OK, one last update, I think I have this right. Still using the w83627thf-* config, but there's no -12V supply and there's no battery since this is a desktop machine, so I commented out label in4 and label in8 and replaced them with "ignore" ... I set the vrm to 2.4 since this is an AMD64 and the single datasheet for AMD64, 64FX, and Opteron all use this same spec. http://www.amd.com/us-en/Processors/TechnicalResources/0,,30_182_739_7203,00.html document #30430 "Amd Athlon64 Processor Power and Thermal Data Sheet" http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/30430.pdf It turns out Vcore was correct all along; I thought it was wrong before because the chip is spec'd for 1.4V/1.0V and I was seeing ~1.7V here. But in fact, I had left Vcore on [AUTO] in the BIOS and it chose 1.7V! I fixed the BIOS setting, so the chip is running cooler now. Asus PC Probe didn't report the 3rd temperature sensor, so I really had no idea what it was, and it wasn't in the BIOS screens either. But in the Asus A8V manual there's a screenshot showing this labeled as Power temperature. Also, SiSoft Sandra reported it as Power / Aux so I'm now using these labels: label temp1 "M/B Temp" label temp2 "CPU Temp" label temp3 "Pwr Temp" I believe it's correct. The M/B temp changes very little, +/- one degree over many hours. With a room at 18C ambient, M/B seems to stay around 27C. At the same time, CPU idle, the CPU temp reads 31C, and the Pwr Temp reads 18C. When I run something CPU-intensive (like super_pi) the Pwr Temp rises a couple degrees immediately, and the CPU Temp ramps up slowly. The PWM controller for Fan2 works fine to control the CPU fan. I haven't tested Fan1 or Fan3 because they're plugged into my Antec power supply and that's controlling them. set fan1_div 8 set fan1_min 1000 set fan2_div 4 set fan2_min 1000 label fan1 "Case Fan" label fan2 "CPU Fan" label fan3 "PSU Fan" I had to set fan1_div to get any reading at all, and fan2_div to get something close to reality. fan3 seemed to come up with correct defaults already. I hope this helps any other Asus A8V owners out there... Jean Delvare wrote: > Hi Howard, > > >>I think the CPU temp readings are a little low compared to >>what the BIOS reports too, but it's hard to verify. > > > See FAQ 4.12: > http://www2.lm-sensors.nu/~lm78/cvs/lm_sensors2/doc/lm_sensors-FAQ.html#Section-4_002e12 > > >>I also started powernowd, and when the CPU goes into low power mode not >>only does the clock frequency drop from 1.8GHz to 1.0GHz, but VCore also >>drops from 1.75V to 1.41V, which triggers an alarm. I suppose I can just >>set a lower minimum for VCore, but I wonder if that might mask a fault >>somewhere down the road? > > > No, that's actually what you should do. I don't know exactly how > powernowd works, but if it generates events when the speed and vcore > change, you might use these to reprogram the limits on the fly. That'd > allow you to have one short range for each nominal vcore, instead of one > large range covering both, which somewhat limits the efficiency of the > monitoring. > > I have to admit that our current architecture doesn't really cover this > situation, so it's left to you to implement the limts change as a hack > if you think it's worth the effort. > > -- > Jean Delvare > -- -- Howard Chu Chief Architect, Symas Corp. Director, Highland Sun http://www.symas.com http://highlandsun.com/hyc Symas: Premier OpenSource Development and Support
