He said Yet it looks like the on chip sensor is more reliable, even at idle. 9C idle from the socket temp sensor, with 25C room temp does not compute! edit pls ignore me, got it back to front, good post Pill
I was thinking about taking my dremel and making a notch in the IHS on my 8350 so I can insert a temp probe. With the probe being that much closer to the CPU die, I wonder how much of a temp difference there would be between that and the socket temp.
Nice! I have to say I quite like Argus. It's very intuitive, and gives a great amount of detail as far as whats being monitored.
Thanks Phil. I was having a problem even finding a utility that would monitor CPU temp. Core temp jumps all over the place and it seems to me to be monitoring usage and not temp.
old info on TCTL value on thuban KTE posted this on overclockers.com regarding CoreTemp on 1055T http://www.overclockers.com/forums/showpost.php?p=6508650&postcount=19
That doesn't really demystify anything, sadly. Not that it isn't helpful, but this does not explain the actual implementation. Which is a shame, because I'd really like to know.
Hi, guy who wrote that post is mistaken, the 1055 only has 2 sensors just like all Phenom II's. The articles he linked to are from 2007, the X6 wasn't released until 2010 nearly 3 years later. Furthermore the articles have factual inaccuracies, for example EE Times states that each core has it's own power plane when in fact the voltage is constant across all cores. AMD subsequently disabled core unganging and reintroduced it again with Bulldozer......
At least the CPU can at least be read accurately using lm-sensors in Linux since it does support IT87xx Super I/O chips, but will require a manual compilation especially for the IT872x chips unless you're running Linux kernel 3.8+. deltatux