Hi,
Most of us here know that the DTS sensor in Brisbane core CPUs is broken and shows wierd (low) temperatures in SpeedFan.
Mr. Milani-Comparetti (creator of SpeedFan) tells me that he is considering adding a built-in +29C temperature offset on the 4850e.
Is +29C realistic? Seems quite high to me.
Has anyone tried to find out exactly how far the incorrect SpeedFan core temps are from the actual core temp on CPUs with Brisbane core?
Brisbane core CPUs and Speedfan
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
i also find the cpu reading from my mobo to be fairly accurate.
from memory i think there was about a 20C offset- the core reading from speedfan seemed to follow the cpu temp quite closely with the offset in place.
i can do some testing tonight if you want confirmation, but i'm afraid that there are differences between batches of cpus.
from memory i think there was about a 20C offset- the core reading from speedfan seemed to follow the cpu temp quite closely with the offset in place.
i can do some testing tonight if you want confirmation, but i'm afraid that there are differences between batches of cpus.
-
RedAE102
- Friend of SPCR
- Posts: 118
- Joined: Wed Jul 04, 2007 3:29 pm
- Location: Lost and Found Bin, Cypress, CA, USA
- Contact:
On a side note, the newest versions of Everest now support proper Brisbane core temperature sensing. I use it on my Brisbane Athlon X2 5400. The main CPU temperature diode reading still shows temperatures around 0°C, but the two core temperature readings show ~30°C at idle (1.2 GHz @ 0.768 V) and ~55°C under prolonged heavy load (2.8 GHz @ 1.225 V). You could get a [illegal] copy with a serial number via bit torrent like I did, or you could just use the 30-day demo version to gather data for a proper offset to add in Speedfan. I kept Everest because Speedfan won't get core temperatures for my HD 4670.
