Best way to monitor temperatures?

[djglenn]

Hey,

As some may know (through various threads) I have recently upgraded my PC..now running all in a Antec P182.

What is the best way for me to get accurate temperatures of components (apart from bios)?

I tried speedfan in the past, but I found it to be way out..

I have a ASUS P5Q board, does asus provide software?

[Oldjim]

I use Core Temp for the CPU core temperatures (better than the normal sensor)

[PeterStoba]

Core or RealTemp

[GoNz0]

everest beta as its free and the best out there afaik http://www.lavalys.com/beta/everestu...yc4xlkn0gs.zip

[Emirzan]

Core temp is good, a program like everest will detect sensors on CPU, hard disks and GPU and display temperatures for you. I don't think the P5Q has temperature sensors for the chipsets (this review mentions it) so if you want to monitor those you'll need to put some temperature probes in.

On my system I use the Scythe kaze master fan controller which also comes with 4 temperature sensors. As far as I can tell they are pretty accurate and you can position them whereever you want, so next to chips, heatsinks, fans etc. There are other cheaper products that also have temperature sensors with a display too.

Fan controllers

edit: HW monitor - Free and shows loads of useful information like temperature sensors and fan speeds

[Mithrandir]

I use the following:

Core temp

ATI tray tools

HWMonitor

Everest

[xAmritx]

Realtemp is the best one and lets you adjust your TJmax.

[SiM]

Realtemp is the best one and lets you adjust your TJmax.

So does coretemp

They are the same... realtemp's only advantage is that it records minimums and maximums too

[TheReaFireblade]

GPU-Z - the graphics equivalent of CPU-Z, shows your graphics temp's

[peterb]

All these (software) monitors use the same internal sensors - built into the CPU and motherboard - so unless you know the calibration functions of these sensors, you cannot determine which is 'best' or the more accurate because you don't know what the actual temperature is at the sensor to make a comparison (and the fact that a readout says 56.7798 degrees (to how ever many DPs) doesn't mean that it is that - at best you can probably say that it is between 55 and 57 degrees - which is good enough - and good enough for comparative readings over time, where a change might be important.

To measure the actual temperaturwe of a component, you need a calibrated probe which you can fix to a heatsink - then you can use the mfrs data for the thermal resistance of the heatsink to guesstimate the temperature at the heat source - but it will only be a guesstimate.

Silican semiconductor junctions can withstand (and function at) temperatures well into the 90C region, and Intel data sheets show that some processors can function with a Tj higher than that - and if that is approached, the CPU will throttle or in extremis, shut down

Obviously you want to avoid that, but while high temperatures may shorten component life, anything below 70 isn't really wotrth worrying about.

As I say - trend is the important thing - if temperature is increasing over time, you might want to investigate why - but there are more important things to lose sleep over thanh whrether your CPU is running at 54 degrees or 56 degrees!

[djglenn]

Thanks for the help!

With core temp I am getting the below temps on the Q6600. Any reason why core 0 is hotter than the rest? Have I incorrectly applied the thermal paste? I followed the AS5 guide to the T......

Core 0 - 33/34oC
Core 1 - 30oC
Core 2 - 28oC
Core 3 - 31oC

[Blitzen]

everest beta as its free and the best out there afaik http://www.lavalys.com/beta/everestu...yc4xlkn0gs.zip

+1
Get the best....Everest

One for the oldies (including me http://www.visit4info.com/advert/Eve...-Glazing/24721

[peterb]

Thanks for the help!

With core temp I am getting the below temps on the Q6600. Any reason why core 0 is hotter than the rest? Have I incorrectly applied the thermal paste? I followed the AS5 guide to the T......

Core 0 - 33/34oC
Core 1 - 30oC
Core 2 - 28oC
Core 3 - 31oC

Slight differernces in loading - differences in the sensor characteristics, possible differences in the heat transfer between each die and the heatsink. - all possible reasons.

Heat transfer differences might (only might) be because of the way the thermal compound was applied - but if you re did it there is no gurantee that the result would be better - it could be worse, and there is the risk of doing damage to something wwhile you are taking it apart and reassembling it.

I'd live with it - and enjoy doing what you built the computer for, rather than worry about a few degrees reported (you don't know that they are actual) temperature differences.

But those are low temperatures - only a few degrees above ambient - if you touch the heatsink, it will be mildly warm to the touch.. now if it burnt you would have something to worry about

[jonh17]

I use HWmonitor and find it probably to be the most accurate.

[peterb]

I use HWmonitor and find it probably to be the most accurate.

How do you know? What have you calibrated it against?

[GoNz0]

core 1 gets more use, and the sensors aint perfect, they all read a little out