Well many of you have seen my now
famous and extremely slutty 165 strut her stuff. At that time I was trying to get my core 1 to break through it's barrier at 2.6ghz. I figured maybe heat was the issue, which was my reason for making it nude. Since then, through many more hours of burn in and testing, I have found that core 1 will still not go past 2.65 or so with any sort of stability. Allow me to elaborate:
- 2600mhz, 1.375v - 24/7 dual prime stable.
- 2650mhz, 1.375v - core 0 24/7 stable, core 1 fails after 6-8 hours depending on how it feels.
- 2700mhz, 1.375v - core 0 fails after 4 hours, core 1 fails after 15mins.
------ upping voltage
- 2500mhz, 1.425v - 24/7 dual prime stable.
- 2600mhz, 1.425v - core 0 24/7 stable, core 1 fails instantly.
- 2650mhz, 1.425v - core 0 24/7 stable, core 1 fails instantly.
And the trend continues...
Now, my load temps under dual prime at 2600mhz 1.375v were around 47c or so with IHS on and are now ~41c-42c and can be maintained even lower with the denki at higher rpm's. At 1.425v, load temps go to around 50-52c with ihs on and 44-45c with ihs off. Behavior has been the same on core 1 even with the temperature difference.
I then examined the possible interference of the PWM chips. Perhaps their extreme heat was causing the issue. So I slapped on some swiftech heatsinks and plopped an old AXP stock fan on top of them bringing load temps down from over 65c to ~38c at the same voltage. Still, core 1 acts the same.
So, clearly with a cool pwm, a good power source (OCZ 600w in sig), good load temps on a naked chip itself, a lot of possibilities have been removed. Why on earth would core 1 actually be MORE stable at near stock voltage than even with a slight increase like 1.425v, especially if heat was kept under control as it has been? Any input would be greatly appreciated.
Linear Mode
