OC'd Q6600 G0 to 3Ghz.. wondering about my voltage settings 680i board

relly

Distinguished
Jan 6, 2011
16
0
18,510
Before I begin, here are my system specs:
OS: Windows XP SP3
CPU: Q6600 G0 Stepping <@ 2.4Ghz stock> w/Tuniq Tower 120 Cooler (fan on full throttle)
GPU: GTX 460 1GB (stock)
MB: Asus P5N32-e SLi with latest bios 1903 (Website-->) http://usa.asus.com/product.aspx?P_ID=vBZLIBtPzYB2bLcb
Memory: 2GB dual channel pc2-6400 Crucial Memory.. (I forget what model but it was definitely on the qvl of my moboi)
Antec 900 Case which has tons of fans, keeps everything pretty chilly..

I have a pending thread in another section here which discusses my next build (2600K based) but I just want to mess around with this rig for now. But I also don't want to blow anything up lol

I finally got my Q6600 @ 2.4Ghz to OC above 2.81Ghz (was never getting pass bios screen if I tried anything) which was through the AI overclocking at 10% (15% and 20% wouldn't boot either). I finally decided to go through manual options and after reading a few topics about my same CPU and mobo.. so I copied some of the volts they used when they said were stable when OC'd at 3Ghz (CPU-Z reads it at 3000.5 MHz and keeps changing but just the decimal spot)(Core Temp reads it at 2999.71MHz)(Windows shows 3GHz)

I believe my OC'd is running stable. I ran intelburntest, which I read was supposedly good for testing OCs. I could only run the standard option, since I have 2GB ram and for some reason memory available never got close.. I passed all 5 tests, averaging around 35 seconds to complete (each stress test). Core Temp shows my CPU is running on each core at 49C 48C 39C 40C on idle. I also had speedfan but I really don't know how to read it.. it also shows CPU temp at 64C (same time I checked Core Temp).

Anyway, since it seems stable (haven't run 3D games or 3D benchmarks yet) so far.. I was wondering if I could keep using the voltage I'm using or gradually go down and keep testing.. Here are the voltages I'm currently using:

Vcore: 1.4375v (I tried something higher at first but it crashed so bad it reverted my bios settings lol)
Memory: 2.1v (since thats what it's rated at)
1.2VT: 1.35v
NB: 1.35v
SB: 1.5
CPU VTT:1.35v

FSB in bios is 1330 (333.30x9)
Memory Timings at 4, 4, 4, 12, 2T and at stock 800Mhz (unlinked)

All Spectrum options disabled, basically everything in that one screen is disabled.. Speed step and GPU ex options disabled..

Are my voltages off? I'm stable at 3GHz for the first time (after so many failed attempts) and I like the speed improvement so far (apps launching much quicker).

Thanks in advance!

Also, I'm going to be updating this thread with further benchmarks to see if my computer locks up or reboots.

 

lothdk

Distinguished
Jan 20, 2010
881
0
19,160
It has been a long time since I fiddled with my Q6600, but if I remember correctly, the max Vcore is 1.5, and from memory, the voltages you have set seems like they match a 3.6 GHz overclock, so you should back down the voltages if that is where you want it at.

Next thing, download Prime 95 and run it 5 - 15 minutes when you are finding (increasing FSB and voltages) your max and stable overclock, once you are there, run it 8+ hours (overnight) to check true stability.

Close to 50c idle is too much, check that you have correctly installed your heatsink, maybe you used too much/little thermal paste?
 

relly

Distinguished
Jan 6, 2011
16
0
18,510
Before the OC my CPU ran about 30-40c idle. The rest of my system is really chilly. 24-35c.

What voltages should I try lowering first?

I ran a 3dmark06 and got my highest score yet.. Almost 2000 points (16,443) higher than my previous test at 2.735ghz (scored 14,674) But core temp read 83c high on full load (1st core was highest, all other 3 were lower)

Maybe the voltages are sending too much power through my CPU.. Which is why I want to know which ones should I start fiddling with first? Thanks for your reply

 

relly

Distinguished
Jan 6, 2011
16
0
18,510
After various Bios changes by dropping the voltage on the 1.2 HT to 1.3v/ NB to 1.25v/Memory to 2.2v (thats what it's actually rated)/CPU VTT to 1.45v/Vcore to 1.375v right now and trying to lower it still.. Core Temp is reporting my card at 40/40/32/32 idle (84/83/65/65 full load) but Everest is reading it at 29/30/24/22 idle (60s in load).

Which one should I believe?
 

lothdk

Distinguished
Jan 20, 2010
881
0
19,160
That is quite a difference, try and download HWMonitor from cpuid.com, it will likely show the same readings as Everest, as those seems to be about right, depending on your average room temperature.

Also, if you have not already, grab CPUZ while you are at that site, it is a great tool to check CPU and memory settings, BIOS revision and so on.
 

relly

Distinguished
Jan 6, 2011
16
0
18,510
Back with an update. There is definitely something causing instability, but I don't know what to rule out. I thought my system was stable until trying to play COD Black Ops. Right as the map loaded, my PC instantly crashed and restarted. Every other test and benchmark with those settings showed stability with zero errors and reboots. So I raised my 1.2 HT to 1.35 and left everything else the same (vcore: 1.337v lol, Mem: 2.2v, NB: 1.35v, SB: 1.55v, VTT: 1.45v & temps @ 36/36/28/28 idle *66/63/54/51 load using Real Temp)

Here are the tests/benchmarks I ran:
Prime95:pure CPU test for 7 hours, No errors
Prime95:pure Memory test for 3 hours, No errors
Used the Real Temp Sensor test with prime95 it completed the cool down test
Sisoft Sandra: Ran CPU, GPU, Memory and basically all the tests and passed with no errors or reboots
Intel Burn Test: Passed (69.5 seconds and the speed was 25.6)
3dmark06 (16453)
Mafia 2 Benchmark @ 1080p (everything on high, AF off, AA x16: average 60.3fps
Counterstrike Source Video stress test @ 1080p (everything on max): 321.07 fps
Call Of Duty: Black Ops single player: After clicking Resume the screen went black and the pc restarted
Call of Duty: Multiplayer: After the map loaded, the same happened. Screen black, restarted on its own.
After bumping up the 1.2 HT to 1.35v the games both lasted longer (got through a few checkpoints in SP and 2 map reloads in MP-roughly about 20-35 minutes of playing) but the same thing happened. Only this time the screen would sometimes freeze with whatever was playing on it, no hard drive activity on my case indicator and caps lock doesn't react-system is frozen forcing me to cold boot)
I reloaded BIOS defaults to 2.4GHz and rebooted, but after 10 to 15 minutes they both crashed again, either freezing or rebooting. So I reloaded the 3GHz OC again since that didn't seem to be the problem.
Mafia 2 did crash once during free roam mode and driving through a tunnel, but only on one occasion. (I've played hours without one crash)
Also my system froze while browsing YouTube after 10 straight hours of Prime95 and I had to cold boot.

I've reinstalled the latest nvidia drivers thinking maybe it was GPU related, but it still crashed. I'm about to reinstall my chipset drivers.. and Call of Duty too. Any ideas?

System again:
q6600 @ 3Ghz (333x9 FSB) w Tuniq Tower cooler
GTX 460 1GB (stock)
2 GB Crucial Ballistix DDR2 800Mhz (4,4,4,12,2T as rated and 2.2v)
p5n32-e Sli Motherboard (nforce 680i)
Antec 550w PSU
Windows XP Pro SP3
All latest driver updates installed

 

relly

Distinguished
Jan 6, 2011
16
0
18,510
Those timings are stock and were always the most stable ones. Messing with them always caused instability and I even noticed a performance loss with some.
My TJmax is 100C, and after changing the voltages around a bit (mainly using a lower vcore) my load temps dropped about 20C.

Back with another update. I reinstalled the chipset and basically every nvidia driver (used driver sweeper/cleaner right before) and ran a 3dmark, it completed it. So I played cod sp and mp for about 3 hours @ 1080p texture on extra with no crash at all. So I browsed the net for a while and decided to run a Mafia 2 benchmark. 15 seconds into the benchmark my screen went black and my pc rebooted. So I ran a CPU prime95 test (using the custom prime 95 settings from a thread here) and it's been testing for 16 hours so far with no errors, I'm just wondering when it's gonna complete the test on it's own, or if 16 hours is enough to show the cpu is at least stable?

I'm about to OC my video card, idk if that'll help but what else am I gonna do?
 
if feel stable at RAM and what software did you use, if something did not work all the better to go safe mode and uninstall software uninstall GPU oc vga driver, after that you just reinstall the vga driver from the cd bundle, the test with your game, if it worked and you can vga soft latest updates, and test again if not successful then you have to go with the vga driver software support provided by the vendor.
 

relly

Distinguished
Jan 6, 2011
16
0
18,510
Mafia 2 crashed again after 24 hours of prime 95 cpu test.

I've bumped my RAM speed to 1066mhz. Running tests to check if its stable. Right now Prime95 memory test.

 

relly

Distinguished
Jan 6, 2011
16
0
18,510
Bump.

I'm still having stability issues. I'm back at stock settings until I figure this out :(

It's not my GPU.. it's definitely the OC voltages that are failing. Should I bump them up?

and what about my RAM speeds? Should I use a certain ratio when trying to be at 3GHz?
 

CoreTemp. Unless your computer is in a really cold room with low ambient temps, the Everest reading of 22 C is definitely bogus.

And the problem with using the BIOS to monitor temps is that by the time you have rebooted and POSTed, the CPU will have cooled.