• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Should have bought I7...

SilverSliver

SilverSliver

[H]F Junkie
2FA
Joined
Feb 23, 2004
Messages
13,178
Well, I bit on the price dropped Q9550 a few weeks back thinking it would satiate my desire for something newer and faster. I conveninently put the fact that my motherboard is a gigantic pile of steaming crap in the back of my mind, and bought the shiny Q9550.

It's a good chip. E0, runs 3.4 on completely stock volts.

But my motherboard is horrendous. Can't get it stable at 3.6 -- despite adding ridiculous volts to the NB. And it won't even boot at a FSB above 444 or so in an attempt to get 3.8. I hate you, DFI. I will never buy one of your boards again.
 
the dfi x38/x48 boards never did oc the 45nm quads too tough. maybe a bios update? consider a dfi or asus p45 board.
 
Newest bios. I figure if I am going to rip apart my rig for a new mobo, it's going an be an X58.
 
thats a whole other story. dfi and asus make a bunch of good p45 boards between $100 and $140. with x58 you are looking at a cpu and ddr3 on top of the board cost.
 
SilverSliver said:
Well, I bit on the price dropped Q9550 a few weeks back thinking it would satiate my desire for something newer and faster. I conveninently put the fact that my motherboard is a gigantic pile of steaming crap in the back of my mind, and bought the shiny Q9550.

It's a good chip. E0, runs 3.4 on completely stock volts.

But my motherboard is horrendous. Can't get it stable at 3.6 -- despite adding ridiculous volts to the NB. And it won't even boot at a FSB above 444 or so in an attempt to get 3.8. I hate you, DFI. I will never buy one of your boards again.
Click to expand...


quit blaming the board for one.. its the northbridge thats the problem.. maybe you should of done some research because this is a known issue with the 45nm C2Q's and its not just intel northbridges that are the problem.. nvidia's northbridges have the same problem as well.. there are only a few northbridges out there that can sucessfully overclock a 45nm C2Q over 400mhz fsb and that number gets lower the more you go over 400mhz fsb..

the p45 and p35 seem to be the most consistant northbridges that allow for 450+ FSB..
 
sirmonkey1985 said:
quit blaming the board for one.. its the northbridge thats the problem.. maybe you should of done some research because this is a known issue with the 45nm C2Q's and its not just intel northbridges that are the problem.. nvidia's northbridges have the same problem as well.. there are only a few northbridges out there that can sucessfully overclock a 45nm C2Q over 400mhz fsb and that number gets lower the more you go over 400mhz fsb..

the p45 and p35 seem to be the most consistant northbridges that allow for 450+ FSB..
Click to expand...

The northbridge is on the board, so therefore, its the board in my eyes.
 
sirmonkey1985 said:
quit blaming the board for one.. its the northbridge thats the problem.. maybe you should of done some research because this is a known issue with the 45nm C2Q's and its not just intel northbridges that are the problem.. nvidia's northbridges have the same problem as well.. there are only a few northbridges out there that can sucessfully overclock a 45nm C2Q over 400mhz fsb and that number gets lower the more you go over 400mhz fsb..

the p45 and p35 seem to be the most consistant northbridges that allow for 450+ FSB..
Click to expand...

The northbridge is on the board, so therefore, its the board in my eyes.

But honestly, would it be worth the trouble getting a "step down" in motherboard to a P45 or the like to get, perhaps 400 daily mhz more out of the chip versus just sucking it up and getting an I7?
 
my dfi x48 and x38 only do 3.2 ghz at stock volts. i couldn't figure out how to go past that for my Q9450's. i guess i should just be happy :p (i am ;))
 
SilverSliver said:
The northbridge is on the board, so therefore, its the board in my eyes.

But honestly, would it be worth the trouble getting a "step down" in motherboard to a P45 or the like to get, perhaps 400 daily mhz more out of the chip versus just sucking it up and getting an I7?
Click to expand...


depends if you want to spend $100 to $150 or around $500 to $600.
 
NukeULater said:
My Abit IX38 Quad GT (intel X38) can hit 467MHz FSB (7.5x) no problem. It will go to 473MHz with a bit of trial and error. My Q9450 tops out about 3550MHz, after that the voltages start to get really high. So I normally run at 440Mhz FSB with a 8x multiplier.

http://www.cpuoverclockers.com/settings/658-intel-q9450-c1-ghz

I guess I was just a lucky one.
Click to expand...

yeah my brother has that same board and hes running his Q9550 at 3.8ghz at a 450fbs with around 1.34vcore and 1.45v for the nb. its not a bad board but its too bad abit went out of business... no more bios updates :(
 
Im running a q9550 at 447x8.5=3800 all day prime95/intelburntest 50 loops %100 stable, on a shit ass 750i northbridge. If pushed a little harder it does 470fsb and runs that puppy at 4000mhz so its not the quad core. That is on a proven to be a Piece of shit asus P5N-D that is currently still plugging away. I dont see how you can say the 45nm quads wont do over 450fsb+ on "only a handfull of chipsets". My work comp also has q9550 and I have ran it at the same exact 470fsb on a 3 different generic Abit ip35-e's, 2 asus p5nkl's , 2 asus p5q P45's, and they all work fine. I have tested both my home and my work q9550 in all 4 different boards with the same results. Sounds like you just need to tinker with it more, also biggest change for me on all of the asus boards has been to fix the v-droop so its more stable voltage going to her. Also tested running corsair xms-2 in all four slots, OCZ reapers 1066mhz 8-gig, Crucial ballistix Micron based Pc8500 8-gig, now settled on corsair dominator pc8500 8-gig set. I think that rules out being limited by running 2gig sticks in every slot, memtest/prime95 blend stable. These boards also been tested with my 2 different q6600 that need piles of voltage to keep running at 3600-3950mhz. Just dont give up and check your voltage droop and document every step you make on a notebook so you dont get into the dreaded stair at the screen fist slam from testing to many settings. Be scientific :D I will send you some bios settings if you would like them, between my nvidia chipsets and p35/p45 combo's you should beable to get a baseline from me. Also my p35 $80 abit IP-35E rocks out my e5200 at 4500-4600mhz just as good as my $180 Asus P5Q P45 chipset board if your looking for a cheap upgrade. Poor little e5200 :p
 
Well, I am currently stable @ 3.6 now. 1.28vcore, 1.4v mch, 1.2 vtt. Ram is at stock timings @ 2.1v. 3.7 was unstable at 1.31vcore, 1.4v mch and 1.3vtt. Core error with OCCT, booted into Windows though.
 
Thats getting there, try 1.45MCH around 1.36-1.38v-core, 447x8.5 should net you 3800mhz 5:6ratio divider and ram at 1067mhz (set divider if you need 800mhz ram speed). This is what im at but running a lower voltage so you might want to just pump the MCH up to 1.45+ and leave core voltage where its at to try 3800mhz. DO NOT RUN alot of vtt as its highly detrimental to 45nm cpu's even with dice/ln2/phase/water. Try 1.3vtt if you seem to not gain by bumping v-core and MCH up as a test. Also look at cpu-z at idle and record your v-core then put a load on it like prime95 or orthos, etc... and see if your getting any v-droop or lower fluctuating v-core while %100 load. If so every board ive come across with bad v-droop is just a bitch to keep stable and you have to overvolt it to come out where you need it while loaded. This is why I recommended 1.36-1.38 to try incase of bad v-droop.

As soon as I had stable voltage from v-droop pencil or hard mods overclocking is a breeze, and if possible you can verify with Digital multi meter to compare to bios and cpu-z output.
I can prime95/intelburn 4000mhz stable but the voltage required is to much of a jump 1.45+vcore along with temps, but lower it to 3950mhz and it will only take 1.39-1.40 to hit it.
 
Thanks for the help Manist.

Here are my current stats. Stable at 3.7. 1.28v, 1.45mch, 1.2vtt -- automatic ram timings. I'm a bit hesitant for any higher vcore, my temps are already climbing -- opinion? edit -- as soon as I posted this, it locked up.

37.jpg
 
Last edited:
SilverSliver said:
The northbridge is on the board, so therefore, its the board in my eyes.

But honestly, would it be worth the trouble getting a "step down" in motherboard to a P45 or the like to get, perhaps 400 daily mhz more out of the chip versus just sucking it up and getting an I7?
Click to expand...



well DFI doesnt make the northbridge..


the advantage with that DFI board is you have a lot more room to fine tune settings.. i wish i had those kind of options on my biostar board.. but i think thats why i keep my DFI LP UT NF3 board around since its so much more fun to overclock on it..


btw get realtemp 3.. and your temps look perfectly fine.. if you were able to atleast do some of the p95 test.. then its most likely cpu voltage..try 1.3-1.32v and see what she can do.. you should be able to hit 4ghz though.. probably going to need a lot more tweaking.. ya still got a little ways to go on the temps before you need to start worrying.... if you are worrying about 71C being high.. the i7 might drive you insane.. :p


one note.. make sure you watch your ram clock that you arent pushing it to far.. though im sure you already know that but figured id state the obvious just incase.. :p
 
Last edited:
First thing I thought about when I woke up, do you have speedstep and c1e and all of that thermal garbage turned off? If I let my quads declock with speedstep and c1e it will not supply the voltage it needs to keep it stable and crash before hitting login screen. Also one very important setting for quad cores is GTL reference voltage set it at 67% if you have that option, also if your already at 63% GTL try 65% before going to 67%. I find that to be the most important setting when cranking out quad cores be it 45nm or 65nm.

Also dont be a shy giving her a little more v-core to get it at 3.7-3.8 if you give it 1.4 in bios its probably about 1.36-1.38 actuall after being under load. Even with my aftermarket heatsink as soon as I hit above 1.4-1.45 for my 3950mhz runs it starts to cook it but its still within safe voltage and temps for a 45nm. I used to pile 1.55v-core on my q6600 when everyone says it was to much but its still chuggin along after year and a half, its 65nm though so it can take it like a tank been prime and 3d06 stable at 3951mhz many times. Also I have been running this poor e5200 (2500mhz stock) at 1.6v just so I can prime95 stable it at 4300-4500mhz depending on how cold I take it, its still kicking even after 1.7v 4600mhz 3dmark meat freezer runs for weeks.
 
I just looked at your screen shot and it looks like you need a little more v-core, tweak that GTL reference voltage first before trying more v-core though. Also you may try running a house fan over your CPU and see if cooling it a little better makes a difference if more GTL and v-vore dosnt get you anywhere, ive seen as high as 80's on my 3950mhz q9550 prime runs so yours isnt to bad of temps.
 
I'll give this a shot when I get back home from work. Thanks for the help. With the prior tips though I was at least able to get 3.8 to post and try and load Windows.
 
I've had much better luck with my g35-based motherboard than the G/P45 series. 3.4Ghz at 1.07 vcore with stock voltages all around elsewhere. 3.8Ghz with stock motherboard but vcore at stock (1.175). Anything more than this, it seems to be POWER SUPPLY limited with my 380w Antec (my power supply can't be more than 5.5" deep so there are limits or what I can use) as 3.9 and 4.0 I cannot get 100% stable in any way. Memory is still running under speed and voltages only affect cooling, not stability. Temperature IBT loads like 50-52C at 3.4Ghz and 60-62C at 3.8Ghz so cooling isn't a problem. I run it at 3.4 as it helps keep my system idling at 116w of power consumption and does everything as fast as the hard drive can keep up with - before looking into these numbers, I was up near 190w at IDLE (including the HD4870's lack of stock idle settings).

I suggest not blaming the motherboard first due to all the other possible problem sources. It's possible there's dirty power being output by the power supply, bad capacitors, etc. Lastly, I have a coworker who has had horrible stability with his X38-based motherboard. His older G33-based board has been much better, both made by Asus.
 
SGSeeker said:
I suggest not blaming the motherboard first due to all the other possible problem sources. It's possible there's dirty power being output by the power supply, bad capacitors, etc. Lastly, I have a coworker who has had horrible stability with his X38-based motherboard. His older G33-based board has been much better, both made by Asus.
Click to expand...

Reason I blame the mobo so easily is that this particular model is very much disliked on the DFI forums. The LT series is considered to be much better, while the DK series has serious issues with FSB walls.
 
SilverSliver said:
Reason I blame the mobo so easily is that this particular model is very much disliked on the DFI forums. The LT series is considered to be much better, while the DK series has serious issues with FSB walls.
Click to expand...

My DK P35 board did just fine with Xeon 3350 up to 3.6. Have you played around with GTL Ref? It's essential for C2Q oc'ing.

Other than that you can give your chip a little more volts. But I wouldn't go much over 1.35-1.4 if you want to keep that chip for a while.

It is true that the X48 x X38 DFI boards exhibited lackluster performance, but you should be able to hit at least 450FSB with the right tweaking.

IMO the X48 and X38 UT boards from DFI were way overpriced for what they were.
 
1.38vcore, 1.3vtt, 1.45mch, GTL ref is stock at 67%. No go on 3.8. Made sure speedstep was off. It will post, gets past loading bar in Vista then black screen.

I should clarify that it's the DFI DK X38 boards that get bad press on the DFI forums, not the P35 or any others.
 
SilverSliver said:
1.38vcore, 1.3vtt, 1.45mch, GTL ref is stock at 67%. No go on 3.8. Made sure speedstep was off. It will post, gets past loading bar in Vista then black screen.

I should clarify that it's the DFI DK X38 boards that get bad press on the DFI forums, not the P35 or any others.
Click to expand...

I found that GTL at .63 or .61 lowered temps and helped stability on quads. You really should try that out. As far as 3.8 no go. It could very will be your chip. The C2Qs are comfortable at 3.6 or so. 4.0GHz is quite rare and requires some effort. I did get my x3350 to boot at 4.0GHz for a suicide run, but that's about it. My wisdom and wallet told me to back that sucka down (paid $380 for that mofo!)
 
Not to derail this thread, but thanks for the tips. I was able to hit 3.6Ghz because of the information in this thread.
 
You must log in or register to reply here.
Back
Top