- Jun 30, 2004
- 16,889
- 2,201
- 126
I think we've collectively posted a lot about the mis-match of BIOS to thermal-sensoring of the Wolfdales, or the flawed TJunction sensors on those cores and even the Yorkfields.
To re-cap, Intel had posted (somewhere) a statement that those core-temperature sensors were "not meant to measure idle temperature [ranges]" -- sort a a disclaimer discouraging a flood of RMA actions.
With my E8600, the cores are "stuck" at 51C until the temperatures rise above that threshold; after that, they appear to work "normally."
Did later production-runs of the E0-stepping eliminate these limits on accurate thermal monitoring? I ask, because I'm weighing the possibility of ordering another CPU with hopes of getting accurate temperatures at the low-end.
This is subjective, but I feel more confident and trusting of the temperature readings at the high-end or "load," when the sensors function well at idle.
On the practical side, the TCase thermal-sensor on these seems to function reliably, and I'll need to search the Intel web-site to get a better idea of the differential between TCase and (average) TJunction core temperature values. Then, I can temper my over-clocking efforts with attentiion to TCase.
I want to punch up my CPU voltage from 1.30V to around 1.34V, so I can see what I can squeeze out of a set of G.SKILL DDR2-900 modules that run at stock 4,4,4,12 settings at the DDR=900Mhz speed. My only other option would be to drop my multiplier (10) to 9 and boost the FSB that way -- probably at the same voltage setting around 1.3V.
To re-cap, Intel had posted (somewhere) a statement that those core-temperature sensors were "not meant to measure idle temperature [ranges]" -- sort a a disclaimer discouraging a flood of RMA actions.
With my E8600, the cores are "stuck" at 51C until the temperatures rise above that threshold; after that, they appear to work "normally."
Did later production-runs of the E0-stepping eliminate these limits on accurate thermal monitoring? I ask, because I'm weighing the possibility of ordering another CPU with hopes of getting accurate temperatures at the low-end.
This is subjective, but I feel more confident and trusting of the temperature readings at the high-end or "load," when the sensors function well at idle.
On the practical side, the TCase thermal-sensor on these seems to function reliably, and I'll need to search the Intel web-site to get a better idea of the differential between TCase and (average) TJunction core temperature values. Then, I can temper my over-clocking efforts with attentiion to TCase.
I want to punch up my CPU voltage from 1.30V to around 1.34V, so I can see what I can squeeze out of a set of G.SKILL DDR2-900 modules that run at stock 4,4,4,12 settings at the DDR=900Mhz speed. My only other option would be to drop my multiplier (10) to 9 and boost the FSB that way -- probably at the same voltage setting around 1.3V.
Facebook
Twitter