need help overclocking

Nvidiaguy07

Platinum Member
Feb 22, 2008
2,846
4
81
im trying to OC my Q6600 on a DS3L board.

My temps never go above 54degrees C while at load. no matter what combo ive tried.

My last attempt i set my FSB to 400 to make it a 3.6MHz overclock.
my ram is rated for DDR2800 and my ratio is 1:1 so i dont have to add any volts to that right? (rated for 1.8-1.9, its set on 1.8 in the bios)(my setting are still factory for the memory 5-5-5-12).

im not sure if i should be adding any voltage to the FSB so i havent touched that at all.

The V-core i have a feeling is whats holding me back, im just confused as to what V-core reading i should be using and dont want to give it too much juice.

the last thing i tried was to set it to 1.5125 in the bios
CPU-Z at full load while running prime95 had a reading of 1.392 ( i forgot to take my idle reading but it had a good amount of v-droop)
and the PC health setting in my bios had my v-core at 1.46

it seemed ok for a while and then froze up after running prime95 for about 20-25 mins.

am i going too high with these voltages?

is there any way i can hurt my card? ( i have TM2 enabled in bios so i think thats protecting me)

im currently trying out a v-core of 5.25 in the bios.

what else could be causing my freeze ups besides low voltage? FSB voltage or memory voltage?
 

chinaman1472

Senior member
Nov 20, 2007
614
0
0
Try +0.1V on the FSB, the RAM, and the MHC (or something like that). Adding 0.1V shouldn't overvolt anything at all.
 

LOUISSSSS

Diamond Member
Dec 5, 2005
8,771
58
91
first of all, i hope you are checking your temps using real temp 2.0 and core temp 0.97. 1.5vCore on a Q6600 is intel's MAXIMUM suggested. most people on air cooling and 70F room temperature don't get good temps with 1.5vCore.
second of all, u said 5.25 vcore in bios, thats wrong... it should be under 1.5

u must first ensure that all ur parts are running stable at STOCK speeds. run P95 v25.6 Small FFT for 24 hours to make sure ur system is stable
to OC: u shouldn't jump straight to 3.6ghz u need to work ur way up there to get a "feel" for ur chip to see what kind of vcore it requires. change ur fsb and multiplier to 333 x 9 for a 3.0ghz clock speed. Run P95 again for 24 hours, then bump to 367 x 9 for a 3.3ghz clock and test again

 

Nvidiaguy07

Platinum Member
Feb 22, 2008
2,846
4
81
ive had my CPU stable at 3.2 GHZ and 3.4 GHZ but i just cant seem to get 3.6 just it freezes up no matter what.

the temps are correct ive used speed fan with the +15 adjustment and also coretemp (i had to use 0.97 the previous version didnt work on vista x64)

ive tried lower voltages but windows freezes up before i can get to my desktop.

my computer is in a room that it mayeb 65-70 degrees with all the fans turned up.

i meant to say 1.525 in the bios, my only guess is that i need more juice because my temps are fine.


 

Nvidiaguy07

Platinum Member
Feb 22, 2008
2,846
4
81
i have a feeling i need to increase my FSB/memory/south bridge voltage but im not sure what is safe to do.

my memory is rated at 1.8-1.9 and the bios is showing something like 1.84 <am i safe to increase by 0.1 or 0.2?

on a DS3L what is safe to OC the FSB voltage to? (stock northbridge heatsink)
 

Nvidiaguy07

Platinum Member
Feb 22, 2008
2,846
4
81
My last try:
bios v-core: 1.51875v
pc health reading in bios: 1.476v
cpu-z idle: 1.456v
cpu-z load: 1.392v

my idle temps were 26-25-20-25
my load temps while running small FFT's in prime95 were about 60-60-58-58 at thier highest

i ran prime95 with no errors and no freezing or anything for 33 mins with 0 errors (completed 2 runs?) things were looking good....

but then i tried doing a benchmark in prime95 and my screen froze!!!

so would u recommend more voltage? i take it the v-core that shows in CPU-Z is the one that counts because 1.5125 is alot.

i increased my DDR2 voltage and FSB voltage by 0.1 just to see if that helps my instability.

like i said i cant get stable and its not the temp that is limiting me...

PLEASE TELL ME WHAT U GUYS THINK!

 

LOUISSSSS

Diamond Member
Dec 5, 2005
8,771
58
91
Originally posted by: Nvidiaguy07
My last try:
bios v-core: 1.51875v
pc health reading in bios: 1.476v
cpu-z idle: 1.456v
cpu-z load: 1.392v

my idle temps were 26-25-20-25
my load temps while running small FFT's in prime95 were about 60-60-58-58 at thier highest

i ran prime95 with no errors and no freezing or anything for 33 mins with 0 errors (completed 2 runs?) things were looking good....

but then i tried doing a benchmark in prime95 and my screen froze!!!

so would u recommend more voltage? i take it the v-core that shows in CPU-Z is the one that counts because 1.5125 is alot.

i increased my DDR2 voltage and FSB voltage by 0.1 just to see if that helps my instability.

like i said i cant get stable and its not the temp that is limiting me...

PLEASE TELL ME WHAT U GUYS THINK!

i think u need to back to ur 3.2 and 3.4 ghz and do some more stability testing. 33 mins is NOTHING for system testing. u need a good 24 hours to say ur pretty stable... on P95 v25.6 anything else and you're just lying to yourself
 

TheJian

Senior member
Oct 2, 2007
220
0
0
Agree with louisss. 33 minutes doesn't even count as starting in prime. You can't begin to even say stable until you hit 12hrs (and I'm being REALLY GENEROUS here). I say 24+ also. To give you an example I ran Serious Sam 2 (i like it for working a graphics card in high res...It's a very stable game, although a bit weak it's rock solid) for 17hrs. But prime couldn't get to 3hrs. I think sam2 could have run for days. Prime is another story. I'd shoot for 3ghz and 24hrs, then go up from there. Patience sucks I know.

Having said that going up one notch on the other voltages is fine. In fact these boards are designed for that (I've sold plenty of the DS3L all solid cap boards mind you, and I have a DS3R myself). I say 333x9 so you can test your whole system though. Everything should be at default volts then (except maybe the cpu). That board is made for 1333 with defaults everywhere. So slapping your 2.4/1066 on 3.0/1333 is perfect for the rest of the system to ensure TOTAL stability. Right now you're throwing so many things into the equation we're all left guessing. Minimize the variables. At 1333fsb/3,0 only the cpu is being screwed with. Just make sure your memory is set correctly for 800. It'd be helpful to know what memory you have. Give us a link to it at say, newegg. Or tell us what it says it needs for voltage at the timings they recommend. Like mine is 1.8v at 5-5-5-18. Standard Jedec. If yours says the same but at 1.8-1.9 then go 1.9. Sure you can run lower maybe, but just go to whatever the high number is on the voltage recommended by manufacturer. You can figure out how low later. If you're running your Q6600 over 1.5 I think you've gone to far and your problem is elsewhere (unless you use liquid/ln2...heh).

You may have a core like a 2160 I had recently where it wouldn't go any more over 2.66 no matter how much I added (well, I could've hit some funky number around 2800/2900 but I wanted nice round figures). Just couldn't get rock solid at 3.0. I went from 1.3-1.45 and just got barely more stable (like almost to logon @1.32, then at logon 1.34, then logon and desktop at 1.4, fired up prime 1.45 but crash in 3minutes...nothing better after that). All other things were fine (because I had done 24hrs at 2.4 previously without messing with ANY thing else but the chip...memory set 800, same volts that passed 2.4 were set for 3.0 etc..all the same but chip). It was easy to see the cpu was the problem. Nothing else changed from the other perfect settings at 2.4 (by perfect I mean only the cpu is screwed with). I just upped fsb from 800, to 1066, to 1333 for 1.8ghz, 2.4ghz, and 3.0ghz). All of the rest are defaults with normal 3.0ghz chips that are 1333fsb. 1333 is a regular bus speed, 800 memory is normal etc. I ended with 8x333 for 2.66. And volts at 1.29. Once I found 2.66 seemed at near top limits I went down on volts until crapout. I was just below 1.3 for rock solid, 1.2950 or something like that. Cheapo PSU's flucuate too much. I could've probably had 1.28 on a PC Power and Cooling PSU :) I love those. But customers won't pony up for them. Anyway, NARROW your scope here. At this point you could have a bad board, crap dimm etc. We have no idea it is solid at 2.4 even. Let alone skipping 3 and heading to 3.6. I love your ambitious attitude but back the truck up a bit first :)