• 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.

Q9550 55C load at default?

E

ebilcake

Limp Gawd
Joined
Jan 3, 2008
Messages
168
Just installed this on my P5K Premium and this is disappointing to say the least, this is also using a Ultra Extreme with 2x120mm fans, my Q6600 didn't even get that hot at 3.4GHz. :(
 
Tried reseating the heatsink, still 55+ at default clocks, oh well, just hope the temps don't shoot up to much when overclocking this thing
 
I don't know what to say... You say you had a Q6600 using the same cooler, and when you OC'd to 3.4Ghz, you didn't get temps as high as you're getting on the Q9550 at 2.83Ghz?

In any case, 55*C at load isn't TOO bad. But that depends on a couple factors:

1) The room temperature
2) The quality of airflow in the case
3) Properly seating HSF with the correct amount of thermal compound

I get ~60*C load temps in Prime95 Small FFT tests with my Q9550 @ 3.77Ghz, but my room temp is around 68*F.
 
Well I tell a lie, the Q6600 did reach around 65C at full load at 3.4GHz, but it was allot lower at 3GHz at stock volts, certainly less then 55C.

So far i'm running the Q9550@3.6GHz it's at 65C with 1.325, but im not very confident it's stable. Air flow is good, 3x 120mm case fans, all around 1200rpm and two 120mm fans on the Ultra, same speed, room temp is pretty cool, i'm in the UK and it is late october afterall.

I'm using the right amount of AS5 as per the instructions on their website and i've seated it twice with no change in temps.
 
The temperature sensors on 45nm chips can be garbage. I wouldn't put too much faith in them. As long as your overclock is stable and you're not thermal throttling then you can safely ignore them.
 
unclewebb said:
The temperature sensors on 45nm chips can be garbage. I wouldn't put too much faith in them. As long as your overclock is stable and you're not thermal throttling then you can safely ignore them.
Click to expand...

The heatsink is still cold to the touch but i'm not sure im comfortable ignoring the temps, whats the max rated voltage for these?
 
if the heatsink is cold to the touch then it's not properly pulling heat off of the cpu, something isnt seated right, check to see if it's properly mounted to the motherboard
 
ubergeek14 said:
if the heatsink is cold to the touch then it's not properly pulling heat off of the cpu, something isnt seated right, check to see if it's properly mounted to the motherboard
Click to expand...

It is mounted properly, theres no other way to mount it, unless I turn it round and point the fans upwards but that would be silly, it's been priming for an hour now and it is feeling a little warm on the top of the heatsink.
 
ebilcake said:
3.8GHz at 1.4V now, 70C!! :(
Click to expand...

1.4v on air is where the yellow line is for many. I'd put the red line at 1.425-1.45 depending on what cooler you had. If those temps are accurate I would be hesistant to push much higher. Having said that if the system is stable and passes everything you throw at it you're kinda good to go.
 
GonzoP said:
1.4v on air is where the yellow line is for many. I'd put the red line at 1.425-1.45 depending on what cooler you had. If those temps are accurate I would be hesistant to push much higher. Having said that if the system is stable and passes everything you throw at it you're kinda good to go.
Click to expand...

It failed after about 20min, needs a little more more voltage, i'll settle for 3.6GHz for now and probaby just save 3.8GHz for benching, not bad but not great either.
 
Yeah 3.6 isn't too bad, I'm sure you wanted a bit more out of the Q9550. Perhaps you just got an average OCer
 
Turns out it's the motherboard having issues above 3.6GHz, doesnt seem to like the memory above 425, memory is rated for 533.

Currently using a 333 strap, worth changing it?
 
Those tempts don't seem TOO bad... what kind of program are you using for measuring the temp? RealTemp is a good one for seeing what kind of differences there are between the 4 sensors.
 
I just realized you said you were pumping 1.4v+ into your Q9550. Be careful with the voltage, that is well over the max safe limit of 1.3625v.
 
ebilcake said:
Is that because yours an an E0 or because mine just sucks?
Click to expand...

i dont think so that e0 stepping makes difference +-10°C.. try to reinstall your cpu cooler it can help :)


and high temperature is the "handicap" of 12mb cache.. > hot hot hot
 
tmaven said:
i dont think so that e0 stepping makes difference +-10°C.. try to reinstall your cpu cooler it can help :)


and high temperature is the "handicap" of 12mb cache.. > hot hot hot
Click to expand...

I've already reinstalled it twice, I can't do anything differently, but it doesnt really matter as the motherboard won't allow for anymore anyway, everytime I increase past 425 it has a mini fit and crashes. :eek:
 
thats wierd you have better cooler but higher temperature.. if i be you i will try normal stock cooler.. if it will overhead itself.. return it to shop
 
I had mine going from 50-90C with heatsink improperly seated. Put it back on correctly, it goes from 33C to 41C at load.
 
How exactly can you mount a 120-Ultra on incorrectly? There is only one way to install them and I know I used right amount of AS5.

Either the CPU is just a poor clocker or the heatsink is defective in someway.
 
ebilcake said:
How exactly can you mount a 120-Ultra on incorrectly? There is only one way to install them and I know I used right amount of AS5.

Either the CPU is just a poor clocker or the heatsink is defective in someway.
Click to expand...

One or more of the screws might not be tight enough... or even if they are as tight as they can be, the contact might not be perfect, and a washer mod might be necessary to improve temps.
 
I don't know about that. Its kinda hard to botch the installation. Uneven, imperfect contact could occur from one screw being too loose or too tight though
 
jimbo6846 said:
One or more of the screws might not be tight enough... or even if they are as tight as they can be, the contact might not be perfect, and a washer mod might be necessary to improve temps.
Click to expand...

All screws are as tight as they will go, they can't get any tighter and none are loose.

washer mod?
 
i have a Q9550 in my possession right now , thats doing the same thing - really high idle / load temps

tried with the current setups:

Stock Cooler (Ran 60ish Idle) (Ha didnt even wanna see prime 95)
Artic Freezer 7 PRO (54-56 Idle) - (Didnt even let it run more than 2 mins of prime95)
Xigmatec Cooler (49 Idle) - (82C Load)
Thermaltake Watercooling setup (44C Idle) - (76C Load)


newegg is taking it back , as a defective processor , its also a C1 Stepper
 
I'm having temp issues with my Q9550 C1 also. Except in my case, its just one core that is a lot hotter than the rest. I've reseated the cooler 3 times now, with no change. Sucks because I'm limited by that one crappy core, but I guess there isnt anything I can do about that....
 
You guys should really be checking to see if you have an option to change the GTL and VTT voltages in your MOBO bios settings. I had no idea it would make such a difference.

It didn't matter as much on the Q6600, but on my x3350 (workstation/server variant of Q9450) it was the difference of almost 10C!!! :eek:

Currently my x3350 is at 3.6GHz as a stable everyday OC. My system idles at 30-40 degrees and loads at 60-70C depending on my room temps.

But before I changed my GTL and CPU VTT volts I was getting 70-80C load!!!

My MOBO is a Asus P5E-VM HDMI.

I'm not near my machine now to tell you guys what my exact settings are, but playing with the GTL and VTT voltages also helped me get less of a difference between the cores as well. I get better priming stability too with the tweaks.

I'm sure many of you know. The LGA775 quads generally have Core 0 a few degrees higher than the rest. I was getting a 10C difference between the other three cores and the first one. I now only have a 3-5C difference across the cores.

My CPU volts are set to 1.42 in bios (in windows and bios hardware monitors it shows 1.32 and less under load) while my VTT is set to 1.4V. I believe my GTL for the CPU is set to .61 in the bios.

Also don't jack up the North Bridge Voltages too high either. Apparently there is some kind of relationship between temperatures of the North Bridge Voltages and the Front Side Bus and CPU temps. I can't confirm this personally, so I took it as a rumor.
 
Flexion said:
You guys should really be checking to see if you have an option to change the GTL and VTT voltages in your MOBO bios settings. I had no idea it would make such a difference.

It didn't matter as much on the Q6600, but on my x3350 (workstation/server variant of Q9450) it was the difference of almost 10C!!! :eek:

Currently my x3350 is at 3.6GHz as a stable everyday OC. My system idles at 30-40 degrees and loads at 60-70C depending on my room temps.

But before I changed my GTL and CPU VTT volts I was getting 70-80C load!!!

My MOBO is a Asus P5E-VM HDMI.

I'm not near my machine now to tell you guys what my exact settings are, but playing with the GTL and VTT voltages also helped me get less of a difference between the cores as well. I get better priming stability too with the tweaks.

I'm sure many of you know. The LGA775 quads generally have Core 0 a few degrees higher than the rest. I was getting a 10C difference between the other three cores and the first one. I now only have a 3-5C difference across the cores.

My CPU volts are set to 1.42 in bios (in windows and bios hardware monitors it shows 1.32 and less under load) while my VTT is set to 1.4V. I believe my GTL for the CPU is set to .61 in the bios.

Also don't jack up the North Bridge Voltages too high either. Apparently there is some kind of relationship between temperatures of the North Bridge Voltages and the Front Side Bus and CPU temps. I can't confirm this personally, so I took it as a rumor.
Click to expand...

Thanks , i'll see if changing them helps, can you recommend a safe setting yo try first or should I just set the lowest?
 
ebilcake said:
Thanks , i'll see if changing them helps, can you recommend a safe setting yo try first or should I just set the lowest?
Click to expand...

As I'm sure you'll discover by experimenting. I wouldn't set the GTL ref values to the lowest. You may be able to boot into Windows, but the system will be flaky.

In my bios, the options are .63x, .61x, .59x and so on. Before .63x it was auto.

I have my GTL Ref set at .61and my VTT is slightly lower than my Vcore for the CPU. You can kill your CPU by setting the vCore or VTT too high.

In reading the SFF forum, there is a thread about the P5E-VM HDMI motherboard. There were a couple of OC'ers able to hit 4.0GHz with Q9450s with this MOBO. I noticed that the successful OC'ers had their Vcores and VTTs at the same voltages or Vcore slightly higher than VTT.

Using the same settings I was able to get the CPU running at 3.8 with some stability and post at 4.0GHz. Whereas if I set everything to default besides the typical Vcore and Northbridge voltage tweaks, I would end up with 3.4-3.5 max OC.

More techical info can be found here:

http://www.edgeofstability.com/index.html

I hope you have a good understanding of electrical engineering.
 
Flexion said:
You guys should really be checking to see if you have an option to change the GTL and VTT voltages in your MOBO bios settings. I had no idea it would make such a difference.

It didn't matter as much on the Q6600, but on my x3350 (workstation/server variant of Q9450) it was the difference of almost 10C!!! :eek:

Currently my x3350 is at 3.6GHz as a stable everyday OC. My system idles at 30-40 degrees and loads at 60-70C depending on my room temps.

But before I changed my GTL and CPU VTT volts I was getting 70-80C load!!!

My MOBO is a Asus P5E-VM HDMI.

I'm not near my machine now to tell you guys what my exact settings are, but playing with the GTL and VTT voltages also helped me get less of a difference between the cores as well. I get better priming stability too with the tweaks.

I'm sure many of you know. The LGA775 quads generally have Core 0 a few degrees higher than the rest. I was getting a 10C difference between the other three cores and the first one. I now only have a 3-5C difference across the cores.

My CPU volts are set to 1.42 in bios (in windows and bios hardware monitors it shows 1.32 and less under load) while my VTT is set to 1.4V. I believe my GTL for the CPU is set to .61 in the bios.

Also don't jack up the North Bridge Voltages too high either. Apparently there is some kind of relationship between temperatures of the North Bridge Voltages and the Front Side Bus and CPU temps. I can't confirm this personally, so I took it as a rumor.
Click to expand...

Wow. Your experience sounds a lot like mine! And as I also have an Asus mobo, I too have those settings. I will definitely give that a try. Thanks!
 
Ok, I don't have an option for VTT, I do have

NB & SB Voltage
NB & CPU GTL Voltage
FSB Termination Voltage

Little help?
 
ebilcake said:
Ok, I don't have an option for VTT, I do have

NB & SB Voltage
NB & CPU GTL Voltage
FSB Termination Voltage

Little help?
Click to expand...

What kind of motherboard do you have?

For NB and CPU GTL Voltage, what can you set? What are the options? Try seeing if different combinations of GTL voltages have an effect on temps for you.


FSB termination voltage I believe should be set to a similar value to your Northbridge voltage.
 
Flexion said:
What kind of motherboard do you have?

For NB and CPU GTL Voltage, what can you set? What are the options? Try seeing if different combinations of GTL voltages have an effect on temps for you.


FSB termination voltage I believe should be set to a similar value to your Northbridge voltage.
Click to expand...

Just found out. Someone on Tom's Hardware says FSB termination voltage is the same as VTT. But I find that peculiar, as my motherboard has separate settings for these. Figures Freakin' Tom's Hardware readers.

http://www.tomshardware.com/forum/248406-29-q9450-voltage
 
btw guys which program are you using for measuring temperature?

coretemp, everest, occt, realtemp.. i have still 47@3 and 50@4 core... it sux
 
Flexion said:
What kind of motherboard do you have?

For NB and CPU GTL Voltage, what can you set? What are the options? Try seeing if different combinations of GTL voltages have an effect on temps for you.


FSB termination voltage I believe should be set to a similar value to your Northbridge voltage.
Click to expand...

That was in the first post, it's the P5K Premium or P35, i'll check but I think they were 0.61, 0.67 etc..
 
ebilcake said:
That was in the first post, it's the P5K Premium or P35, i'll check but I think they were 0.61, 0.67 etc..
Click to expand...

Sorry about that. (Mobo question.)

Well, if you have values like .67, .63, .61. I would try the .63 and .61 settings. I believe on my Asus mobo Auto defaults to .67, so I only show .63 and .61 in my bios settings.

My DFI Mobo however shows .67, .63, .61.

Either way I got my temps down by using .61 as my GTL Ref setting and 1.4 on the VTT.

As for monitoring programs. I get my temps from Realtemp, Asus Monitor Probe and the Bios Hardware Monitor.

Apparently Realtemp should be the program to use on Intel 45nm procs, as the internal probes are reporting 10c higher than they should be.
 
You must log in or register to reply here.
Back
Top