780 GTX overclock results

Not sure if I posted here but since I unlocked AB for voltage control I can now run 1228/7100 on the cards for benches with 85% fan speed (75% is a bit noisy but at 85% cards seem to settle down well).

Originally, without any volts I could only do 1135/6600. Volts increased to 1.25 to achieve the above overclocks.

Not bad. I remember you batching about the 1135. That's what my reference 780 did on stock voltage. 1170 something with 1.21. My amp card did 1228 at 1.21v.
 
after over a week of testing in multiple games/benchmarks, I think my Asus DirectCu II @ 1241/1675 is stable. I've tried skyn3t bios but stock gave better memory results and it was 2-3c cooler.
I am getting about 70c@GPU/85c@VRM absolutely max (long gaming period with 100% load, closed case front and case fans set to 7V, GPU fan maxing on silent 70%)

Very happy with the GPU. much better than R9 290 (@290X) I owned before
 
Not bad. I remember you batching about the 1135. That's what my reference 780 did on stock voltage. 1170 something with 1.21. My amp card did 1228 at 1.21v.
Cards were really starved for voltage but I think I also lost the core lottery since cards won't go past 1228 even with 100% fan speed and 1.275 volts. I even tried 1241 and they ran for a bit but crashed.
 
tdr, bsod, reboot, black screen, hard freeze, corrupted screen ... it can go down a lot of ways!

Is there anything specific about a bsod? Like a lack of volts or something lol like that?
 
Is there anything specific about a bsod? Like a lack of volts or something lol like that?

What did the minidump say (error code) for the BSOD? A bSOD from a GPU overclock is really not that common. You may have bad RAM/gpuRAM.
 
What did the minidump say (error code) for the BSOD? A bSOD from a GPU overclock is really not that common. You may have bad RAM/gpuRAM.

This is what I got last time

attempted_execute_of_noexecute_memory

I set the card to default and I'm gonna run for a couple hours on heaven to see if it recreates the issue even at stock
 
I would uninstalll drivers, clean out the reg with CCleaner after and install the drivers fresh again, with a clean installation through nvidia custom install. Then after run stock clocks in valley/heaven loop fro 45mins to an hour.

If you get BSOD after this run MEMTEST86 for a hour and a video memory stress test -- there are several programs to use (google).
 
Best I can get is 1306 at 1.28v with reference cards in SLI, but too hot for 24/7.

I run 1228 1.21v in summer and 1267 1.25v in winter (or if I want a bump for demanding games)

I've slightly adjusted fan curve so it's a little steeper, but it's not far from default. I don't find them noisy or intrusive at all, but they're certainly not as silent as stock settings.
 
Back
Top