Some notes after putting together 3 new systems in the last month or so.
CPU temperature as reported by the motherboard is almost meaningless to compare to other systems, even of the same model.
* P5B Deluxe reports CPU temps roughly the same as Core Temp![]()
* P5W-DH reports CPU temps on average 18 degrees cooler than Core Tempedit: Early rev P5W with 6600
* Later rev P5W + 6400 = 9 degrees cooler than Core Temp
* both P5B and P5W systems begin to throttle exactly 3 seconds after hitting 82C Core Temp. Core Temp, calibrated or not, is reading the sensor that matters.
(all measurements at 22.5C ambient, average stable temps after 20 mins) All systems measured without any use of Rightmark, CrystalCPUID or SpeedStep. Both Systems with Intel 6600 C2D
Saving power while overclocked: (replacing SpeedStep)
* early revision P5W reports CPU temps 4 degrees lower than later revision. (tested with same CPU, fans and case)
* P5W + 6600 can drop 3C idle temps by using RightMark CPU Clock or CrystalCPUID to drop Multiplier from 9x to 6x (keeping same voltage)
* P5W + 6600 can drop 7C idle temps by using RightMark CPU Clock to drop Multiplier as well as throttle using ODCM.
* P5W + 6600 can drop 15C idle temps by using RightMark CPU Clock to enable C1E as well as throttle. (Drops Voltage to 1.1v – use with caution when overclocked – use in combination with ODCM throttling) This does not work on P5W + 6400 => crashes due to less multiplier options.
* P5B - Rightmark or Speed Step has no temperature effect whatsover that I can find. Dropping multiplier or throttling does nothing. Given that Idle CT temps are 10+ cooler than the P5W this implies that something is already enabling power saving, even with all C1E, SpeedStep disabled in BIOS.
A lot of this is already known and discussed, but I thought I would add my experiences.
All temps at 22.5 ambient, and measured by digital T-Balancer thermometer.
P5W-DH –67MOAG – 1.65v vMCH max revision
P5W-DH –68MOAG– 1.85v vMCH max revision
P5B Deluxe Wifi – 1.03G revision
Bookmarks