Nope it's a GTX 1080 without Ti. It's a GTX 1080 Palit Super Jetstream.

@LoverAli94: I've used flypool and nanopool for some time to make a comparioson. Nanopool never gave me a high hashrate but the ZEC amount I was not what I was expecting. Flypool gave me a lower hashrate but the right ZEC amount.

About ping: I get 20-40 mostly. Sometimes when something is wrong I get around 200-300ms but my rejected shares are not really higher. I think in theory it's bad to have a high ping but it's not significant.

I run my 1080 mostly with 55% TDP (126W) +180 core (1809-1822 MHz) +1000 memory. Efficiency I get is 4.45 Sol/Watt (565 Sol)
This is my energy efficiency setup. I benefit mostly from the +1000 memory (P0).

I can go up to 640 Sol but it's noisy and not a great efficiency anymore. I use my PC while mining (via onboard GPU) so I'm interested in low noise primary.


@jarkle: What settings (tweaks) do you apply in Afterburner? I have the same numbers after restarting the miner while using Afterburner.

run with '--pec 1' parameter.

I see. I've also started with mining on Linux. But as much as I wanted to mine on Linux, I could not find a way to OC the card proper with Linux tools.
I do not know of a way to change the P2 state to P0 in Linux (Maybe someone does?).

But here is a solution for Windows if you want to try. This removes the memory clock limit during CUDA usage (P2 state).

1. Get the latest nVidia Profile Inspector (should be always latest: https://ci.appveyor.com/project/Orbmu2k/nvidiaprofileinspector/build/artifacts)
2. You propably need to run this as admin, but not sure about that.
3. Look for section 5 (5 - Common), change 'CUDA - Force P2 state' to off like in this Screenshot:
4. Restart CUDA app (Miner)

You should now see the P0 state active and also an increase in your vram clock. Since +1200 is really high I could imagine a sudden crash. Except this OC is proven stable in games.

What are your absolute clock and memory numers while mining?

Nice hashrate. But if I may ask, what is your absolute core/memory clock?
Since you run in P2 mode and not P0, your memory clock should be heavily limited, so that +1200 mem could have limited effect.

GPU's MODEL NAME: Palit GeForce GTX 1080 Super Jetstream
VRAM: 8GB GDDR5X
CORE/MEMORY CLOCK: 2126-2139 MHz (1.05V) / 6003 MHz (1501 MHz)
HASHRATE: ~645 Sol/s
MINER SOFTWARE: EWBF's CUDA Zcash miner 0.3.4b (fee deactivated via --fee 0)
OS: Windows 10 x64 Education - Version 1703
DRIVERS: 384.80 (Beta) / CUDA 8.0
TDP(OPTIONAL): 91% - 209 Watt

This is not my all time setup. Just a benchmark what you get if you raise the vram +1000 MHz (6000 MHz) with high core clock and let the FAN run at 100% with 58°C (25°C ambient). I probably could run at this clock but my girl would kill me.  

My normal clock is about 1900 MHz @ 142W and 60°C, which results in 580 Sol/s.

Screenshot:

Edit: Thanks for your effort in making the benchmark site.