MSI RTX 3070 Gaming X Trio - Same Model but different hashrate

Hello Everyone

I have 2 identical graphic cards and the hashrate is “a lot” different with the same oc settings.
How can i change that?

use fixed core instead

and yet its the solution to your problem.

increase your power limit so the card isnt throttling itself, or do what i said first, use a fixed core s o the card uses less power and doesnt hit the power limit you set, and theyll get the same hashrate, or just call the solution a useless reply lmao

Just what @keaton_hiveon wrote… Put your core clock on +1100 instead of -1100 and than their hashrate should be the same.

1 Like

Lower temperatures under 40 helped with GPU 0. But it’s just not an answer why 2 identical cards with the same settings perform different.
With the current settings, I get good temperatures, besides the old RTX 3070 220W (MSI RTX 3070 Ventus 3x)

125 w is perfect with Phonix miner. If i use T-Rex miner I can lower them to 120 W with the same result on the 250 w cards, but on the old 220 w card I start getting invalid shares (before only Stale shares).

If i raise the clocks to +. Will i not get higher temperatures?

It must be a manufacturing problem or because that card is second hand.

Every card is different. So power consumption wont be identical. Setting a power limit of 125w for each card won’t give the same hashrate. As you already wrote if the card becomes cooler, hashrate will rise. This is because internal thermal resistance becomes higher and GPU needs less power. There is nothing wrong with your card.

But if you don’t listen to us and set a positive +1100 core clock, we cant help you. The positive core clock will probably lower the power consumption and temperature. Good luck

Thanks for a perfect answer and explanation. I have tried the +1100 and after 10 min. it looks like the diffrence is 0-1 degrees.

62.36 MH seems a bit high?

If you get invalid shares. You should lower the men speed. I personally wouldn’t go above 2400

I only get invalid shares (Ethermine) with T-rex miner on the 220w card (Ventus x3). On Phoenix miner it’s only 1-2% Stale shares on the Ventus x3. The other cards (Gaming X Trio) works fine.

Do you know why that card Ventus x3 sometimes is showing 0% fan, but the fans are still spinning.
But the cards temperature is rising to 50+ degrees. My only option here is to “shut down & boot in 30” and keep doing that, until the fans speed is showing again.

I have read in other threads that there are some bugs with the nvidia and linux regarding fan read out. I had a rig with 6 3060ti/3070, at startup sometimes one card didn’t report fanspeed. after reboot it did. But i haven’t seen that it stopped during mining.

I would advice to test it for a day with memory speed on 2400 for that card. Maybe this will help.

This is the setting that I use for my 3070 ventus 3x

With your OC settings i could not get more than 60 MH and when i use + core settings, my power supply is too weak for 125w, so this was my best settings with PhoenixMiner.

and my best OC settings with T-Rex miner.

Looks :ok_hand:. What you can do for fine-tuning lower the core with steps of 15 (1085, 1070, etc) until hashrate drops. The lower the core clock, the lower power consumption will be.

Thank you.
I have tested some more clock settings, but the Power Limit is still different. So i guess I can choose +1100, +2600 and 125 with the same hash rate or -1100, +2600 and 125 with the same Power Limit and 0-1 different hashrate.

With the same hashrate and different power consuming i am saving 20-30w :+1:

Thats the silicon lottery :slight_smile: … Some cards are better than the others and needs less power for the same hashrate. Power consumption also depends on temperature. The hotter a card gets, the more power power it will consume (not much, but it will). Good luck with mining

1 Like

I have a question, why are you keeping the cards so “cold”. I mean wouldnt they work fine around 50 degrees? I have one RX 570 which works 51-52 degrees with 47% fan speed currently. Isnt it ok like this ? I see people going over 60 degrees.

I can see that CPU 3 sometimes is going down 0.5-1 MH, when the temperature hits 40+ degrees and I get more stale shares.

I live in Thailand and the air condition is on 22 hours per day at 25 degree (0,13 usd per kWh), so I guess it didn’t matter to have my hardware as cold as possible.

Besides that I also have a XCH plotter where the CPU and NVMe drives runs hot.

Why do you worry too much about 0.5-1 MH? Like what everyone said every card is different. You should be aiming more on stability and efficiency rather than a small difference in hashrate. You also need to check the speed on pool. That ‘slower’ card might be really faster in the pool you might be surprised