Originally Posted by GanjaSMK
It should be really easy to figure this out. All you need to do is turn off your PC for about 5 minutes, let it cool down. Then when you turn it back on, go directly into the BIOS and check your hardware monitoring section (whatever it may be called). It should list at least two temperatures, if not more. The two should be CPU/MB or CPU/NB. Once you can clearly see the temperatures of those on a cold boot, just boot directly into Windows from there. Then you can check which temperatures are which.
If you're using onboard video, then I highly doubt you're getting your SB reported to you. Most likely you have CPU/MB(or)NB/GPU temperatures in HWMonitor. It's only on most of the really expensive boards that you'll get SB temps, and even then it's never really a concern (unless you're experiencing problems related to the SB).
EDIT: Also, that XBit labs review of the 1055T is a one-time review. It's not like they spend weeks on end to produce the review, maybe a few days tops. Their conclusions of temperatures do show there is a sensor problem, but it doesn't conclude that everyone will see the same results.
You are assuming that I believe what is posted in the BIOS to be accurate. As I said, I have already isolated TMPIN0 as the System Temperature and TMPIN1 as, most likely, the CPU Temperature listed in the BIOS. The latter was harder to isolate as TMPIN1 and TMPIN2 are well matched at steady state. I used the CPU temperature alarm in order to draw my conclusion as TMPIN1 crossed the 60C threshold. TNPIN2 always responds more slowly. Moreover, I believe that the fictitious core temperatures may be calculated based on partial input from TMPIN1.
I have a hard time digesting your comment with regard to the distinction between NB temperature and the integrated GPU temperature. Two locations on this chip are not physically separated by enough difference to yield any noticeable temperature difference, even instantaneously. Your argument would seem to imply that TMPIN1 and TMPIN2 refer to these particular values. I, on the other hand, would argue that one of these must be the CPU temperature. If you can present an argument as to which one is the socket temperature based upon observations, I would greatly appreciate it.
Perhaps, a more mature BIOS revision will distinguish all three temperature values that appear in Everest and in other software. However, for now, I trust Everest more than the BIOS.
I can relate to your post. First of all, why are your ambient temperatures so high? Do you operate your system out of a kitchen, solarium, boiler room or next to a laudry room, run heating units all day or receive continuous exposure to sunlight? North East American temperatures are in the upper teens. Therefore, indoor temperatures should be in the lower 20s.
Easy Tune 6 reports temperatures directly from the BIOS. For example, my system and CPU temperatures idle at 41 and 43 and reached 46 and 65 during the hottest part of the summer. The 65 C was a transient which dropped to 62-63 C as fan RPM increased from 5500 to 7000. This is also the temperature that sets off the BIOS alarm. However, at 65 C, Hardware Thermal Control should throttle the CPU. To the best of my knowledge and in accordance with the FLOPS ratings, it did not. I will verify this again, next summer.
Originally Posted by doritos93
First of all, I want to thank you for taking the time to formulate a specific response to my question. Hopefully we'll be able to find a solution to this problem. All summer, ambient temps have been around 30+ degrees. So the 23.9 - 25 right now is actually good. Honest, these are what temps usually are like around here and my machine is in a normal room about 3-4 feet from a window blowing in cool air. I laughed really hard when I read "boiler room", even if you were serious.
Now here is the confusing part. Everest identifies the latter as the Motherboard temperature (TMPIN1 in HWMonitor). The System temperature is idnetified as the AUX temperature in Everest (TMPIN0 in HWMonitor). It is safe to assume that this is the South Bridge temperature.
However, Everest identifies a third temperature (TMPIN2 in HWMonitor) as the CPU temperature. This temperature also rose to 62-63 C during the hottest part of the summer, but more slowly than TMPIN1. It now idles at 45 C.
Now, we must do some detective work as TMPIN1 and TMPIN2 should reference the CPU and North Bridge diodes in no partcular order. Since my temperatures are almost indistinguishable except during transitions from idle to load and vice versa, let us assume that TMPIN2 @ 57-61 C is your Northbridge temperature. I assume that your screenshot is from a Prime95 or Linpack run. (What readings do you get at idle?) In that case, is there any reason why your Northbridge should be so hot? There is no GPU integrated in the low-power Northbridge and your chipset features two heatsinks connected by a heatpipe, more than sufficient to provide adequate cooling. Moreover, this design enables the chipset to be cooled by the CPU cooler. Finally, 2.72 GHz is not unreasonable for your Northbridge link frequency. The evidence does not support a conclusion that TMPIN2 is your NB diode temperature.
Assuming that your Apogee GTZ / MCP655 is mounted correctly, the evidence would not support a conclusion that TMPIN2 is your CPU temperature, either.
More information is required. Please provide me with the full spectrum of temperature readings that you experience from the moment you boot your computer, at idle during normal use and under stabilty tests. For now, this is all the information that I can provide. I hope that it is useful.
Originally Posted by doritos93
Well looking at Everest right now (idle) CPU = 37, MB = 29. ET indicates the same values but inverted. The AUX temp (which is the labeled as a system temp for me as well) = 43.
BIOS indicates same idle temps as ET, as you said.
I mounted the waterblock 3 times, was careful with the paste too. One thing worth mentioning is that my loop was unused for about 1.5 months because I had to RMA the board. During that time, I started the loop periodically to keep the coolant from getting sludgy. Since then, I've emptied the loop and replaced it with fresh water. Unless I really screwed up the waterblock by leaving still water in there for too long, my cooling solution shouldn't be in question. Although, I am considering slapping on the stock cooler as a basis for comparison.
The board is fresh from RMA. The only thing I notice is that theres a white spill on the NB heatsink but nothing that prevents it from booting.
As an experiment I pointed a small fan at the NB and SB and funny thing is no temperatures dropped at all (or hardly noticeable). On a different machine, pointing a fan right at the NB caused temps to drop dramatically.
I'm a little scatter brain this morning so if I think of anything else I'll reply later.
Thank you again.
In summary, we can establish the following observed temperatures:
Assuming that there are no special modifications to your North Bridge, I believe that the CPU temperature should exceed the NB temperature most of the time. It also appears that TMPIN2 is more volatile than TMPIN1, which is typical of CPU behavior. In my case, TMPIN1 is extemely volatile because of the overclocked IGX integrated on the North Bridge.
You probably should check some of the posts below in order to see what the Apogee GTZ is capable of. Riskitall84 is currently running a 95 W processor in this configuration and his temperatures seem to be much more reasonable than yours with clock speeds well in excess of 4 GHz. Perhaps, he can offer you some advice on how to lower temperatures, though I am not certain that you will be able to achieve his overclock. Your motherboard has had several issues overclocking above base clocks of 275 Mhz, especially with early BIOS versions.
Originally Posted by doritos93
Sorry the link you sent is dead. Like I said, unless the block broke somehow during the 2 months it was sitting doing nothing, there shouldn't be a problem.
I'm entirely stable @ 3.9 GHz so the board is doing pretty well for itself. There are only 3 bios revisions so far so I think you're confusing my board with another, similarly named 790X Gigabyte board.
I still convinced that since the PC speaker doesn't start blaring when TMPIN2 goes over 60 is a dead give-away that the temp sensors are not mapped correctly in HWMon / Everest. I mean, when did these applications start having priority over the readings in the BIOS? (ET=BIOS) I'm starting to think that mobo manufacturers haven't quite worked out all the kinks with these newer Phenoms.
Let me know what you think
This is the information that I have regarding your motherboard's overclocking potential using the F2 BIOS:
If these reviewers are credible and no substantial improvements have been introduced in the F3 revision, a 300 MHz reference clock is out of the question.
As far as the identity of the temperatures, I dare say that I am at a loss. The 61 C load temperature is clearly not a spurious reading. If it is not the CPU temperature, could it be the temperature of the North Bridge or the voltage stabilizer? I believe that these links suggest otherwise:
In this case, the BIOS readings would be mismapped and the alarm would not sound because the sensor that it is referncing is actually aboard the Northbridge, not the CPU. I cannot really say much more. Good luck!
Further to our discussion, it appears that there is no clear consensus regarding the identity of TMPIN1 and 2 on your motherboard. Here are some links:
Some believe TMPIN1 to be the CPU socket temperature; they argue that TMPIN2 is the Northbridge temperature, which is not adequately cooled by aftermarket coolers that blow air parallel to the chipset surface. One owner even attempted to measure the NB temperature directly. The stock fan supposedly does a much better job by venting air normal to the surface of the die. This explanation sounds reasonable, but you mentioned that adding the same fan to the chipset lowered TMPIN2 by less than a degree.
Others suggest that Everest correctly identifies the CPU and Motherboard (Northbridge) temperatures.
Both arguments have merit, but both are also flawed. Based on the description in the first link, I am convinced that our motherboards share the same set of sensors. However, I must note that my TMPIN1 sensor has indicated 65 C without throttling or compromising performance in stressful benchmarks. Perhaps, there is some degree of miscalibration.
I require one further clarification. In March, when you were using the F2 BIOS, how did your configuration differ from the present one. I assume that you were using a 955 or 965 BE at the time. Same cooling? Which modification was responsible for lowering TMPIN2 from 80 to 61 C?
I will provide you with one more reference and advise you that I, personally, am in the Everest camp. Please compare the previous link:
to this one:
If you look carefully, you will notice that TMPIN1 falls and rises abruptly. In fact, it is very well correlated with the temperature variations in the cores. Both of these behaviors are charateristic of readings from an internal diode.
TMPIN2, on the other hand, lags temperature changes in the cores and requires almost a minute to reach steady state at both 0% and 100% processor states. This is the nature of a socket temperature, which is mediated by some thermal resistance.
The team at Xbitlabs did a good job of examining this while performing tests on their ASUS M4A89GTD PRO/USB3 motherboard (there should be no ambiguities with that board).
You may notice another interesting property. At 0% load, the difference between TMPIN2 and the core temps is much greater than at 100% load. There are two reasons. The socket temperature was not given sufficient time to reach steady state prior to state changes from 0% to 100% load. Secondly, thermal conductivity exhibits some temperature dependence in electronic materials. Based upon these observations, I am fairly certain that TMPIN1 is an internal diode temperature measurement from within the Northbridge and that TMPIN2 is the CPU socket temperature.Edited by Headless Fansprings - 9/23/10 at 2:02pm