I'd like to know too. I'd like to know what the relation between the gpu and memory overclock is as well. How do you decide what one to keep pushing? Is there a certain ratio you should keep?
 
Also to the OP: have you seen this thread?
http://forums.evga.com/tm.aspx?m=1975128

Just as a frame of reference my plain reference 780 was able to get +160 on the core and +200 on memory. This gives me 1201MZhz on the core and 6400MHz (3200*2) on memory. I had it up to 1254MHz on core but it wasn't 100% stable so had to back it down.

If you back down AA to like 2x instead of 4x you should get 60+ most the game cept random times or even turn AA off completely. For me when you are playing online for example game is so fast paced you won't notice the edges being jagged. Truthfully i use high preset in the game instead of ultra.
 
my gtx780 acx
i have +50mhz gpu, +250 ram. atm havn't tried to push it anymore then that. it will top out 1175mhz with boost. With that said every gpu is different some overclock better then others, its kinda all about luck of the draw.

Higher temps can hurt the stability, but I would think 80C, while quite high, wouldn't be high enough to affect the stability greatly.  Keep in mind, every card is different and will behave different during overclocking.  It could be just that your card isn't one of the best overclockers.  However, even the most horrible cards I would think would get more offset than you were able to get. 
 
Another thing to keep in mind, is that your power supply is most likely getting near its capacity.  Depending on your system draw, I could easily think you are pushing close to 500W which is over 80% of your rated usage.  This is getting outside the peak efficiency band of the power supply, which is typically in the middle of its rating, say between 40-70%, of course it varies between brands and quality of power supply.  Also, if you power supply is older, they tend to degrade and lose efficiency with age.  Unstable or inconsistent voltage supplied to your GPU would be more of a concern regarding stability.
 
The purpose of applying the additional +38mV is to help with the ability to overclock and remain stable.  As one pushes the clocks higher and higher, it requires more and more power to remain stable.  This additional overvolt feeds this need for more power.  With a custom BIOS, such as with Classified cards and the like, more voltage can be set, but with the standard BIOS we can only apply +38mV, which usually equates to 1.2V as that is the max voltage limit set by Nvidia.
 
I have had crashes too without getting the notification, so I don't know if it was there but it disappeared before I could see it, or it was due to other reasons.  I highly suspect that 99% of the crashes we experience when we have overclocked cards are due to the driver crashing.
 
Lastly, regarding the memory clock reading, is the 1502Mhz reading at idle or during load?  I assume you are seeing the value in Precision X?  If that value is at idle, check what it reads during load.  It should jump up to 3000 for our 780s as standard.
 
Hope I answered all your questions, but feel free to shoot me a PM or here and would be happy to help out more.

Update: I installed the 2 new case fans and they've made a big difference in temperature. I'm now reaching no more than 64 degrees under load.
 
I tried OCing again I've been able to reach a GPU clock offset of +90 with +25mV for a GPU clock of 1188 MHz. When I use +38mV it tends to be unstable, and whenever it recovers from a crash PrecisionX resets the voltage back to +0, leading me to believe there's an issue with the power draw.
 
However, as soon as I touch the mem clock offset at all, it starts crashing right away. Any ideas what could be going on?

Trabe3, how are you recording what the core clock boots to? In 3dmark it isn't listed that I can see, and in Precision X, I can't see the GUI while the 3dmark runs.

I have dual monitors, but 3dmark blacks out the 2nd monitor when running the test for some reason...

Here's the results I got from running my GTX 760 4GB (Reference blower) using 3dmark on the Firestrike "Extreme" test for best results(score) without crashes so far:
 
Power Target 115% (MAX slider)
Temp Target 95° (MAX slider)
GPU Clock Offset +200MHz
Mem Clock Offset +240
 
GPU Temp hit 62° max
 
GPU-Z Report:
GPU Clock 1180Mhz Memory 1622Mhz Boost 1233Mhz
Default Clock 980Mhz Memory 1502MHz Boost 1033MHz
 
Any suggestions, or additional tests or configs to look at or run?

EVGA 780 ACX OC
Power 105%
GPU +90Mhz
Memory +70
Custom fan settings (60% at 50' - 90% at 80') and Voltage +25 mV.
Temp never higher than  67-68 C. Zero crash or problems from week.
 
My test showed: 106% power unstable under Heavens or Valley. Every step over +90 GPU or +70 Mem effect in higher max fps in benchmarks but, what is more important, much lower min fps (like 5 or 12). I tested with 105% power from 60/60 to 200/500, 150/300, 120/150, 120/120 etc end finished with 90/70 100% solid.
 
Unigine Heaven Benchmark 4.0
stock EVGA
min - max - aver - points
28,5 - 154 - 78 - 1977
OC
34   - 165 - 83 - 2103
 
Benchmark set: 1920 + ultra

Post released from Filter
 
Being an over 6 year old thread ... I'm not sure if you'll get a reply
 
I'm not sure How well EVGA Precision X1 works on the older cards .... or are you Using Precision X OC ?
 
 
I'd web search for OC guides
 
780 are Before NVIDIA's baked-in Boost control.  These older tech GPUs - required more testing & usually a Custom BIOS tweak for "Best Performance"
 

kronophormen
Nope, I'm currently using the precision X1, which is the replacement of the precision X OC. I tried to find some guides related to this GPU but i couldn't find none.

 
web search: EVGA GTX 780 SC ACX & OC guide   -or-  EVGA GTX 780 & OC guide
 
 
EVGA GeForce GTX 780 SC ACX review - Overclocking the ... 

[Official] NVIDIA GTX 780 Owner's Club - Overclock.net ...
 
 
Basic Steps ---> Keep it Cold as possible --> cold room, case fans @ 100%


As always there is no guarantee it will OC past factory specs.
 
With the OC software, adjust and Test and repeat --- small steps
 
Make sure your MB, CPU, RAM are stable First
 
1) slide the Power and Temp Target full right
 
2) Run baseline benchmark software Fire Strike or Heaven
 
3) Set custom fan curve, so that GPU Temp in C +10 = Fan %   50C temp = 60% fan (or evem more like +20)
 
4a) Increase your GPU Clock in small amounts maybe 100 MHz, then follow 50 MHz steps till unstable and back up in 25 MHz steps till stable
 
4b) Increase voltage and see if stable OC is higher
 
5) Take Max stable GPU MHz and reduce by 10% for stability
 
6) Repeat with RAM OC with GPU already OC
 
7) Heat matters on the GPU, if your RAM OC causes too much heat the GPU will slow.  Better to have max GPU clock than RAM clock
 
8) Benchmark software is synthetic, your favorite game may need different setting to be stable
 
 
More Voltage = More Heat
 
If your GPU heats up with longer term use - several hours of gaming or Folding -- then you have to cool it better or reduce the OC or Voltage because GPU MHz speed is King

 
Sorry to say that I do not have the "magic numbers" ----- it is an adjust & test situation -- "silicon lottery"
 
-- GPU can all be different, and the rest of your system does make a difference
 
Temps Matter - Room & GPU are related  DeltaT -- the difference in Temp is what we look at
 
Have FUN with your OC adventure