5 Replies Latest reply on Feb 13, 2011 3:17 PM by Doc_SilverCreek

    Processor Thermal Margin Issue

    Amtrak288

      My DX58SO motherboard which is running the BIOS it came with (upgraded it and had nothing but problems, only to downgrade it back to what it came with), shows the Processor Thermal Margin as 0 degrees C at a cold start-up with all other temps listed as normal.   The Intel Desktop Utilities program listed the margin as 128 degrees C and the Intel Desktop Control Center will not display a CPU temp at all, only says 28 with a Delta Symbol then the letter "T" after it and the indicator as at the coldest point.  Is there something wrong with this motherboard?  This is the second DX58SO that I've owned, the first one was sent back to Intel by warranty for unrelated problems so I have another board on the shelf ready to be used, but I really don't want to have to go through the whole process of switching out another motherboard, any ideas here???

        • 1. Re: Processor Thermal Margin Issue
          Doc_SilverCreek

          Hopefully on of the desktop guru's can give you a dinfinity answer.

           

          In the server space, until you update the code stack, (most noticable the FRUSDR) the full therimal sensor range data does not get loaded.

          Basicly the system has no idea what processor you have installed, so it can't corerectly flag the therimals.

           

          I don't know if the desktops prefore the same way or not.

          • 2. Re: Processor Thermal Margin Issue
            Amtrak288

            If this is the case, how would one update the "code stack", and also, what exactly is the code stack?

            • 3. Re: Processor Thermal Margin Issue
              Doc_SilverCreek

              I just did a quick scan of the DX58SO TPS and did not find any reference to SDR or adjustable thresholds, so this may just be a server feature.

               

              The code stack for a server is

              BIOS code

              BMC code

              ME code

              FRUSDR code

               

              Any of which can be installed individually but are normally all installed together as part of a OFU (one flash update)

               

              Therimal margin is the amount remaining before the CPU throttles or shuts downs so you are right to be very concerned with a reported 0 margin.

              Either your CPU is overheated (CPU heat sink not clipped into mother board?) or the temp sensor does not appear to be function (bent CPU pin?, mother board circuit failure? BIOS code not configured to support?)

               

              My recommendation would be to configure a minimum system (nothing connected that is not required to boot to BIOS set-up)

               

              Check that the heat sink is fully clipped into mother board

              Remove and check the CPU pins in the socket and make sure nothing is on the CPU contacts. (use extreame care! the socket pins will bend if you touch them at all)

              Check the components around all mounting holes for damage. (slipped screw drivers are the cause of many board failures)

              Update the BIOS, even if it is to the same version you already have on the mother board.

              Clear cmos / restore defaults (F9)

              • 4. Re: Processor Thermal Margin Issue
                Amtrak288

                The heat sink is fully connected to the motherboard and here's how I know, when I put this motherboard in, I didn't have it connected and the machine went into thermal shutdown (didn't push down on it hard enough), that only happened once last May when I installed the board.  That never happened again and there are no other events logged in the BIOS's event log.  Also, when I run a 3rd party application called CPUID Hardware Monitor, it DOES show the CPU Temp as 39 degrees C, and lists all 4 of the core's temps ranging from 39 to 43 degrees C.  It lists the CPU's fan RPM at about 1829.  No system temp is more than 40 degrees C.  Also, when I powered up the machine this morning after having been kept in a 68 degree room for 7 hours powered off, I powered it up and it read the Processor Thermal Margin as 0 degrees C with all other listed temps in the high 20's.  This machine has a total of 7 cooling fans installed in it and has not shown any performance issues outside of Windows hanging after a bunch of WHEA-Logger errors were reported (that's not related to this).  I'm really afraid to disassemble this thing again, I've already done this once as this is the second motherboard I've used.  How do you reset the CMOS on this motherboard anyway, can't find info about it.

                • 5. Re: Processor Thermal Margin Issue
                  Doc_SilverCreek

                  http://downloadmirror.intel.com/18128/eng/DX58SO_TechProdSpec.pdf

                  Table 23. BIOS Setup Configuration Jumper Settings

                   

                  Clearing CMOS will reset all BIOS defaults so make sure you have looked first to see any that you have changed.

                  (ie, SATA ports  - if your using AHCI, and the default is legicy, the HDD OS will not be reconized)

                   

                   

                  I tend to believe CPUID

                  (I like that tool, consider their donation site http://www.cpuid.com/softwares/cpu-z.html i would hate to see them go away or start charging for updates)

                   

                  Are you reading the margin with a tool inside the OS or is it a bios set-up screen?

                  It may be that the monitoring software is not loading the right IPMI drivers.

                  Windows and Intel's are not always compatable. (CPUz loads their own version (smart of them)