Topic: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 1016. (Read 6590565 times)

I can't really understand one one of these GPUs are jumping on speeds, 23-29Mhs, until today everything worked fine.
Any ideas what can it be?
I use -ethi 8 and -mode 1
no any other specific commands

Anyone using -ttli?

Currently I am using using -li 1, with low energy roms.

Anyone has optimised settings for low intensity mining?

My goals are less temps and power consumption - I don't mind less hash.

you need to undervolt

Hey all, I'm having a really strange problem with Claymore running through nicehash. I have a dual 280x rig, and no matter what I do, it always seems to crash after about ten minutes. I've upped the voltage, tried underclocking, tested each card individually, etc. I have 15.12 drivers, I have 16gb of virtual memory, I've set the max_allocation value to 100 and set -nofee to 1 (I know this doesn't pay the dev, which I'd like to do, but someone said it could help" Still either freezes windows or crashes to a white or blue screen. Are there any known fixes for this?

I've changed the values in the config file, but I can't seem to get intensity below 5, could that have something to do with this?

thanks

89C with 100% Fan???

Dude, thats not gonna live long, you should lower that down a bit.

@mafaz
They are indeed RX 470 4GB cards
newegg link - https://www.newegg.com/Product/Product.aspx?Item=N82E16814131705&cm_re=powercolor_red_dragon-_-14-131-705-_-Product
CPU is a old Phenom II x965 Black
Motherboard is a old Asus M5A99X EVO
Ram is 8 GB of Corsair Vengance
PSU 850W Bronze running 4 of the cards
As to stability they ran for 48 hrs with no issues and only 2 Incorrect Shares I just rebooted the rig today cause I stole a card for my main PC as it's ancient GT 620 Died.

I have 5 of them 1 (93% Asic Quality) is running at 55W the other 4 (69-81% Asic Quality) are running at 125W+ all are Samsung ram.
Windows 10 Anniversary
AMD Driver version 16.11.5
Core clock set to 1242 (Was actually the magic number across the board for these)
Mem (Samsung) set to 2100 using the strap of (This was the magic strap for me. As using the 1500 Straps always crashed the cards over 1800)
777000000000000022CC1C0010626C49D0571016B50BD509004AE700140514207A8900A00300000 0191131399D2C3617
Powertune % set to 50 max (Never exceeded +25% on the poor ASIC cards)
I also have about 9-10 Hrs in each cards Bios finding the best TDP (W), TDC (A), and Max Power Limit (W) for each card. Very slow fine tuning but this is also my first round modding bioses with a goal in mind.

Image of MSI AfterBurner

Images of 4 of the cards the 5th is in the machine I am typing this on so it's hashing slower when I'm at the PC.

Why work it? The Poles are well digging and interest drip with us! And so everything is fine. Sea, women, margaritas !!!!

cant really believe this....
what processor do u use?
How many rams?
Motherboard?
core voltage and power limit?

how many incorrect shares?
and what about stability?
you sure your talking rx470? if yes is it MSI, Sapphire or which?

1242 Core/ 2100 Mem

2100 MHZ mem using the 1625 Strap the card was unstable and was maxing out at 27.7MH using 1500 Straps but the 1625 strap got me to 29.9 with SC at I 20 but my goal for the rig was over 30MH per card so I thought about it for a single card in a rig the best I could find was a SC I of 20 which got eth to 29.9 but when they are put into a 5 card rig SC I of 17 brought all 5 cards up to over 30MH.

I don't know my wall power but I am using a 850W Bronze PSU CPU is a old Phenom II x4 965 Black.

1500 straps? over/under clocks? my xfx black 4gb(elpida) 470s i can get a max stable of 1150/1975 and 28.3mhs ..with sia at any intensity that goes to 28.6 ...not worth the extra power.

That is Great!  What are your watts at the wall?

And I got it 30MH per RX470 4GB
The trick for the last 1 MH was finding the perfect SC Intensity for me it was 17

can someone help, best settings for sapphire 480 OC 8gb , stock bios - mining only eth
on stock settings I get 24.8 mh/s

drivers 16.12.2

p.s. i don't care about electricty, just stable settings


edit: hmmm tried this one
core 1100
memory 2180

and I get this
ETH - Total Speed: 51.315 Mh/s, Total Shares: 21, Rejected: 0, Time: 00:28
ETH: GPU0 26.569 Mh/s, GPU1 24.746 Mh/s

why is GPU0 have much more hashes than GPU1 ?? they have identical settings, temp etc.

Hi Claymore, may we please get an update on Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v7.4 (Windows/Linux), currently this is the most profitable way of mining and usually the most consistent. I'm pretty sure this should also gain from asm improvements.

Do we need to set virtual memory to 16GB even if we have 16GB/32GB RAM available on system?

How did u calculate this?

Man, i really need a power meter 

This is 200W from the wall!

It wasn't the actual motherboards just the pci-e ports. Ports stopped working after smoke came out of them. Yes i am using using powered risers (ver 7s).
25watts per card is too much price to pay for +1.5mhs anyway. I wont push them to hard.

Wich part of motherboard did you burn? 24 pin maybe? It is because of unpowered raisers besause too much current goes through MBO to power the gpu's. You must use powered risers. I also have burned MBO and power suply 24-pin connector when I had more then 4 gpu's with unpowered raisers.

GPU-Z is reporting about 120-125W GPU only power draw.
It's strange you would burn motherboards, pcie + 8 Pin power should be enough for quite a bit more than 120w.
Whats ur settings to go below 100w, and what is the hash rate at that wattage?