• 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.

Core i7 w/V8 temp issues

N

Nexion<ChesteR>

n00b
Joined
Nov 26, 2005
Messages
21
Just took the heatsink off after letting it sit overnight, after re-application using less paste. Was running at 49-50C idle.

This looks strange:
imag0150m.jpg

imag0151s.jpg


Re-applied again using the bead method down the length of the heatsink per arctic silver, running at around 42C idle. I checked it with a razor blade and you can see how it is "arced" downward in the middle of the CPU. I thought this was really odd, I had newegg send me a new one and got an RMA for the other. I received the new one in yesterday and it is "domed" in the center pointing upward. I applied AS5 again using the bead method down the length of the CPU. I'm running hotter overall on all cores on this new CPU, whereas the old CPU would run 42-40-38-37, this is running 42-41-42-38. Running Burn-In Test with the case side panel on I max out at 65C on the CPU temp test. Do temps across each core matter? Which CPU should I keep now that I can assume the flatness imperfections are "normal"? Recommendations on what I should do to get temps better? I'm fairly experienced with computers, but this is my first rig I've put a decent amount of money into. I appreciate the guidance.


Specs:
Silverstone Raven RV-01 case
eVGA 3X SLI Motherboard
Intel Core I7 920
eVGA GTX285 1GB
Corsair Dominator DDR3 1600 6GB Ram
Corsair Dominator memory fan
Corsair 1000w Certified PSU
WD 1TB Black HD
Coolermaster V8 heatsink

66201117065607936622707.jpg
 
Last edited:
idle temps dont matter at all. many intel 45nm cpu's have sensors that stick in the downward direction anyway. keep whichever one runs cooler at load. or if you are going to overclock (i see no reason not to) keep the one that overclocks better.
 
Whichever one you keep I would consider a lapp job to flatten it out for better HS contact and check the HS too.
 
Your temps aren't that bad anyways. I have the same mobo/cpu/hsf combo and my temps aren't that much better than yours. Stock idle was 39-40c stock load was about 60c. I have it at 3.4ghz now and idle temps are mid 40s with load temps up to about the 65-70 range depending on ambient. While your temps may not be the best, it's still well within acceptable boundaries. I never did check the mount job on my v8 as the temps seemed fine to me, maybe mine needs a lap too! :(
 
its not worth lapping your heat sink, unless you lap your cpu also. its not worth lapping your cpu because it no question voids your warranty, and devalues it, just for maybe a couple degrees.
 
so what about the cores being overall hotter, can each core be any random temp, or is it typical for them to cascade down from core 1? tryign to decide if i want to swap out the CPU and remove the heatsink (for the 5th time :p)
 
Are you measuring temps with a thermal probe or software because software is inaccurate...
 
WarlordOne said:
Are you measuring temps with a thermal probe or software because software is inaccurate...
Click to expand...
Thermal probes are less accurate than software unless you feel like drilling a channel into your heatspreader and sticking a thermocouple inside.
 
My opinion after seeing 1001 Core i7 screen shots is that Intel might not be setting TJMax consistently across all 4 cores.

When running Prime Small FFTs, there is almost always a 5C difference between core 0 and core 3 with the two center cores reporting somewhere in the middle. I use Small FFTs when testing this because it is the best program for equally loading each core.

This observation is based on RealTemp data which unlike some programs, actually reads the data from all 4 cores and correctly organizes the temperature data regardless of the APIC ID that the bios sets. Not all programs are doing this correctly.

Here's the latest beta of RealTemp:
http://www.fileden.com/files/2008/3/3/1794507/RealTempBeta.zip

Whether you agree with my theory or not isn't important. The sloped looking temperatures on these CPUs is normal. When they are not sloped looking then I get suspicious.

Lapping your CPU might improve your temps by a degree or two but it will void your warranty so usually isn't worth it. Lapping won't change the 5C difference at full load between core 0 and core 3.
These sensors are worlds better than the Core 2 45nm sensors and are very repeatable. Core 0 tends to be the most accurate sensor when using the TJMax that is written into the processor. Core 3 TJMax might be deliberately set 5C higher to better control thermal throttling but that's just a theory.
 
Last edited:
unclewebb said:
My opinion after seeing 1001 Core i7 screen shots is that Intel might not be setting TJMax consistently across all 4 cores.

When running Prime Small FFTs, there is almost always a 5C difference between core 0 and core 3 with the two center cores reporting somewhere in the middle. I use Small FFTs when testing this because it is the best program for equally loading each core.

This observation is based on RealTemp data which unlike some programs, actually reads the data from all 4 cores and correctly organizes the temperature data regardless of the APIC ID that the bios sets. Not all programs are doing this correctly.

Here's the latest beta of RealTemp:
http://www.fileden.com/files/2008/3/3/1794507/RealTempBeta.zip

Whether you agree with my theory or not isn't important. The sloped looking temperatures on these CPUs is normal. When they are not sloped looking then I get suspicious.

Lapping your CPU might improve your temps by a degree or two but it will void your warranty so usually isn't worth it. Lapping won't change the 5C difference at full load between core 0 and core 3.
These sensors are worlds better than the Core 2 45nm sensors and are very repeatable. Core 0 tends to be the most accurate sensor when using the TJMax that is written into the processor. Core 3 TJMax might be deliberately set 5C higher to better control thermal throttling but that's just a theory.
Click to expand...

So the CPU I'm running now is not sloped, I guess I should swap to the original?
 
Post a RealTemp screen shot while running Prime95 Small FFTs. Load temps are more important than idle temps. The two end cores almost always vary by 5C and the two center cores are somewhere in the middle. This can vary depending on how you install your thermal paste and whether the heat spreader is convex or concave. I'm just saying that this is normal. Why it happens doesn't matter. It just happens.

I'd keep what ever CPU overclocks better. That's important to me. Maybe cooler temperatures are more important to you.

It doesn't matter too much because they probably run quite similar.

I tend to use a modified version of the AS5 method. I draw a bead like they recommend and then use a little more and arrange it in a rectangular area directly over top of where the cores lay. I find this gives more consistent temps with less break in or burn in time needed but long term both methods are probably similar for results.
 
Last edited:
unclewebb said:
Post a RealTemp screen shot while running Prime95 Small FFTs. Load temps are more important than idle temps. The two end cores almost always vary by 5C and the two center cores are somewhere in the middle. This can vary depending on how you install your thermal paste and whether the heat spreader is convex or concave. I'm just saying that this is normal. Why it happens doesn't matter. It just happens.

I'd keep what ever CPU overclocks better. That's important to me. Maybe cooler temperatures are more important to you.

It doesn't matter too much because they probably run quite similar.

I tend to use a modified version of the AS5 method. I draw a bead like they recommend and then use a little more and arrange it in a rectangular area directly over top of where the cores lay. I find this gives more consistent temps with less break in or burn in time needed but long term both methods are probably similar for results.
Click to expand...

well I already have the motherboard out unfortunately. I appreciate all the help, but I think right now I need to decide if i want the CPU with the mound in the middle of the spreader, or the concave across the center so the spreader is "arced". Mathematically, I believe the arced one will probably have more coverage area. What are your opinions? If you believe I should put the mobo back in and re-test it before swapping the CPU, I will. No one else has helped this much :p
 
As I said, core 0 tends to be the most accurate and at idle, both of your CPUs were reporting 42C. At full load I'm not sure how these two CPUs compared and to do a fair comparison you really need to control all of the variables that control your CPU temperature. A quick test of each without the thermal paste having time to cure and using a program that doesn't equally load the cores might not be a very fair comparison.

Are both CPUs the exact same? I prefer the one that has stamped on top of it, SLBEJ which is revision D0 to the earlier SLBCH which is revision C0.

If you don't ever plan to lap your CPU, the one with the cores or the center of the heat spreader making better contact with the heatsink makes sense. If it was my CPU, I would probably take the other CPU that is higher at the edges of the heat spreader and then do a quick lapping of it to flatten it out.

It's a coin toss. I'm sure there are advantages and disadvantages to both. If you don't plan to set some new world records with this rig and only want a computer that works then either CPU will be fine.
 
i just repasted this one and lowered the temp by 1*C. 39-37-38-34. I think im just going to stay with this one, i guess i dont see much benefit in swapping to the other for that 1C or so drop on the 3rd core. I appreciate your help!
 
You must log in or register to reply here.
Back
Top