Bitcoin Forum

Alternate cryptocurrencies => Mining (Altcoins) => Topic started by: Gunna01 on June 29, 2017, 10:50:39 PM



Title: Does Motherboard\Hdd\Ram affect hash?
Post by: Gunna01 on June 29, 2017, 10:50:39 PM
Instead of taking my rig down to flash cards I'm using an old PC to flash.

I copied BIOS from a RX570 thats getting 25mh/s and flashed it to an identical model\brand however that card started mining at 23mh/s using identical settings. Does mobo\hdd\ram play a part in hash rate?

Also is it normal to lose hash when reducing core clock in afterburner? As soon as I reduce core clock my hash starts to drop. Memory clock is at 2010, power limit 0 and core voltage 0


Title: Re: Does Motherboard\Hdd\Ram affect hash?
Post by: Vann on June 29, 2017, 11:03:46 PM
As long as there are enough system resources, HDD, CPU, RAM and PCI-E lane speed makes very little to no difference in hash rate. Almost any system will work. The biggest diffrence is how many cards you can get to work without a ton of grief. Core clock has a big effect in hash rate. Even more so for compute intensive algorithms like ZEC compared to memory dependent algorithms like ETH. Also every card has a different ASIC value from the factory which determines how well and easily the card overclocks. You can check the ASIC value with GPU-Z.


Title: Re: Does Motherboard\Hdd\Ram affect hash?
Post by: Juggar on June 29, 2017, 11:07:05 PM
I run 10 x AMD E-350 rigs (windows 10) with 64 gb SSD and 2 GB ram and I swear they are the slowest fuckin things.

They get the job done though, I have noticed that Claymore miner cant create a DAG Epoch with teamviewer is actively viewing the screen on that rig.

Teamviewer takes up 50% CPU on AMD E-350.


Title: Re: Does Motherboard\Hdd\Ram affect hash?
Post by: Gunna01 on June 29, 2017, 11:43:33 PM
As long as there are enough system resources, HDD, CPU, RAM and PCI-E lane speed makes very little to no difference in hash rate. Almost any system will work. The biggest diffrence is how many cards you can get to work without a ton of grief. Core clock has a big effect in hash rate. Even more so for compute intensive algorithms like ZEC compared to memory dependent algorithms like ETH. Also every card has a different ASIC value from the factory which determines how well and easily the card overclocks. You can check the ASIC value with GPU-Z.

That's what I thought however there was a clear reduction of 10% hashrate

So how do you reduce power draw if you cant underclock the GPU core?


Title: Re: Does Motherboard\Hdd\Ram affect hash?
Post by: Vann on June 29, 2017, 11:54:33 PM
As long as there are enough system resources, HDD, CPU, RAM and PCI-E lane speed makes very little to no difference in hash rate. Almost any system will work. The biggest diffrence is how many cards you can get to work without a ton of grief. Core clock has a big effect in hash rate. Even more so for compute intensive algorithms like ZEC compared to memory dependent algorithms like ETH. Also every card has a different ASIC value from the factory which determines how well and easily the card overclocks. You can check the ASIC value with GPU-Z.

That's what I thought however there was a clear reduction of 10% hashrate

So how do you reduce power draw if you cant underclock the GPU core?

You can undervolt the card with the top slider in Afterburner after you enable voltage control in the settings. 1150 MHz core clock with a -98 mV undervolt and 2075 MHz on the memory clock gives me 28 - 29 MH/s on a RX 570 4GB depending on the card.


Title: Re: Does Motherboard\Hdd\Ram affect hash?
Post by: grape_tectonics on June 30, 2017, 01:11:46 AM
There's also the memory quality of each card to consider, ECC working overtime can easily account for 10% of hashing performance. Check both cards with hwinfo while they are running to see how many memory errors they are coughing up, each error will either reduce overall performance or worse - ruin a share. If you really want the max out of each card, you got to find the sweet spot of memory errors vs memory speed for each of them, I've found that keeping them under ~100 per minute usually works best.