Overclock.net banner
1 - 14 of 14 Posts

·
4GHz
Joined
·
1,767 Posts
Discussion Starter · #1 ·
Currently sitting (unstably, according to Prime95 and LinX) at 4GHz, and I'm not sure where to go from here. I'd like to make it to 4.25GHz on my P5Q Deluxe, but right now, vcore is at 1.52v and 4GHz isn't comfortably stable. Not to mention, it loads at 83C, which is pretty eurgh (Ambient ~22C).

Surely just adding more vcore isn't the way to stability? I know I'm on a C1 Q9550, but that can't be hampering my efforts that much, can it?

All the fancy settings (PLL and NB and clock skew and whatnot) have been left on Auto or have been otherwise untouched - perhaps there lies the secret to 4.0+.

Also, what are the odds I'm limited because I'm running 4 sticks of DDR2?
 

·
Premium Member
Joined
·
4,776 Posts
I'm not familiar with that board but, I can tell you this: if your vcore is at 1.52 actual (like in CPU-Z) you are waaaay past the limit. That's a helluvalot of voltage bro.

Try this: take 2 of your RAM sticks out and then try and get the cpu stable, with only 2 sticks in and WITHOUT oc'ing the memory. (Keep the RAM at or even below the stock speeds so you only have to trouble shoot the cpu for now.) That way, when you do get the cpu stable, all you have to do is pop the other 2 sticks back in and then mess with the NB voltages.

I'm sure someone with experience with that board will help you out eventually.

Wish I could help out more bro.

Edit: Q9550 + P5Q DELUXE.

P5Q OWNER'S.

Good luck.
 

·
Registered
Joined
·
1,616 Posts
I'd say you found your problem on your own!
Quote:
Also, what are the odds I'm limited because I'm running 4 sticks of DDR2?
This will put a lot of added stress on your cpu, and the fact that your mixing brands of ram. Just not a good idea!
These factors will limit your ability to overclock!
Quote:
Memory
[4x2GB|DDR2-1000/1200|5-5-5-15][G-Skill|OCZ]
Not good!

Take out 1 set, and test again!
 

·
Registered
Joined
·
2,934 Posts
1.512 way too high for 45nm...

thats near the max for 65nm.. your gona burn your chip up.

thumb.gif
 
  • Rep+
Reactions: yoyo711

·
Registered
Joined
·
1,616 Posts
Quote:
PLL and NB and clock skew
PLL - 1.62v
NB - 1.45v
clock skew - delay 100

Take your vcore, and keep lowering it till it wont boot up, then raise it a few ticks, and test for stability, with only the G.Skill memory installed.
 

·
Premium Member
Joined
·
4,635 Posts
I have managed to run my P5Q Pro @ 506MHz FSB and it was stable in games,3DMark 11,Cinebench but it would error on cores i P95 Large FFT test..

P5Q Deluxe is better mobo and it has better CPU GTL and NB GTL options in BIOS,but still I'm almost sure that 500MHz FSB stable in P95 Large FFT test is not possible on it..

Maybe ~ 485MHz FSB is more realistic overclock..

And KEEP Vcore UNDER 1.4V (UNDER LOAD) AT ALL TIME..

KEEP YOUR CORE TEMPERATURES UNDER 75C AT ALL TIME..

CHEERS..
 

·
Premium Member
Joined
·
4,776 Posts
Quote:
Originally Posted by Inteller;14243597
PLL - 1.62v
NB - 1.45v
clock skew - delay 100

Take your vcore, and keep lowering it till it wont boot up, then raise it a few ticks, and test for stability, with only the G.Skill memory installed.
OMG, I didn't even see that.

@ DEC42: yeah bro, you want to take one of those mismatched sets out for now. Leave one set in and do your oc'ing. That could be causing you some serious issues right there.

And just my
2cents.gif
but 1.45v is the highest I've ever gone in CPU-Z. But my temps have never gone above 58c under load (PRIME95).

You are definitely in the Danger Zone at that voltage man. Couple of years back AlMighty did an experiment with a 45nm chip and at 1.6v it burnt the pads. 1.6v 24/7 The chip still worked but the pads were burnt and that means the cpu socket itself was soon to go too. Your hardware, just a heads up for you.

Good luck.

___________________________________________________________

And if you don't mind my asking:
What TIM are you using?
What fan/s do you have on that Mega?
Is it tightened ALL the way down?
 

·
Premium Member
Joined
·
4,514 Posts

·
4GHz
Joined
·
1,767 Posts
Discussion Starter · #10 ·
Quote:
Originally Posted by crashdummy35;14246010
OMG, I didn't even see that.

@ DEC42: yeah bro, you want to take one of those mismatched sets out for now. Leave one set in and do your oc'ing. That could be causing you some serious issues right there.

And just my
2cents.gif
but 1.45v is the highest I've ever gone in CPU-Z. But my temps have never gone above 58c under load (PRIME95).

You are definitely in the Danger Zone at that voltage man. Couple of years back AlMighty did an experiment with a 45nm chip and at 1.6v it burnt the pads. 1.6v 24/7 The chip still worked but the pads were burnt and that means the cpu socket itself was soon to go too. Your hardware, just a heads up for you.

Good luck.

___________________________________________________________

And if you don't mind my asking:
What TIM are you using?
What fan/s do you have on that Mega?
Is it tightened ALL the way down?
Well, I used to have it at 1.47v on my BioStar TP35D2-A7, but that was because the vdroop on it was unholy. I guess I'll have to be more diligent with my vcore on this board.

I don't think I really want to run the chip at 1.5v+ 24/7, so, don't worry too much about that
wink.gif
.

TIM is AS5, long cured. Fans are some 120x25mm Antec case fans that I don't know all too much about. I'm certainly sure that the Mega's tightened all the way down.
 

·
Registered
Joined
·
997 Posts
DEC_42: I had stability issues at 470FSB+ on my P5Q too, and it was the RAM failing, so I'd isolate RAM test in P95 to confirm that in your case. The solution for me was to set RAM timings to Auto; helped even more than setting performance level to 11 which just gimped performance without adding stability. PLL is definitely key, I just instinctively set it to 1.6 now whenever I venture over 450 FSB
Quote:
Originally Posted by crashdummy35;14246010
You are definitely in the Danger Zone at that voltage man. Couple of years back AlMighty did an experiment with a 45nm chip and at 1.6v it burnt the pads. 1.6v 24/7 The chip still worked but the pads were burnt and that means the cpu socket itself was soon to go too. Your hardware, just a heads up for you.

Good luck.
I find it funny you say that because we have the same chips and I take mine to 1.65 regularly and have never had any problem. If you read the thread you linked to, you'll notice the guy had 120 degrees at one point which is far more likely to be the culprit of any sort of damage than simply running a lot of juice through the chip.
 

·
Registered
Joined
·
3,983 Posts
Quote:


Originally Posted by KingT
View Post

I have managed to run my P5Q Pro @ 506MHz FSB and it was stable in games,3DMark 11,Cinebench but it would error on cores i P95 Large FFT test..

P5Q Deluxe is better mobo and it has better CPU GTL and NB GTL options in BIOS,but still I'm almost sure that 500MHz FSB stable in P95 Large FFT test is not possible on it..

Maybe ~ 485MHz FSB is more realistic overclock..

And KEEP Vcore UNDER 1.4V (UNDER LOAD) AT ALL TIME..

KEEP YOUR CORE TEMPERATURES UNDER 75C AT ALL TIME..

CHEERS..

Listen to this guy.
 

·
Premium Member
Joined
·
4,776 Posts
Quote:


Originally Posted by Petrol
View Post

DEC_42: I had stability issues at 470FSB+ on my P5Q too, and it was the RAM failing, so I'd isolate RAM test in P95 to confirm that in your case. The solution for me was to set RAM timings to Auto; helped even more than setting performance level to 11 which just gimped performance without adding stability. PLL is definitely key, I just instinctively set it to 1.6 now whenever I venture over 450 FSB

I find it funny you say that because we have the same chips and I take mine to 1.65 regularly and have never had any problem. If you read the thread you linked to, you'll notice the guy had 120 degrees at one point which is far more likely to be the culprit of any sort of damage than simply running a lot of juice through the chip.

Well, I'm just trying to help the bro out "conservatively".

We have the same chip but mine is total crap. Takes me 1.462v in CPU-Z for 4.15GHz, any less and it fails PRIME95. It's been at 4.05GHz with 1.440v CPU-Z for the last 2 years. I just broke it down for a cleaning about 2 weeks ago and I checked the pads and my socket and everything looks cool. Still only reaches ~59c in PRIME and never goes above 47c in a game, and that's cause I'm using the PK1 TIM that comes with the Mega for the moment; temps were about 4 degrees lower using my old OCZ FREEZE.

almighty had some high temps because he was having trouble with his loop. The chip stayed under water the entire experiment. But in the end it burnt at 1.6v CPU-Z.
__________________________________________________ ____

@ op, so what's the latest?
How's the oc going?
Also, if you've got the room in your case and you have an old fan laying around, try making a shroud for the push fan on your Mega. I always thought that it was more trouble than it was worth but, in the end, it lowered my load temps by ~4 degrees on average. Shrouds work...no doubt about it. Check out DUCKY's tut on making the shroud and why it helps.
 

·
4GHz
Joined
·
1,767 Posts
Discussion Starter · #14 ·
Got a Delta for the Mega. Damn loud, but the shroud helped. Vcore is 1.39, PLL is 1.64, and NB is 1.46v, but it's still not LinX stable. Loading at 77/78C, though. I'm pretty damn sure that the Mega's mounted right, but perhaps I need to remount the whole thing.
 
1 - 14 of 14 Posts
Top