Winnie 3200+ Temps(Asus A8N-SLI vs Gigabyte K8N Ultra-9)

[addyngan]

Hi,

I am upgrading my system to AMD64, and I have tried two motherboards so far which give me significantly temperature.

My current config:

AMD Winchester 3200+, Asus A8N-SLI Deluxe, ATI X800XL, 2x512MB Kingston.
Antec Sonata, Nexus 120mm on HD cage, Stock case fan, stock PSU, NB47J on nforce4 chipset, Zalman 7000-AlCu at 5V (min setting)
Hitachi 160GB SATA, Maxtor 160GB IDE x2 on RAID 1

Currently the CPU idle temperature is about 37, and load is 46. MB temp is 32. I have replaced the Asus chipset fan with the Zalman NB47J, and i hope the front fan will help cooling the chipset.

Anyway it seems this CPU temp is quite high for the winchester 3200+ , and more importantly, i have much lower CPU temperature when i install the Gigabyte K8N Ultra-9 (with stock passive nforce4 heatsink), with everything else the same. With that setup, I have idle 31, and 40 load.

The only significant difference between the two boards, as far as I can tell, is the location of the graphics board. On the Gigabyte board, the PCIEx16 port is much further away from the CPU (basically the 2nd slot of the SLI board), while on the Asus, I am forced to use the slot closer to the CPU. Do you think that affect the airflow and thus the higher temperature? Or do u think i may have messed up sth during the heatsink installation? I am using AS5, but given my limited experience, i can't say if i did a good job of applying it.

Addy

[merlyn]

Calibrate Your CPU Temp Reporting

[m0002a]

Anyway it seems this CPU temp is quite high for the winchester 3200+ , and more importantly, i have much lower CPU temperature when i install the Gigabyte K8N Ultra-9 (with stock passive nforce4 heatsink), with everything else the same. With that setup, I have idle 31, and 40 load.

You are assuming that the two motherboards report the temps from the same sensors, or that the bios does not change the temps (for whatever reasons) before reporting them to you. You can not reliably compare reported temps from two different MB's.

[addyngan]

i see.. i thought the temperature comes straight from the CPU and was expecting reasonable comparison.

Anyway, what do u think the error of the temperature would be? +- 5C? And is my 37C/46C idle/load temperature in good shape for my setup?

If i were to get some thermal probe, would it help improving the accuracy? Any suggestions for a good probe?

Addy

[Tibors]

The probe would be less accurate. If you follow the instructions in the article linked in the second post, then you get the most accurate numbers.

Some mobo's have been reported to be as much as 10°C off.

The temps are good if the system is stable.