-high-90s, and at one point in a previous use session with the High Performance power plan it measured a whopping 58 C.

I hope that I haven't doomed this CPU by either using it at first without a BIOS upgrade (I did so soon afterward) or by poor thermal paste application? I've been told to get better thermal paste, but the worst-case scenario seems to be to get a replacement through Newegg's policy. As such, I'd appreciate any recommended solutions.

[Dominus Atheos]

58c is nothing, don't worry about it. Just set the BIOS option to turn off at a certain temp. (I have mine set at 70c)

As for the BSODs, try reinstalling the old proc and see if that solves the issue. If it doesn't, try a system restore and/or clean OS reinstall. Even if it does solve the issue, I'd still recommend a clean system recovery with the quad core installed before returning it.

[Edward Yee]

I should note that the BSODs were seemingly at random, so how I'd test for it with the old processor I'm not sure, and I'd need to reapply thermal paste anyway.

I doubt that System Restore would help (it didn't seem to help me with previous issues, and Startup Repair seems useless), though I could give it a shot... and I'm not opposed to a clean OS reinstall, though that'd be tedious.

"A clean system recovery"?

For what it's worth, this claimed to find several errors, although Windows Memory Diagnostic turned up nothing; I'll be trying Memtest86 and Memtest86+ though later tonight. If I do need to replace hardware then hopefully it's only the DRAM -- cheap and obsolete.

[Zaune]

As far as the extra heat goes, it might be worth double-checking you haven't got a bundle of cables messing up the airflow.

[tezunegari]

I don't think the thermal paste is at fault - unlike older AMD CPUs the x4 965 has a heat spreader on top (a metal plate increasing contact area with the cooler).

Have you tried resetting the BIOS to factory default?
How to do so should be described in the manual.

[Napoleon the Clown]

Have you upgraded the graphics card, too? If you've got a higher-end GPU it's gonna put off a fair bit of heat. 40* isn't an ideal idle temp, but it's not an emergency, either. What's the motherboard temperature at during all this?


You may well be lacking enough airflow. If such is the case, you'll either need a better case or to install water cooling. I've got the 3 ghz version of what you have and it runs nearly 10* C cooler than you're running, and I wouldn't think 400 mhz is going to be that big an increase in heat.

[TC27]

In my experience STOP errors are usually related to an issue with your RAM - did you re-use the same RAM modules when you upgraded/

Have you fiddled with any memory voltage or ratio settings?

[Uraniun235]

What power supply do you have, and what CPU did you upgrade from?

[Dominus Atheos]

Dear god we need a tech support subforum.

I should note that the BSODs were seemingly at random, so how I'd test for it with the old processor I'm not sure, and I'd need to reapply thermal paste anyway.

How often do they happen?

I doubt that System Restore would help (it didn't seem to help me with previous issues, and Startup Repair seems useless), though I could give it a shot... and I'm not opposed to a clean OS reinstall, though that'd be tedious.

if you are getting driver related errors, doing a system restore to before they started should help.

I'd definitely recommend aclean reinstall before spending any more money.

"A clean system recovery"?

For what it's worth, this claimed to find several errors, although Windows Memory Diagnostic turned up nothing; I'll be trying Memtest86 and Memtest86+ though later tonight. If I do need to replace hardware then hopefully it's only the DRAM -- cheap and obsolete.

So what were the results?

[Edward Yee]

It seems that the desktop is finally running stably... after several changes. Physically speaking, I moved the sticks from the slots nearest to the CPU socket (#1 and #2) to the adjacent slots (#3 and #4). I also upgraded the BIOS from F5 ("stable") to F6B ("beta"), the most recent version, and finally -- after seeing that the "factory" defaults for both were actually "auto-setting" my RAM timings to 5-5-5-18 -- I manually set them to 4-4-4-18, as is specified on the sticks themselves (albeit "@ 2.2 V"). Memtest86 also turned up no errors on the RAM sticks.

The BSODs were recurring at varying frequency depending on the power plan it seemed (the higher the CPU clock multiplier, the more likely), but after the aforementioned changes I haven't had a single BSOD or computer freeze up.

I've found that on the Power Saver plan (in the power options) the CPU temperature seems to drop down to the mid-30s, particularly if the computer is idle, while on Balanced or on High Performance it tends to go to 40s or 50s depending on the load. Motherboard temperature tends to be within a few degrees of the CPU temperature. If these temperatures are actually relatively normal / "not bad" for what I've got, then I'm happy.

Dominus Atheos, we may be able to belay that tech support subforum yet.