45nm Temp Monitoring

I just put together my first 45nm system (E8500) and I'm curious about which software to use. I've tried Real-Temp and Core-Temp so far.

Real temp shows that I'm currently running at 31/32C but when I use Core Temp it's showing much higher temps at around 40C.

In Bios at idle I'm at 38.5C...

So I'm asking what is correct... also this is my board ASUS P5Q-E... perhaps it's having trouble recognizing the temps of a 45nm chip?

Thanks-

Real temp seems to work the best for my Asus...

In the BIOS...your CPU will be running 100% so your temps will always be higher in the BIOS then at idle in windows...

Also try Hardware Monitor...I like that also for the simple fact it gives reading for both temps and voltages...

I use both Real Temp and Hardware Monitor. HWMonitor shows 10c higher than Real Temp and bios. Real Temp shows closer to bios temps.

Thanks. So are you both saying that for 45nm chips, because I've read that the sensors are different and motherboards have trouble picking up on them.

The 45nm CPUs are believed to use a TJmax of 95c, which would mean Realtemp is correct.

core temp just relesed a new version for that reason

for me it's realtemp & speedfan

adjust tjmaxoffset in coretemp and then temps should similar to realtemp.

meionm said:

adjust tjmaxoffset in coretemp and then temps should similar to realtemp.

Click to expand...

I am curious on this as well. What should my TJMax be for my new 8400? To the 95c as posted by WorshipMe? The standard TJMax was 105 in CoreTemp if I recall correctly.

WonderingSoul said:

I am curious on this as well. What should my TJMax be for my new 8400? To the 95c as posted by WorshipMe? The standard TJMax was 105 in CoreTemp if I recall correctly.

Click to expand...

I go with 95c as readings for idle temps are close to bios readings and more believable. Doubtful that cpu is 42c at idle regardless if it is overclocked or not.

yes 95c
http://www.xtremesystems.org/forums/showthread.php?t=179044

That's weird

Core temp reads:

48-49-47-47

Real temp reads:

34-34-41-41

Hardware monitor reads:

Same as Core temp:

Since this is a fully 24/7 loaded 65nm Q6600 @3GHz under a TR SI-128 heat pipe cooler, I'm going with Core Temp.

I've been reluctant to buy a 45nm quad because of all the stuck sensor issues, etc.

That's weird

Core temp reads:

48-49-47-47

Real temp reads:

34-34-41-41

Hardware monitor reads:

Same as Core temp:

Since this is a fully 24/7 loaded Q6600 @3GHz under a TR SI-128 heat pipe cooler, I'm going with Core Temp.

Click to expand...

That's because you have a 65nm CPU. The 65nms use a 105c Tjmax to calculate core temps. 45nms use 95c.

I use core temp, the extra 10C is safety for me.

meionm said:

I go with 95c as readings for idle temps are close to bios readings and more believable. Doubtful that cpu is 42c at idle regardless if it is overclocked or not.

Click to expand...

Using 95 as tjmax is making my temps LOWER. I don't think this is correct because my temps were already remarkably low to begin with (low 30s on idle in about 23.8 ambient).

E8400 with Tuniq Tower
P35-DS3L
P182 case with top fan blowing directly into the Tuniq

There is a HUGE thread on XS from the creator of Real Temp that explains why his temps read lower than others. If you'd like to read it, it's in the Intel section. As pointed out here and other threads here, it has to do with the TjMax setting. I use Real Temp as well.

Axis

WorshipMe said:

That's because you have a 65nm CPU. The 65nms use a 105c Tjmax to calculate core temps. 45nms use 95c.

Click to expand...

Doesn't the B3 Q6600 have the 105°C Tjmax while my G0 has 95°C?

Maybe Real Temp is assuming 105°C for my processor at stock configuration.

Edit: That was it. Tjmax was set 85-85-95-95

Now it's reading a much more reasonable
45-45-44-44

2nd Edit: Since Intel hasn't seen fit to disclose the actual Tj max of their desktop processors, it's a guessing game and most apps guessed 100°C for 65nm processors.

Sooooo... I've decided to go with Real Temp as at least he did some more extensive testing.

http://www.techpowerup.com/realtemp/

Audioaficionado said:

Doesn't the B3 Q6600 have the 105°C Tjmax while my G0 has 95°C?

Maybe Real Temp is assuming 105°C for my processor at stock configuration.

Edit: That was it. Tjmax was set 85-85-95-95

Now it's reading a much more reasonable
45-45-44-44

2nd Edit: Since Intel hasn't seen fit to disclose the actual Tj max of their desktop processors, it's a guessing game and most apps guessed 100°C for 65nm processors.

Sooooo... I've decided to go with Real Temp as at least he did some more extensive testing.

Click to expand...

2 cores had a lower tjmax? that could explain as to why 2 cores are always hotter than the other 2..

-Blue- said:

2 cores had a lower tjmax? that could explain as to why 2 cores are always hotter than the other 2..

Click to expand...

No, both cores within a quad have the same TjMax.

Axis

I can't understand why people are still debating what TjMax is for the E8400 and other 45nm processors.

I find it pretty much impossible to argue with this test:
http://www.xtremesystems.org/forums/showpost.php?p=2883315&postcount=573

If you don't believe my biased testing then you can always check out some independent testing by rge over at XS:
http://www.xtremesystems.org/forums/showpost.php?p=3085792&postcount=1525

As for the Q6600, it came in two varieties. The original one was the B3 stepping which consisted of two E6600 Dual Core CPUs strapped together. Guess what? The E6600 B2 has a TjMax=85C so why do people keep thinking that the B3 was different than that?

Anyone that remembers the early days of the Q6600 knows that with the Intel cooler, it was possible to get up to the thermal throttling point when stress testing. What was Intel's solution? Release a Q6600 - G0 and increase the TjMax temperature to 95C to decrease the chance of thermal throttling.

I spent a pile of time testing 65nm and 45nm, Dual and Quad Core processors. Before changing TjMax to some number that the competition picked out of a hat, how about reading the RealTemp docs and at least spend some time learning about what the problem with these sensors is.

http://www.techpowerup.com/realtemp/docs.php

I can't think of any processors that RealTemp is using the wrong TjMax. I can't say the same thing about the competition.

that could explain as to why 2 cores are always hotter than the other 2

Click to expand...

This frequently happens in a Q6600 and has nothing to do with TjMax. There is nothing unusual about having core0/core1 running at a hotter temperature than core2/core3. I think it has more to do with how the IHS was attached to the cores. When running fanless I've seen a temp difference of 10C between the two sets of cores but with a modest overclock and a Tuniq on high, the difference is only about 2C or 3C. Same processor, same sensors, same thermal paste. This goes under the title of s-h-!-t happens.



Don't judge whether TjMax is correct by your idle temperatures. These sensors, out of the box, aren't capable of accurately reporting idle temperatures so that comparison is meaningless. Bios temps also have nothing to do with Windows idle temps. The CPU load is completely different and in Windows the processor is executing wait instructions which don't create the same heat as when you are in the bios.

Hey Uncle, nice to see you stray away from the brood over at XS.

Axis