Given intels secrecy regarding even cpu temps, I would plan on being stuck with guesses for tjmax for a long time.

I prefer real temp "guesses" as they are at least estimates based on a repeatable measurement, with some small possible error. Using mobile's to guess can lead to apparent random results.

For example all programs use 85 tjmax for E6600. Casing temps are 85C, when DTS=0, suggesting any gradient from core to casing at low voltage, underclocked settings, with no heatsink are less than 1-2C, ie less than measurement error. But for E8400, Casing temps are 95C when DTS=0, under same testing settings. So if coretemp and Everest were correct, than a 10C gradient exists from core to case on E8400 but 0C gradient exists on E6600, which is mathematically impossible, one of the two is incorrect.

And working out formulas as here post 916, http://www.xtremesystems.org/forums/...179044&page=37
I am pretty sure coretemp did way overestimate tjmax of E8400.

I would call Realtemp an estimate backed up by repeatable measurements and intel formulas. I would call what CT, E, etc are doing as "guessing".