Thanks philipma1957! I will.

@Looper_U @sxemini After some days of testing with the 2070 I managed to get 44MH/s with 125W.
I use these settings:
- core: -500
- memory: +1200
- power: 71% (the lowest)
With higher memory (+1300 or +1400) I got also 45MH/s but there were lots of rejected shares.

I have read on the internet about people getting 44-45MH/s from both 2070 and 2060s with 110W.
I own both those cards and from my personal experience I think those results are too optimistic, especially for the 2060s.
Maybe those values were achieved for a short time with unstable settings or maybe I'm just unlucky with the silicon lottery.

Anyway now that I've restored my 2070, I'll do some tests in the next days and see if I can get something better.

As far as the 2060s the best I get is 38.2MH/s @100W. I'm talking about STABLE settings for a card running 24/7
I'm currently using these settings on the 2060s:
- Core CLK: stock
- Gpu CLK: +1000
- Power Limit: 71% (the lowest)
I tried to raise power but I never reached something like 45MH/s. Do you have any suggestion?

Hello everyone I can not believe it, but I just solved my problem.
The low hashrate reason was in deed the BIOS being wrong. It caused the GPU to have a wrong power drawing from its 8-pin port.
I wrote to Palit/Gainward tech support and they gave me the correct BIOS.
Now the GPU gets 41.5 MH/s with 125W as it should!!!
I just want to say big thanks to everyone who helped me here. I couldn't do it without you guys!
I will leave this thread with [SOLVED] tag. I think it can be useful in case anyone in the future has the same or similar problem.

Hi, I have 5x 1080 ti for ETH mining and I can guarantee they're great., but there are a few things you must consider before buying.

1) They get 40+ MH/s as other cards do (RTX 2060 super, 2070, 2070 super), but power consumption is higher. They draw 170W even with good tuning, that is 30-40W higher than other newer cards, so if electricity cost is a problem, be warned

2) 1080 ti price is good and maybe roi (return of investment) is faster, but this is a 2017 model. When you buy 5 years old used cards, failure probability is higher and this can definitely affect the roi

3) We don't know for sure, but ETH mining is likely to end this year and 1080 ti was great for daggerhashimoto algorithm, but probably won't be as good with new profitable coins (ERGO, FLUX, RVN...). I'm not talking about performance, it's more a matter of efficiency (performance vs power consumption)

Thanks again for your help! These links are really helpful.
One last question: on techpowerup.com there isn't my gainward gpu firmware, but there are others from different brands very similar.

This is my gpu: https://www.gainward.com/main/vgapro.php?id=1025&tab=ov&lang=en
And this is a similar gpu from EVGA: https://www.techpowerup.com/vgabios/207314/evga-rtx2070-8192-181023-1

They both are dual fan RTX 2070 with 8-pin power port.
Do you think this EVGA firmware is going to be ok on my gainward model?

Thanks in advance

Thank you for your message!
Unfortunately among my other gpus I don't have another one exactly the same. I'll try and see if I can download its firmware from the internet.
I don't have experience in flashing gpus. Do you have some tutorial to suggest?
Thanks in advance

Thank you for your reply!
The Gainward gpu you linked is almost equal to mine, but it's not mine.
This is mine: https://www.gainward.com/main/vgapro.php?id=1025&tab=ov&lang=en
The only difference seems to be video outputs. All values are the same...

Anyway at the wall gpu is drawing something like 120W instead of 200W shown by miner and GPU-Z.
I know that 200W is crazy for a 2070, but that's basically the power consumption it should have with a 114% Afterburner power limit OC.
Another thing I forgot to mention is that gpu base clock during mining with OC is very volatile. It seems thermal throttling, but temperature is not high (less than 50 Celsius).

In the end I agree with you: I bought this as used and previous owner may have flashed it with a wrong version.

I'll see if I can get some help from Gainward website and facebook page, but I'm afraid that official support does not cover firmware flashing. I hope to find some help among other users, who maybe experienced something similar.
In the worst case I'll keep the card as it is with 114% power limit and 38 MH/s hashrate.

Hello I have a Gainward Nvidia RTX 2070 with power issues.
It draws less power than it should and it gets only 9 MH/s when mining ETH with DaggerHashimoto instead of 40+.
I tried different miners, different drivers, different risers, different slots and nothing changes. This happens with stock settings (no OC).

This is the problem: this gpu has only one 8-pin PCIE power port, while GPU-Z says there is another 6-pin PCIE port, that actually does NOT exist.
With several tests and a Watt-meter at the wall I discovered that the ACTUAL power drawn by gpu is the sum of:
1) power that flows through 8-pin port
2) power that flows through riser
Both GPU-Z and any miner show a higher drawn power, because they consider a 3rd component coming from 6-pin PCIE port, that, as written before, does NOT exist. This higher power is only virtual, since Watt-meter at the wall shows a lower power. Of course I took into consideration idle power and system power while doing the math.

At the moment I solved this way, but it's not that great. I set OC with MSI Afterburner at 114% power limit, giving the card a decent power and a decent hashrate (38MH/s).

My idea is that this gpu has a wrong firmware, that makes the whole PC think it has 8+6 pin power ports. Could that be the case? If that's true how can I solve? All my other 6 gpus work great.

Some details of my rig for you to know:
- OS: Windows 10
- Miner: NBMiner, T-Rex and others
- PSU: Corsair AX1200i + Corsair HX1200
- Biostar tb250-btc pro

Thanks in advance for your help and patience reading this huge text.

I'm starting to think it's something in the BIOS, too.
Is it safe using phoenix miner? I've nothing against it, but I read that NiceHash cancelled from their list for safety reasons.
As @batsonxl asked, what's your driver version?
Thanks everyone for your help!

GPU-Z says that it's Samsung memory.
I'm pretty sure my Afterburner (4.6.3) is the latest, since I've downloaded it from MSI website something like 1 week ago.

Every test has been made only with RTX 2070. GTX 1080 ti was always disabled or removed from PC.
At the moment driver version is 461.92. I also removed it with DDU and tried an older one (446.14), but nothing changes.

Thanks both Gorosden and batsonxl for replying!
Yes of course it's not RTX 2700, but RTX 2070. My mistake, sorry.

I tried your settings:

> core clock: -502, memory clock: +750, power limit: 60%
> system crashed (I think because memory clock was too high)

> core clock: -502, memory clock: +0, power limit: 100%
> nothing significant happens: core clock stuck @300MHz, memory clock @7000MHz and hashrate @9MH/s

> core clock: +1100, memory clock: +0, power limit: 100%
> same as before: core clock stuck @300MHz, memory clock @7000MHz and hashrate @9MH/s

In order to get core clock work properly I'm forced to set power limit over 100%

Hi everyone I'm new in the forum and I hope this is the right place for my help request.
I have a Gainward RTX 2700 and I'm wondering why my hash rate is only 9MH/s when mining ETH. I use NBMiner with Daggerhashimoto algorithm.
If I don't do any OC, my GPU mines with core clock of only 300MHz causing a very low hashrate. During this process memory clock stays @7000MHz as it should be.
In order to get a decent hashrate I'm forced to do some OC with MSI Afterburner. I don't touch neither core clock or memory clock, I just raise power limit to 114% (max possible) and then GPU starts to mine quite properly. Core clock stays between 1300-1400MHz and hashrate reaches of 38MH/s.
I'm pretty sure that if I was able to raise power limit beyond 114%, I could also get higher core clock and higher hashrate. Something like my GPU is thirsty for power.
My concern is more about power consumption rather than temperature. In fact with 114% power limit GPU needs 200W (and that's a lot!), while temperature stays safely under 45°C with 60% fan speed.
It is as if something is nerfing my GPU only when mining, because during stress test with FurMark everything goes fine according to values declared by manufacturer: core clock 1620MHZ, memory clock 7000MHz and 170W power consumption.
Thanks in advance to anyone who's willing to help me find a solution! 38MH/s is fine for me but I would like my GPU to need much less power.

Here there are some more details about my system:
- OS: Windows 10
- CPU: Intel Celeron G3930 with 16GB ram
- HDD: Corsair 256GB SSD
- PSU: Corsair Ax1200i
- MoBo: Biostar TB250-BTC-PRO
- GPU_1: RTX 2700 (described in the topic - 38MH/s @200W)
- GPU_2: GTX 1080 ti (42MH/s @170W with P0 state and ETHlargementPill both enabled)