joeyao
02-19-2003, 09:32 AM
As we image before, Abit Technical support give me reply that KD7 series mobo support reading the temp of on-die diode.
after download the lastest vesion MBM, and select "LM90 remote" in "temp"- "should display board sensor" option , it seem that the on-die temp(named "core temp" in KD7-E bios) can be read in windows enviornment. but after a little test, this core temp is something like a trick:
I use a VCOOL software(like cpuidle)to lessen the cpu load, but the core temp is still high , the same as when entering windows without using VCOOL, so I confirm the "cpu usage" drop down to almost zero when opening VCOOL. It is a pity that the wire of my temp sensor(Tt Smart Fan2) was broken, so I can not prove further that the change of core temp is not synchronous with the change of the temp of the back of CPU.
In my imporession, the temp of back of cpu will drop down markedly immediately when opening VCOOL.
Suppose that the core temp is a real one, its reaction should be more faster than the in-socket sensor and should drop down when the cpu load is lightened. But if the core temp is a combination and calibration (on-die temp & in-socket temp plus calibration), it should be a trick.
It seem that no mobo successfully offer the solution of temp reading now.
after download the lastest vesion MBM, and select "LM90 remote" in "temp"- "should display board sensor" option , it seem that the on-die temp(named "core temp" in KD7-E bios) can be read in windows enviornment. but after a little test, this core temp is something like a trick:
I use a VCOOL software(like cpuidle)to lessen the cpu load, but the core temp is still high , the same as when entering windows without using VCOOL, so I confirm the "cpu usage" drop down to almost zero when opening VCOOL. It is a pity that the wire of my temp sensor(Tt Smart Fan2) was broken, so I can not prove further that the change of core temp is not synchronous with the change of the temp of the back of CPU.
In my imporession, the temp of back of cpu will drop down markedly immediately when opening VCOOL.
Suppose that the core temp is a real one, its reaction should be more faster than the in-socket sensor and should drop down when the cpu load is lightened. But if the core temp is a combination and calibration (on-die temp & in-socket temp plus calibration), it should be a trick.
It seem that no mobo successfully offer the solution of temp reading now.