ALCPU Forums

The Coolest wrote:

3. I can only see a single temperature reading.
    AMD processors based on the Phenom and Phenom II (Athlon II, Sempron II, Turion II, etc.) only have a single thermal sensor.
    Thus Core Temp will only display a single CPU temperature reading. There is no way of getting a per-core reading on these processors.

The above confuses me because CPUID Hardware Monitor Pro provides the following readings to me:


CPUID Hardware Monitor Pro - AMD Phenom II X6 1090T Temperatures and Powers

If CPUID Hardware Monitor Pro has correct readings, as above, then can those readings be added to Core Temp?

The CPU has a single sensor, you can easily see that each core always displays the same value.
I decided that in Core Temp the space could be used for more important information, such as per-core frequency.

The Coolest,

I apologize for the previous image that appears to support an incorrect conclusion. Please view the following image made today (may need to be zoomed for some viewers):


CPUID Hardware Monitor Pro - AMD Phenom II X6 1090T - 2010-08-26; 1.31 pm

Core #1 20.5°C
Core #2 20.8°C
Core #3 21.0°C
Core #4 21.3°C
Core #5 21.0°C
Core #6 21.3°C

Those readings change with changes in my system (CPU) voltages and individual core loads.

Ok, I'll look into this in more detail. Since most programs read temperatures in a linear fashion, one core at a time. By the time the program gets from the first core to the last, the CPU has heat up a little. But as I said, I'll look into this more, because there might be something I'm missing.

Thank you for considering my request and expending some of your efforts to it.

Please understand that I have observed in the CPU Usage History on the Performance tab of the Windows 7 Task Manager that at times the same load appears to be shared by at least two and maybe three cores. I just checked and observed two cores appearing to share one task/thread while another pair appeared to share a second task/thread but the fifth and sixth cores appeared to not be sharing their tasks/threads with any other cores. Alternatively, I am able to assign any individual cores or any combination of cores or all cores to the same process/thread. However, I have not done so for any of the images I have posted here.

To show all my core temperature readings do not follow specific patterns, I am including an image in which they are in a somewhat reversed order to my last previous and two images in which the readings are varied:


..................................2010-08-27 2.40 am....................................................................2010-08-27 3.04 am....................................................................2010-08-27 3.08 am..................................

Thank you for all of your hard work, and thank you, again, for considering my request and giving your efforts to it.

Once again. I don't know the interval at which the readings are taken in HWMonitor, but the thermal sensor gets updates all the time, the minor differences you see between the cores are negligible and could easily be a result of a delay between reads.
There is no point displaying 6 values of the same sensor in Core Temp. In HWMonitor it makes sense to simplify a user's understanding of what they see, and the readings are arranged in a list. In Core Temp, since there is no list, you only have so much room for all the information to display about the processor, and wasting space on duplicate readings is not the most efficient way to go.
All Phenom processors only have ONE SINGLE SENSOR. All of the 6 core temperature readings are reporting the same value exactly, especially if you round them to the closest integer value, which Core Temp does anyway.

I just confirmed it with HWMonitor's developer, you see different temperatures because of the delay between readings taken by each core.

Thank you for your assistance. It may be that you and HWMonitor's developer have the correct information.

As noted in my signature, my motherboard is the ASUS Crosshair IV Formula. My desire is to understand the complexities and full functionality of the system I built, including the CPU, so that I operate it within safe limits and avoid premature failure of the components. In that regard, I reviewed a couple of AMD Phenom CPU manuals. The following information stirred my excitement:

http://support.amd.com/us/kbarticles/Pa ... tures.aspx

Last Updated
1/20/2010
Article Number
CPU-10

Troubleshooting:
If the system is not experiencing any instability, it is possible that the utility used to measure the temperature of the processor is inaccurate. The following troubleshooting suggestions can help determine if the issue is related to the temperate monitoring utility:

1. Determine how the utility obtains the processor’s temperature
Utilities that monitor the processors temperature can obtain the measurement in several different ways: through thermal sensors on the motherboard or digital temperature signals (DTS) on the processor. If a motherboard uses thermal sensors to obtain the processor’s temperature, depending on where the thermal sensor is placed, the temperature reading can be inaccurate.

. . * For best results, use a monitoring utility that extracts the temperature reading from DTS

I also reviewed the document at http://support.amd.com/us/Processor_TechDocs/43375.pdf.

In those studies and in addition to the above DTS information, I discovered there are certain instructions that will dump various information about the CPU into registers. I also installed and reviewed AMD Overdrive, which reveals a lot of information about the CPU, motherboard, memory, etc. and allows tweaking, such as the multipliers of individual cores. So, with the ability to tweak individual cores and track individual core speeds, it seems only logical that individual temperatures should be readable. However, I was not able to find such information and it appears that AMD engineers have not considered that with the ability to set individual core multipliers different from each other the heat buildup within each core also will be different. Since I do not know the physical location and proximity of all the cores, I am unable to know whether that is true or whether they directly influence each other and produce an average CPU temperature. It does seem that is indicated by the above statement of singularity: "extracts the temperature reading from DTS." Nonetheless, I am left confused by the above statement of multiplicity: "digital temperature signals (DTS) on the processor." I will leave that to you technicians because I must get back to completing my bachelor degree.

Again, thank you for your time and all of your effort on my behalf.

Best wishes,
James