I have an overclocked system whose thermal behavior is baffling. It is an ABIT IS7-E motherboard with a Prescott P4 3.0 GHz and two 512 MB OCZ 2,2,2,5 Platinum Rev. 2 memory modules. The PS is a new OCZ 500. CPU cooling is provided by a Thermaltake Big Typhoon. The system is overclocked to 3.74 GHz. I have overclocked many ABIT IC7 and IS7 motherboard based systems so have a lot of experience with the BIOS settings. The memory divider is set to 5:4 and the memory timings to 2,2,2,5 -- in other words the memory is running within spec. Memtest has run overnight -- nearly a hundred passes -- with out error several times. In ordinary operation the system has never crashed and is apparently completely stable. FanSpeed shows a CPU temperature in the low 40's in idle.
The thermal anomaly shows up when I run Prime95 torture test. FanSpeed shows the CPU temperature increasing to 53 to 55 -- it varies within these bounds -- for a few hours. The only anomaly during this time is that a couple of times the temperature plot dropped in a spike to 0 or thereabouts but resumed at 53 to 55. No errors occur in the Prime95 calculations and it completes quite a few blocks of the FFT iterations. In otherwords for several hours under the Prime95 torture test the load temperature stays steady in the 53 to 55 range -- well below the thermal limit of the Prescott -- and the computations proceed without error. Then suddenly the temperature will shoot up and the audio alarm for overheating will sound. I can still power down the system -- Start, Turn Off etc. so the system is still working. The system is undamaged by the event since I can reboot and resume. That makes me think the temperature excursion may not be real, but whether it is or not, I don't understand what is happening.
Any ideas?
Thanks.
The thermal anomaly shows up when I run Prime95 torture test. FanSpeed shows the CPU temperature increasing to 53 to 55 -- it varies within these bounds -- for a few hours. The only anomaly during this time is that a couple of times the temperature plot dropped in a spike to 0 or thereabouts but resumed at 53 to 55. No errors occur in the Prime95 calculations and it completes quite a few blocks of the FFT iterations. In otherwords for several hours under the Prime95 torture test the load temperature stays steady in the 53 to 55 range -- well below the thermal limit of the Prescott -- and the computations proceed without error. Then suddenly the temperature will shoot up and the audio alarm for overheating will sound. I can still power down the system -- Start, Turn Off etc. so the system is still working. The system is undamaged by the event since I can reboot and resume. That makes me think the temperature excursion may not be real, but whether it is or not, I don't understand what is happening.
Any ideas?
Thanks.
Facebook
Twitter