Quote from: R0mi on December 26, 2017, 01:24:19 PM

Quote from: RayH on December 22, 2017, 12:34:22 PM

Hi All, I have a couple of questions if anyone can help

Can anyone tell me what the P Values in Overdriventool change. I believe P7 is the memory frequecy of the card and P3 the cards Ram? but what are the others.

also, I always stop the miner before making any changes in Overdriventool. Is it ok to make changes while the miner is running?

Thanks

Yes, it is OKAY to make changes in OverDriveNTool while the miner is running. Let's say you have 4 Vega cards and one is the proverbial mutt. If you are using the Jericho Jones script, you might opt to have the script load a pretty conservative, lower hashing rate OverDriveNTool profile just so that Cast or XMR start mining smoothly. Then you can go into OverDriveNTool and apply a different profile for the three cards that work great, one by one, while the miner is running. You should see changed hash rates for those cards in under a minute. Or you could create two different profiles, and load the slower one for the mutt card, and the better one for the three good cards using the VideoTools option explained within the script and at the vegaminingblog.

Changing settings in OverDriveNTool while the miner is running can help you determine which card(s) is or are the dog(s) (your miner will crash after you hit "apply" for that card, even if it did not crash for the other cards).

You can fine tune your setup in this way. But be aware that OverDriveNTool numbers the cards differently than the miner program. So change one card's settings and see which card # in the miner window is affected.

Do note that adjusting mem clocks by 20-30mhz will only yield tiny hashrate improvements in CryptoNight.

However reducing voltages will yield significant power savings and will help with cooling thus preventing thermal throttling. Therefore it is well worth doing - but you can't do that with the OverDriveNTool. To achieve this with Vegas you actually need to apply a regmod with a particular voltage for each card, restart your computer and miner and see if BSODs. Then try a different voltage, then repeat the whole process for each card. Much fun!

Of course, a card may decide to BSOD after an hour of running in which case you are back to square one

Which I why I ended up running fairly conservative voltages, sacrificing power savings...

Or did someone find a cleverer way of doing this?