X3350 / Q9450 Overclocking Thread

[bryanW1995]

Originally posted by: cheinonen
My Q9450 finally arrived yesterday so I build my first new desktop system in a few years (have been running laptops for a while). I just changed the bus speed up to 400 for a quick OC to 3.2 with stock voltage on everything and started Prime95 before going to sleep. It ran fine for 9 hours with load temps of 60-63 across the cores (idle around 44-46), and CPU voltage of 1.125 (CoreTemp) or 1.17-1.18 (CPU-Z). Tonight I might try to set the bus speed up to 425 or so with the stock voltage and see if it does fine, but at the moment it's still whisper quiet, which is more important than that last bit of speed.

go to the bios to see what your vcore is. coreTemp is allegedly reporting default vcore, not the vcore you're currently running. Of course, coretemp shows me at 1.0375 while my ip35pro won't let me set vcore at lower than 1.25 in bios. I think I'll trust the mobo before coretemp. You're probably at 1.25 in bios, too, but I'm curious to see.

 

[BigMoosey74]

My q9450 isn't the greatest chip, but I have it stable at 3.4ghz with with 1.184 voltage.

That is a pretty good voltage. I would like to OC my X3350 like you did to keep it under 1.2V

 

[runawayprisoner]

Okay... new development, guys. If you are using an ASUS motherboard, and your motherboard lets you adjust your strap control, try setting your strap control at 266MHz and then set your FSB to 451 or anything above that!
We have investigated a bit more into the weird FSB walls on some boards, and have found out that strap control was somehow messing with it. Details... are all over the place in Q9450/X3350 threads in XtremeSystems. Special thanks to ChaosMinionX for finding out the FSB strap problem.
Now just get back into your BIOS again to see if you can't squeeze that extra 200 - 400MHz out of the chip! It has so much potential, you know!

 

[Tempered81]

Prisoner: do you have the best q9450/3350 Stable Overclock @ 3.9g?

 

[runawayprisoner]

I'd believe so? But that may change soon now that the X38/X48 folks have found their problem and a solution. I believe we'll see 4GHz stable soon.

 

[bryanW1995]

well, good news and bad news. good news is that 9450 is here, stock vcore is only 1.225 vs 1.25 for the X3350. bad news is that I was cleaning the zalman and freakin' lost one of the tiny fan screws, so now I'M screwed. This stock hsf is REALLY RINKY DINK. I am ok at 3.0 right now at least, but this is just depressing...

 

[PoopyPants]

one thing you guys want to be VERY VERY careful about and you can read about this all over the place, but as usual XS is the best place.

the Xeon and 9450 are dieing left and right, and i mean at an alarming rate at anything over 1.4 volts.

i will warn you right out,, be VERY VERY careful when hitting 1.4v

the 45nm chips cant handle that much voltage, you have to remember the smaller nm components inside are not the same 65nm parts... much smaller size and more of them,, voltage is this chips enemy,, avoid it.,

 

[bryanW1995]

I'm at 1.38v in bios, 1.27v after vdroop on the X3350. I only ran it once at 1.41, and that was to try to get orthos to run at 3.704. It failed after a few minutes and I promptly dialed it back down to 1.38v for 24/7 use. The default vcore on my x3350 is 1.25, I'm only at +.13 so I figure that I'm pretty safe there. My Q9450 has a default of 1.225, however, plus it's on an ip35-e, so I'll be a little more careful with it.

 

[PoopyPants]

My Q9450 has a default of 1.225

not sure thats right chief.. those chips dont have a default VID that high,, or if yours is, you better ebay it and get a good one.
VID on those shouldnt be above 1.1v from what i've seen atleast.
maybe i'm wrong ??

the IP35-E may not be handling it well.

 

[bryanW1995]

hmmm, well my ip35 pro won't let me go below 1.25v on my x3350 and my ip35-e won't let me go below 1.225v on my Q9450. My e6750 on the ip35 pro wouldn't allow me to go below 1.35v. Must be an abit thing.

According to coretemp my default vcore on both cpu's is 1.0375, I just have heard in many places that coretemp is incorrect. I'm really not concerned with default vcore or undervolting the chips, however, so this is no big loss to me.

 

[Xpoc]

1.125 here. On the box it says MAX is 1.225.

 

[runawayprisoner]

Originally posted by: PoopyPants
one thing you guys want to be VERY VERY careful about and you can read about this all over the place, but as usual XS is the best place.

the Xeon and 9450 are dieing left and right, and i mean at an alarming rate at anything over 1.4 volts.

i will warn you right out,, be VERY VERY careful when hitting 1.4v

the 45nm chips cant handle that much voltage, you have to remember the smaller nm components inside are not the same 65nm parts... much smaller size and more of them,, voltage is this chips enemy,, avoid it.,

Actually, only E8400 chips are dying. Some E8500 chips have died as well, but that's to be expected since they were pushed to death.
We are having FSB walls with X3350 and Q9450 chips so... for the time being, nobody would need above 1.4v for anything at all. Well, nobody except for me, that is.

 

[BigMoosey74]

Originally posted by: PoopyPants
one thing you guys want to be VERY VERY careful about and you can read about this all over the place, but as usual XS is the best place.

the Xeon and 9450 are dieing left and right, and i mean at an alarming rate at anything over 1.4 volts.

i will warn you right out,, be VERY VERY careful when hitting 1.4v

the 45nm chips cant handle that much voltage, you have to remember the smaller nm components inside are not the same 65nm parts... much smaller size and more of them,, voltage is this chips enemy,, avoid it.,

Dude, that is why I would never purchase a used CPU. If you get one from someone like Runaway...no offense...the d*mn thing has been worn down from over volting.

 

[Cutthroat]

My Q9450 (C1) can run at stock speed with vcore set to 1.1v in the BIOS which is the lowest vcore allowed, load line calibration is enabled so it shows up as 1.08 in Windows, Coretemp says the VID is 1.0625v. Currently I'm at 3560MHz (8x445) @ 1.275v, that's the best I can get.

 

[runawayprisoner]

Originally posted by: BigMoosey74

Originally posted by: PoopyPants
one thing you guys want to be VERY VERY careful about and you can read about this all over the place, but as usual XS is the best place.

the Xeon and 9450 are dieing left and right, and i mean at an alarming rate at anything over 1.4 volts.

i will warn you right out,, be VERY VERY careful when hitting 1.4v

the 45nm chips cant handle that much voltage, you have to remember the smaller nm components inside are not the same 65nm parts... much smaller size and more of them,, voltage is this chips enemy,, avoid it.,

Dude, that is why I would never purchase a used CPU. If you get one from someone like Runaway...no offense...the d*mn thing has been worn down from over volting.

No offense taken. But really, you should read up on the actual issue. There are people who run those chips at above 1.4v 24/7 and their chips have been running strong ever since launch day. A group of more unfortunate users banded together to find out the cause and have found out that changing the PSU helped re-stabilizing the processors in most cases. So a lot of "degraded" chips have lived once again.

As for the voltage, we thought it was vCore, but just now, Intel has confirmed for Anandtech themselves that Anandtech's QX9650 died not from over-volting vCore but from over-volting VTT. That might mean that up to above 1.4v in vCore, it's still safe. But VTT should not be too high. This is still being investigated.

 

[DeathSniper]

My Q9450 is currently @ 3.2Ghz with 1.2v

 

[aigomorla]

okey i thought id post this. I was running around looking for proof and it was right in our own forum.

This is expecially directed @ runaway. Sorry bro, your voltage and overclock seriously scares me and it seems like you dont know what your really doing:

http://www.anandtech.com/weblog/showpost.aspx?i=428

Your VTT is uber high my friend. You will kill your chip. You have been warned.

 

[bryanW1995]

vCore: 1.42v
vPLL: 1.76v
vFSB (VTT): 1.50v
vNB: 1.71v
those numbers are kinda scary...but if you have the $$$ then why not try to be the best?

aigo and mark, do you guys get major vdroop on vtt? I don't remember seeing any on my e6750, but I'm at 1.31 vtt in bios and uGuru only shows 1.20. Other than vcore I don't see vdroop on any other voltages. I've been doing some checking and abit does claim that the current bios is still not "fully" compatible with 45nm quads, however, so I'm unsure if this is a design spec, a board failure, a software glitch, etc...

 

[aigomorla]

Originally posted by: bryanW1995
e numbers are kinda scary...but if you have the $$$ then why not try to be the best?

aigo and mark, do you guys get major vdroop on vtt?

Im sorry i dont use anything besides DFI LT series boards now on my quads.

So the word vdroop got deleted in my vocabulary. ROFL....

But on the serious side, i really do only use DFI LT series. I went on a buying spree to replace the rest of my asus boards with the LT P35 series.

And the vdroop is uber low.

 

[runawayprisoner]

Originally posted by: aigomorla
okey i thought id post this. I was running around looking for proof and it was right in our own forum.

This is expecially directed @ runaway. Sorry bro, your voltage and overclock seriously scares me and it seems like you dont know what your really doing:

http://www.anandtech.com/weblog/showpost.aspx?i=428

Your VTT is uber high my friend. You will kill your chip. You have been warned.

No prob. I guess I scared some people with the X3350 chip results, right? I am more careful with the Q9450 now. The chip would run stable at

vCore: 1.38v
vPLL: 1.60v
vFSB: 1.36v
vNB: 1.51v

The above volts (that Mark posted) are...? You guessed it right. The maximum the motherboard can do. If those were 24/7 volts then yes, they would be scary. But they were only to boot and check out vCore for like a minute or two. I know the longevity of the chip can no longer be expected once it's been exposed to higher voltages, but seriously, I don't plan to keep the chip for longer than two years.

 

[GuitarDaddy]

Received my X3350 yesterday codes are SLAX2 L806A744

Seems like a pretty good chip, 3.2ghz was easy and prime stable with less than 1.2v, 3.4ghz took a couple of vcored bumps to get it stable and temps were in the high 50's after an hour of prime. I was able to get 3.6ghz prime stable but I had to bump vcore up to 1.38v inbios 1.32-1.36v idle per cpuz which drops to 1.25v under load, and the temps hit 69-70c.

So far the P965 P5B-E is holding its own with this chip, I'm up to 450mhz FSB on default NB volts and VTT

I'll probably stop here because of the horrendous vdroop on this board. I feel certain this chip has more, and I may try the pencil vdroop mod on the board and go for 3.8ghz after the thermal paste sets for a few days. I may even drop back to 3.2 or 3.4 for 24/7 use for now because I'm not too comfy with the idle vcore at 1.32-1.36v although the loaded vcore of 1.25v seems in the safe range.

What do you guys think? 1.36v too high for idle?

3.6ghz

 

[the unknown]

edit: wrong page

 

[bryanW1995]

Originally posted by: GuitarDaddy
Received my X3350 yesterday codes are SLAX2 L806A744

Seems like a pretty good chip, 3.2ghz was easy and prime stable with less than 1.2v, 3.4ghz took a couple of vcored bumps to get it stable and temps were in the high 50's after an hour of prime. I was able to get 3.6ghz prime stable but I had to bump vcore up to 1.38v inbios 1.32-1.36v idle per cpuz which drops to 1.25v under load, and the temps hit 69-70c.

So far the P965 P5B-E is holding its own with this chip, I'm up to 450mhz FSB on default NB volts and VTT

I'll probably stop here because of the horrendous vdroop on this board. I feel certain this chip has more, and I may try the pencil vdroop mod on the board and go for 3.8ghz after the thermal paste sets for a few days. I may even drop back to 3.2 or 3.4 for 24/7 use for now because I'm not too comfy with the idle vcore at 1.32-1.36v although the loaded vcore of 1.25v seems in the safe range.

What do you guys think? 1.36v too high for idle?

3.6ghz

that sounds pretty similar to my x3350. It runs 24/7 at 1.38 in bios, 1.31-1.32 at idle and 1.27 vdroop at load. I booted just fine at 3.704 but couldn't get stable for more than a minute or two even at 1.41vcore in bios, so I called it a day at 3.6.

 

[runawayprisoner]

1.41v in BIOS could just be 1.34v real, you know. Then again, 1.34v was enough for my X3350 to do 3.90GHz stable... so...

 

[bryanW1995]

ok, I've been reading about abit bioses and decided to flash my ip35-e with the Q9450 to bios 16. When I did I noticed that my vdroop significantly decreased, from about .12v at load to .05v at load. So I got crazy and went ahead and flashed the ip35 pro now, too. It has significantly improved vdroop, it went from (1.38-1.27=.11) to (1.38-1.31=.07). Now I'm feeling cocky and am trying 8x463 again... wish me luck! I might even go for 8x469 at 1.40v in bios if this remains stable overnight. Of course, my temps in this freakin' room are getting a little bit out of control with the two quads and wifey's laptop, so I might just drop it back down to 3.6 and decrease vcore to 1.34 or so. Either way I win!