Re: PhoenixMiner 5.1c: fastest Ethereum/Ethash miner with lowest devfee (Win/Linux)

Anybody using 1070? Two years ago all of the 1070s were pushing a constant 30-32 mh/s. Now, only one (EVGA) of them is pushing 29 mh/s, the rest (AERO MSI) are in the low 22-23 mh/s. I understand the 1070s are affected by the increased DAG size as well as nvidia's drivers. However, how come one of my card still pushes similar numbers to two years ago, is the EVGA using a more efficient memory or what causes this? I've tweaked mem/clock and TDP but nothing. Thanks.

OS: Win 10
Newest Nividia drivers
Latest/Beta PM.
Straps/vrtimings, etc. all applied

Thats not true. 1070s are NOT affected by DAG size so far.
All mine are hashing at 30 Mh/s with no problem.

It all depends on your system (power usage, OS, memory, driver...)

You are wrong, every Nvidia Pascal Card, reduce the hashrate with every new DAG epoche. It's the same bug as the AMD R9 290x and R9 390 in the past.
The reduction is not much, but it will happened, You see it, when you let your miner benchmark in epoch 0 and for example in epoch 300. Test it and you will see it dude;)

The user was referring a hashing speed of 22mh/s on a 1070.. and thats not related to DAG size. He has another problem, as i have all my 1070s at almost 30mh/s (@~125w)

I recall reading posts and links to nvidia forums stating that 1070s do indeed have a reduction of total mh/s output due to the DAG size. Forgive me as I don't have the link on hand.

Update: I unplugged everything, tried one gpu at a time and ran at 22.5 mh/s and that one odd at 29 mh/s. I plugged everything back, now all are reporting 29 mh/w and one is stuck at 23 mh/s. Any guesses? Because I have no idea what's going on.

To the other poster, my gpus are running at 65C-70C, it's very cold here and one of the reasons I decided to turn it back on. My settings are +100/600+ @70TDP (I tested up to 80 but there's a tiny bit of increase .5).

Thank you all

With reasonable ambient temps and normal airflow you shouldn't be getting those kind of temps. You probably have dried out paste that is no longer forming a good conduction layer. If you are comfortable take the cooler assembly off and repaste each card with a decent silver compound - doesn't have to be expensive. Usually all cards will need this if they've been going 2 or 3 years nonstop.

Thanks for the suggestion. However, I still have about a year left of warranty and servicing them myself would probably void it (the gpus have stickers on top of the screws).

Quote from: lindvior on November 17, 2020, 01:33:31 PM

Quote from: DooGZ on November 17, 2020, 12:16:46 PM

Quote

My 1070's can easily reach 30-32mhs around 2018, current speed is 25-27mhs with the same OC as before, needed to increase core speed + voltage + "-straps" to reach 29-30mhs again.
Same with 1060's, from 24mhs, went down to 21-22mhs with the same 2018 OC.

RX's Polaris are not affected.

First post here

Same here.

I am using EVGA 1070 and my speed now is 25 - 26.

It was 29 - 30 in 2017 / 2018

Most of us 1070s owners are reporting about the same as you.

Quote from: Binary100100 on November 18, 2020, 10:23:52 PM

I have eight GTX 1070 Founders Edition cards mining ETH.
Using HiveOS on 440.82 drivers I'm averaging 28.25MH/s per card.
My temps are good (~68C maximum) with the fans running at 75%.
I have 150 Core 1400 Memory and 124 (w) Power.
That's probably the best that I can do at this moment. I'm constantly tweaking and re-adjusting though. I'm open to suggestions.

I'm under Win 10 and have about the same settings, the only difference is my gpus are 1 or 2 C higher than yours. Anyway, hopefully our posts help other 1070s owners.

My EVGA 1070 rig still running PM 5.1C.

It work perfectly

Thank You Phoenix