Now first remember i have only tested this on the quad core Kentsfield. Stevty had comments like this back inthe days of P-D smithfields....
All cores loaded for at least an hour...all temps are core temp reported...all vcores are set in bios and actual load vcore droops .08-.09v...C2D chips on same mobo only saw .03v max droop. All rails are tight on PSU.
________________________________________________________________________
Basically when I set my QX6700 to 1.4v bios(1.32v actual under load) and 3.2ghz....
Average temps are 48-50c at idle and 63-66c under load....
Using rightmark utility and fan control I was able to reduce fan speed to a crawl and get temps to hit 78c and NO throttling occurs....
________________________________________________________________________
When I set it to 1.4125v in bios (1.33v actual under load) and 3.3ghz
Average temps are 51-53c at idle and 66-69c under load...
Again using Rightmark and reducing fan speed I raise the temps...At around 75c I begin to see throttling in cores 0 and 3...It does this by reducing the multiplier to half...It does this very regularly in a pattern and like clockwork...Temp hits 75ish and it throttles for about 10 seconds until it drops to 70c and then goes back to full juice...throttling does not happen on all 4 cores...just 2 cores equally...The wouldn't effect times in test if you use affinity. I did not at first so it drug all times down equally...
________________________________________________________________________
When I set it to 3.346ghz and 1.425v in bios (1.34v actual under load)
Average temps are 52-55c at idle and 68-71c under load
At this speed I barely had to reduce fan speed as throttling appears to happen right around 71c....similar pattern as above though throttle periods are longer and more frequent...
I raised fan speed back up and it still would occur buty not as often as the temp just spiked to 71c occasionally...Threw open the window and let some 45f air come in and temps shot down to low 60's and it ran for 4-5 hours with no throttling...
__________________________________________________________________________
At vcore of 1.45v I can run 3.4ghz but throttling is happening almost out of the gate when loading...
So stevty was right...the higher the vcore the lower the point at which throttling occurs...this must have been TM1 throttling cauise I can only disable TM2 in the bios
This QX6700 is a B0 steppng 4 chips so the newer chips may definitely run cooler. It appears in my case I need a better cooler then my Ninja Scythe if I really want to go much past 3.2ghz...
It boots at 3.46ghz fine into windows and can even post at 3.5ghz into bios...Too much vcore and throttling will occur so frequently is actually runs slower then when I leave it at 3.2ghz. The throttling at 3.3 and 3.35ghz makes drop 15-30% in performance.
May try water...may not...may just be happy with 4 cores of 3.2ghz....in my testing 3.3ghz will not run at 1.4v so I am forced to raise vcore after 3.25ghz.
*****I need to try this on some of the C2D chips...many of us are giving them 1.45v-1.475v****
Interesting to see where throttling occurs....
All cores loaded for at least an hour...all temps are core temp reported...all vcores are set in bios and actual load vcore droops .08-.09v...C2D chips on same mobo only saw .03v max droop. All rails are tight on PSU.
________________________________________________________________________
Basically when I set my QX6700 to 1.4v bios(1.32v actual under load) and 3.2ghz....
Average temps are 48-50c at idle and 63-66c under load....
Using rightmark utility and fan control I was able to reduce fan speed to a crawl and get temps to hit 78c and NO throttling occurs....
________________________________________________________________________
When I set it to 1.4125v in bios (1.33v actual under load) and 3.3ghz
Average temps are 51-53c at idle and 66-69c under load...
Again using Rightmark and reducing fan speed I raise the temps...At around 75c I begin to see throttling in cores 0 and 3...It does this by reducing the multiplier to half...It does this very regularly in a pattern and like clockwork...Temp hits 75ish and it throttles for about 10 seconds until it drops to 70c and then goes back to full juice...throttling does not happen on all 4 cores...just 2 cores equally...The wouldn't effect times in test if you use affinity. I did not at first so it drug all times down equally...
________________________________________________________________________
When I set it to 3.346ghz and 1.425v in bios (1.34v actual under load)
Average temps are 52-55c at idle and 68-71c under load
At this speed I barely had to reduce fan speed as throttling appears to happen right around 71c....similar pattern as above though throttle periods are longer and more frequent...
I raised fan speed back up and it still would occur buty not as often as the temp just spiked to 71c occasionally...Threw open the window and let some 45f air come in and temps shot down to low 60's and it ran for 4-5 hours with no throttling...
__________________________________________________________________________
At vcore of 1.45v I can run 3.4ghz but throttling is happening almost out of the gate when loading...
So stevty was right...the higher the vcore the lower the point at which throttling occurs...this must have been TM1 throttling cauise I can only disable TM2 in the bios
This QX6700 is a B0 steppng 4 chips so the newer chips may definitely run cooler. It appears in my case I need a better cooler then my Ninja Scythe if I really want to go much past 3.2ghz...
It boots at 3.46ghz fine into windows and can even post at 3.5ghz into bios...Too much vcore and throttling will occur so frequently is actually runs slower then when I leave it at 3.2ghz. The throttling at 3.3 and 3.35ghz makes drop 15-30% in performance.
May try water...may not...may just be happy with 4 cores of 3.2ghz....in my testing 3.3ghz will not run at 1.4v so I am forced to raise vcore after 3.25ghz.
*****I need to try this on some of the C2D chips...many of us are giving them 1.45v-1.475v****
Interesting to see where throttling occurs....
Facebook
Twitter