HHWiNFO64 is Excellent!

[parsec]

Hi Martin, I'd like to tell you that I've been using HWiNFO64 for a while now, and it is just fantastic! First, I use it on my ASUS SaberTooth P67 board with i7-2600k CPU. IMO, all the Sensor readings are perfectly accurate, and thanks for including all the "Thermal Radar" sensor readings (all eleven of them) in the display. The System Summary display is working great now, IMO earlier versions of HWiNFO32 were not showing the correct clock speeds of the CPU cores, which never seemed to show the CPU cores in the C1E state, at 1.6GHz. That has been corrected, and it seems just perfect now. The core speeds really change quickly, don't they, and I imagine it was a challenge keeping up with them. I have the Turbo multipliers at 45 on my CPU, and the display shows that perfectly. I can't see anything not quite right whatsoever, very excellent work, just great and thank you for providing this great program! I also use HWiNFO64 on my SaberTooth X58 PC, with an i7-930 CPU, and it seems perfect on that system too.

I have a few questions and comments, if I may. I've noticed in the Sensor Status display, that there are two or three general CPU temperature values shown, one or two in the "Intel CPU" section (CPU Package, and also CPU IA Cores), the other in the "Nuvoton NCT6776F" section (labelled CPU). The CPU temperature value in the Intel CPU section do not match that shown in the Nuvoton NCT6776F section, the latter always reading at least 10 degrees C lower than the former.

Please note that I am not saying that this is a bug in HWiNFO64, I do not believe it is a problem at all. For whatever reason, the Nuvoton chip provides, IMO, an incorrect CPU temperature. I have actually seen it regularly report a CPU temperature that is less than my room/ambient temperature. While I use a high performance CPU cooler, I know it is impossible for the CPU temperature to be below the ambient temperature.

I am also mentioning this because recently an EFI/BIOS update from ASUS for the SaberTooth P67 caused the program provided by ASUS for temperature monitoring (the Monitor function of AI Suite II) to use the CPU temperature reading from the Nuvoton chip, where previously it used the temperature reading from (I assume) the Intel CPU. The "new" CPU temperature reading in Monitor matches exactly what HWiNFO64 shows for CPU in the Nuvoton section, where it previously matched what HWiNFO64 showed in the Intel CPU section. I consider this situation to be an error by ASUS, and not a trivial one because the CPU temperature value is used by their PWM fan control software to set the CPU cooler's fan speed. The result of this is an under-cooled CPU, and I have been warning users of this in the ASUS forum for this board.

I give credit to you Martin and HWiNFO64, for allowing me to verify and understand what is happening due to this unfortunate situation with the ASUS EFI update. The correct and incorrect CPU temperature readings in Monitor, when used with the previous and latest version of the EFI, match exactly what I see in HWiNFO64, and I was able to identify IMO, that the change was due to using the Nuvoton CPU temperature reading in the EFI update. Thank you so much for all this!

In the Nuvoton section of HWiNFO64, there are four temperature readings I don't understand. They are labeled "Auxiliary", and "SYSTIN1", 2, and 3. I can't match them with any of the other temperature readings in Sensor Status, any idea what they are supposed to be from? I don't consider this to be a bug or problem, just wondering about them.

Again, thanks so much for this program, I use it exclusively now, and is my first choice in hardware monitoring programs.

 

[Martin]

Thank you, it's nice to hear that my tool(s) helped someone.
The temperature values in the "Intel CPU" section come directly from the CPU internal sensors located in the particular section (i.e. the CPU Cores section, uncore or GPU part of the CPU package). These temperatures are usually the highest ones.
I'm not exactly sure what CPU temperature is measured by the Nuvoton chip, I would need to see the mainboard schematics how the chip is connected (and this is not possible). Usually the CPU temperature reported by a sensor chip like Nuvoton is measured from a diode that is placed close to the CPU (i.e. beneath it). So that explains why the internal CPU temperatures are always higher than those measured by a diode close to the CPU. However, it doesn't explain why the Nuvoton CPU temperature is lower than room temperature..
The other Nuvoton temperatures are values reported from the rest of the sensor inputs. I'm not sure whether they are valid at all or where they belong to (depends on how the chip is connected, but this is proprietary and not disclosed by ASUS).

 

[parsec]

Hi Martin, You are welcome. I am aware of the realities of "temperature" readings on Intel CPUs, as I have read some of the technical documents for the various families of CPUs. I also read your post in this forum regarding that, which has real insights into the realities of dealing with deriving temperatures from the information that is available. As you know, a displayed temperature in any program is calculation, using TjMax, or similar figure, and the value in a register indicating the number of increments below that maximum temperature.

Regarding my question and your reply about the Nuvoton CPU temperature, would that work in a similar manner? Then again, I do not know how, for example, a mother board temperature is obtained from it's sensor, which may be quite different.

The ASUS SaberTooth P67 has eleven "sensors" located in various places on the board itself. While they all have names given by ASUS, I know they are not necessarily measuring the named part, chip, etc. A graphical display included with this board shows (hopefully accurately) the location of these sensors. While the values displayed may not be accurate or truly indicate their namesake components temperature, they are at least useful by indicating relative changes in temps in the area of the sensor. For example, the "PCIE-2 sensor is located between the two PCI-E slots where one or more video cards are installed. Users of this board have reported high temperature readings from that sensor when using high end video cards or dual cards in SLI mode, with less than adequate PC case ventilation. IMO this board has pioneered presenting temperatures in regions of a mother board we have never known before.

I have done some checking of surface temperatures of components using an infrared thermometer, and comparing those to the temp reading displayed for the component or area shows they are in decent agreement. IMO these reading are useful and not simply a gimmick.

Why the Nuvoton chip displays at times (not all the time) a temperature lower than ambient can be explained to some degree in my case. I have a well ventilated case and large CPU cooler, and with all the CPU power savings options enabled and the PC at idle, the "Intel CPU" temps are at best a few degrees above ambient. Myself and other owners of this board have determined pretty well that the Nuvoton CPU temp reading is usually 10C less than the Intel CPU reading, and can be over 20C less than the Intel CPU reading. If my ambient is 75F, and the Intel CPU temp is 90F, I can see a Nuvoton reading of 66F. That is the Min temp shown by HWiNFO64 for the Nuvoton CPU temp as I type this. The displayed value may be purposely to low by fudging the numbers, who knows?