My Q6600 and P35-DS3L blue screen when OCed to 2835Mhz

[neosapien]

I've had my B3 revision 1.325V VID Core 2 Quad 6600 since early December, and since then it's been running stably at stock speeds, 266x9 (2400Mhz), with a 1.10625V vcore, and the FSB:memory ratio was 4:5, so the memory was running at 667mhz.

Today, I get the idea to OC it to 3.0Ghz (333x9) and figure that 333x9 at 1.325V should be easily attainable. So, I set the PCI-E bus to 100mhz in the BIOS and set the FSB:memory ratio to 1:1, boot into vista, use Gigabyte's Easytune software to reset the vcore from 1.10625V to 1.325V (I was afraid of doing it in the BIOS, as I didn't want to do something wrong and have to reset the CMOS using the jumper), and start increasing the FSB in increments of 4mhz and then running prime95 in small FFT mode for 5 minutes.

When I set it to 315mhz FSB and run prime95, I get a blue screen. Strange. So, I figured that the FSB and the MCH might need a bit more voltage (though it probably shouldn't - this motherboard is supposed to be able to run a 1333mhz FSB right out of the box, and 315 is only 1260mhz). So, I set the MCH and FSB overvoltage in the BIOS to +0.1V, reboot and reset to 315x9 and 1.325V in Easytune, then start up prime95 again. After about 1 minute of prime95, I get a fatal error in the third core, and a few seconds later, get another blue screen.

So, I suppose it's not the FSB, since I upped the voltage and still got the same result. I don't think it's the memory, since I ran this memory for 4 months at 667mhz without any problems, surely 630mhz is a walk in the park for it. And surely it can't be the vcore. 1.325V for a B3 Q6600 ought to be able to do 3.00Ghz, much less 2835Mhz. In Graysky's how to OC a quad thread, he says his B3 does 3.00Ghz with 1.2625V.

 

[BlueAcolyte]

1) Don't OC using Easytune, god... They have the BIOS for a reason.
2) The B3 quads are pretty hot running, how about that heat?
3) Maybe it's just your chip? Just keep tweaking away.

 

[neosapien]

1) I'm new to overclocking, and am slightly apprehensive about being unable to POST after attempting to overclock, and then having to unplug my power supply, discharge the capacitors, and then play with the jumpers to reset the CMOS to the factory defaults. That's why I used Easytune, so if I crash, then when the system resets it will be back to good old stock speeds.
2) 2.40Ghz 1.10625v vcore: 37c idle, 48c load. 2.835Ghz 1.325v vcore: ~40c idle, 59c load.
3) You mean tweak the vcore by going above 1.325v?

 

[core2slow]

Originally posted by: neosapien
1) I'm new to overclocking, and am slightly apprehensive about being unable to POST after attempting to overclock, and then having to unplug my power supply, discharge the capacitors, and then play with the jumpers to reset the CMOS to the factory defaults. That's why I used Easytune, so if I crash, then when the system resets it will be back to good old stock speeds.
2) 2.40Ghz 1.10625v vcore: 37c idle, 48c load. 2.835Ghz 1.325v vcore: ~40c idle, 59c load.
3) You mean tweak the vcore by going above 1.325v?

the DS3L will automatically goes back to default setting if it's unable to POST. You dont have to mess around with jumpers or anything.

 

[krylon]

Sounds like you should just sell your hot poor overclocking chip to me. For cheap.

 

[gingerstewart55]

Originally posted by: krylon
Sounds like you should just sell your hot poor overclocking chip to me. For cheap.

LOL!


Yes, I think he meant try raising your vcore a bit more....not unusual to have to give your cpu a bit more juice to get higher speeds out of it.

I believe the B3 stepping has a vcore range up to 1.5V, if I remember correctly.

I got my G0 Q6600 up to 3.8GHz at 1.45V.....but I'm running on water and my idle temps are low to mid 30's and full load temps never broach mid-50's. I've played with it at 1.4V at 3.8GHz, but prefer a little more to make sure under stress it doesn't have problems. I'm probably overvolting a little more than necessary, but I haven't had one crash even after hours of gaming, benching, anything.

Don't be afraid to adjust in the BIOS.....it'll work better that way than depending upon that software stuff.

 

[Markfw]

what you want on a B3 is a MINIMUM of 1.35 UNDER LOAD. You may need to set the bios as high as 1.47 to do that, but try in increments. Start with 1.4 in bios (vdroop), +.2 on memory and GMH and make sure the memory divider is at 2.0 Try first time for 3200, or 355 x 9. At those settings, its a good place to start. You might need a little more vcore, but it should at least post.

 

[neosapien]

Do I need to install Gigabyte's Xpress Recovery in order for my BIOS to reset to normal if I do something that makes it unable to POST? Or is that feature there whether Xpress Recovery is installed or not?

 

[VirtualLarry]

Originally posted by: neosapien
Do I need to install Gigabyte's Xpress Recovery in order for my BIOS to reset to normal if I do something that makes it unable to POST? Or is that feature there whether Xpress Recovery is installed or not?

You would be wise to NEVER install that Xpress Recovery stuff. At least from the horror stories that I've read in the past.

I think what he is talking about is that if the CMOS settings for freq/voltage are wrong, then it won't POST, it will auto-reset to default and then POST. Sometimes, at least.

 

[Markfw]

Originally posted by: VirtualLarry

Originally posted by: neosapien
Do I need to install Gigabyte's Xpress Recovery in order for my BIOS to reset to normal if I do something that makes it unable to POST? Or is that feature there whether Xpress Recovery is installed or not?

You would be wise to NEVER install that Xpress Recovery stuff. At least from the horror stories that I've read in the past.

I think what he is talking about is that if the CMOS settings for freq/voltage are wrong, then it won't POST, it will auto-reset to default and then POST. Sometimes, at least.

It does it for me all the time !

 

[neosapien]

I am attempting to establish that it is the vcore which is causing me to crash at 2835Mhz 9x315, and not the memory or the FSB. So, I went into the BIOS and set the CPU to 1.325v and 315x6 (1890Mhz) and am running Prime95. If I can get the FSB to 333 and the memory to 666 while the CPU multiplier is at 6x, then I'll know beyond a shadow of a doubt that to get past 2835Mhz I have to raise the vcore.

@Markfw900 & gingerstewart55
I don't want to OC to 3.20Ghz or 3.80Ghz, I want to get 3.0Ghz (9x333). I just want a mild overclock.

I must have a terribler chip than I first suspected if 1.325V can only get me to 2835Mhz.

Update:
10 minutes stable in Prime95 at 1.325V - 315x6.
Update 2:
333x6 was stable in Prime95. It appears that the inability to do 2.83Ghz must certainly now be a vcore issue, and not a FSB or memory issue as I first suspected. I updated the BIOS from F6 to F7, which is the latest non-beta BIOS, and then I set 315x9 1.325V in the BIOS and booted, and am in Vista now with those settings running Prime95, and it's been stable now for 11 minutes, which is a big improvement from the near-instant BSODs I was getting when setting it to 315x9 1.325V in Easytune. Also the CPU heats up to 60C in Prime95. Youch. Also, my USB MS Wireless Comfort Keyboard 4000, which operated flawlessly at 2.40Ghz, appears to operate intermittently at 2.83Ghz, missing some keypresses so I have to repeat them.

 

[neosapien]

Wow, you guys were right about Gigabyte Easytune sucking. When overclocking from within the BIOS, I was able to do 333x9 at 1.325V, which was stable in prime95 for an hour before I started minimizing the vcore. At 1.275V, I got a blue screen almost immediately after starting prime95 small FFTs. At 1.28125V, blue screen after 15 minutes in prime95 small FFT. Now I'm testing 1.2875V in Prime95 small FFT, and so far (1h17m), no prime95 error or blue screen.

 

[neosapien]

1.2875V resulted in a blue screen after 2 hours and 28 minutes (strange a blue screen and not a prime95 rounding error, i wonder what that means), so now I'm testing 1.29375V in prime95 small FFT.

 

[Idontcare]

Originally posted by: neosapien
1.2875V resulted in a blue screen after 2 hours and 28 minutes (strange a blue screen and not a prime95 rounding error, i wonder what that means), so now I'm testing 1.29375V in prime95 small FFT.

Bluescreens are because Windows has an issue, not because Prime95 had an issue.

Remember windows is relying on that CPU to remain functional while Prime95 is torturing it.

Just because Prime95 torture doesn't cause a rounding error in Prime95 doesn't preclude your CPU from being ever so slightly unstable resulting in an OS dll or background thread becomming corrupted and crashing windows in parallel.

This is why all errors are essentially thrown into the same bucket of "unacceptable". You are testing for system stability (hardware and OS), not just Prime95 stability.

 

[Markfw]

I had to set my bios as high as 1.45 on one of my motherboards, due to vdrop and vdroop. Once in windows, under full load, if you vcore=vid in coretemp, its a good indicator of proper voltage to OC from my experience.

 

[neosapien]

Success! 1.29375V was 8 hours stable in Prime95 small FFT, and 12 hours and 47 minutes stable in Prime95 blend, so it looks like I have reached my goal of 3.00Ghz and minimizing the vcore. Thanks for your help guys!

oh, temps. ~38c idle, 62c load

Q6600-B3 @ 3.00Ghz 1.29375V CPUZ