My RX580 Benchmarking and Results: 30.5Mh/s @ 77.5W (+ 1130MH/s Dcr w/ 125W)

[speedyb]

what should i change to reduce the wattage the gpu uses since i know have my memory OC and my gpu core underclocked. am using srbpolaris is a value for tdp but also a value for max power limit and small power limit? or if its the VDDC i should change what is a good vddc for a 1100mhz core on a rx580 an rx480

I'm using the voltage values recommended by 1stminingrig when I downloaded the rom from there and it's been working like a charm for 90w, I haven't done any further customisation but it can be a good starting point for you

http://1stminingrig.com/best-bios-rom-sapphire-nitro-rx-580-8gb-oc-hynix-memory-30-mhs/

what undervolty worked for you mine seems to crash when i undervolt it around 100 and 50 where is stable shows no watt difference

[1714064068]

1714064068

[1714064068]

1714064068

[1714064068]

1714064068

[coinnoob_za]

Have you guys tried the new AMD Beta driver just released specifically for Mining?
I got my MSI RX580 Armor 8GB with hynix memory (flashed with modded VBIOS - 1750 straps copied to 2000)

With Epoch 139 now my card was down to 28.3MH/s. With the new driver I am getting 30.5MH/s

I found that after installing the driver do not patch the driver.
Use AB 4.4.0 beta 16
The undervolting does not work well for AB, but i used Global Wattman that installs with the driver, Set Voltage manually to 990mV but it didnt work, so changed it back to Automatically. After that it applied the 990mV correctly.
So the ADL API is very buggy with the new driver but the driver definately solves the hasrate drop on Polaris. (Do not specify the clocks or voltage in claymore miner as it does not work)
AB sets the Clocks fine, it's just the voltage that is a bit of a problem.

GPUcore is 1250, Mem 2200.

[windsor_43]

I need help? help when it comes to undervolting my system when dual mining.

I have 6 Red Devil Golden Sample Edition cards, i am using claymore 9.7, afterburner beta 16

My issue is when i am single mining #Eth i can undervolt via afterburner to the lowest settings which 950 (ish) and system follows this settings.

When i am dual mining without setting afterburner to limit voltage, my system uses up about 1450w, so i try to limit voltage to 950/1100/1300 and none of them are respected. i am a little stuck on why this is and just seek my help from those who are having this issue.

[androstan1234]

Is it really that much more  work to set voltage in vbios.

[greatauror28]

Is it really that much more  work to set voltage in vbios.

Not really, as long as your mining software will read it - which is not the case for Claymore 9.8 + AMD blockchain drivers.

I've set mine via SRB but Claymore ignores it. Ended up setting core voltages via Trixx and it's working like a charm.

Using 6x Asus Dual OC Rx580 8GB doing 29.3 eth and 733 dcr @ 118w (gpu-z).

[CryptoSquad]

HI

Can anyone tell me what hash rate i can expect from the below?

https://www.asus.com/uk/Graphics-Cards/DUAL-RX580-O4G/

Will i need a custom bios? If so which one? With a custom bios what hash rate can i expect? What power usage should i expect per card?

I bought 8, but only want to use a single 1200watt power supply. So guessing i may not be able to run all 8 cards?

Also, in my NVIDIA Rigs ive been using a specially designed Nvidia linux image to run the rigs.

What am i best running on this new RX580 Rig?  Windows? Or is there a dedicated AMD Linux image i could work with?

As much info as possible please!

Thanks in advance.

[greatauror28]

HI

Can anyone tell me what hash rate i can expect from the below?

https://www.asus.com/uk/Graphics-Cards/DUAL-RX580-O4G/

Will i need a custom bios? If so which one? With a custom bios what hash rate can i expect? What power usage should i expect per card?

I bought 8, but only want to use a single 1200watt power supply. So guessing i may not be able to run all 8 cards?

Also, in my NVIDIA Rigs ive been using a specially designed Nvidia linux image to run the rigs.

What am i best running on this new RX580 Rig?  Windows? Or is there a dedicated AMD Linux image i could work with?

As much info as possible please!

Thanks in advance.

Can anyone tell me what hash rate i can expect from the below?

https://www.asus.com/uk/Graphics-Cards/DUAL-RX580-O4G/ All of my Dual OC 580s (8gb) are Samsung.

Will i need a custom bios?Yes if you want a better hashrate
If so which one? You can buy mine, or do trial-and-error yourself
With a custom bios what hash rate can i expect? With my 8GB, I do 29.3 ETH plus 733 DCR
What power usage should i expect per card? Mine pulls 160w at the wall

I bought 8, but only want to use a single 1200watt power supply. So guessing i may not be able to run all 8 cards? GPUs rull full power upon start up, so no. Even undervolted, you'd need 1300w at least. I use 2x 750w.

Also, in my NVIDIA Rigs ive been using a specially designed Nvidia linux image to run the rigs.

What am i best running on this new RX580 Rig?  Windows? Windows 10 would be the easiest.
Or is there a dedicated AMD Linux image i could work with? If you know Linux, I suggest going that route. If not stick with Win10.

As much info as possible please! One side says it's not profitable to even think of starting to build a rig nowadays. The other side says you can mine other coins if ETH goes south.

Thanks in advance.

[FirstRigger]

Hi,

everybody can help me please?
RX580 Sapphire Nitro+ 8Gb Hynix, Win10 64bit, bios mod, ati patch, ~29.5Mh/s.
AsRock H170m Pro4 mb, 550W Enermax ps.
Aug-23 blockchain driver.

Claymore 10.0 start.bat:

GPU_FORCE_64BIT_PTR 0
GPU_MAX_HEAP_SIZE 100
GPU_USE_SYNC_OBJECTS 1
GPU_MAX_ALLOC_PERCENT 100
GPU_SINGLE_ALLOC_PERCENT 100

EthDcrMiner64.exe -epool eu1.ethermine.org:4444 -ewal 0xe030ABB211fd08a62800de181788D48511677fFE.Mini1 -epsw x -dpool stratum+tcp://lbry.suprnova.cc:6256 -dwal JoJocy.Mini1 -dpsw x -dcoin lbc -etha 0 -dcri 33 -tt 65 -ttdcr 75 -ttli 79 -tstop 85 -fanmin 30 -fanmax 100 -cclock 1130 -mclock 2150 -cvddc 850 -mvddc 850 -platform 1


Now, my win10 is freeze and freez after reboot too. I must to disable the RX580.
HW info show on picture:

https://www.dropbox.com/s/lztsyyftberh0on/problema.JPG?dl=0

[bouldou]

@FirstRigger : try to disable dual mining, does it freeze too ? Maybe your 550W Enermax is not enough... would be strange but try

[Brockshad]

I'm getting 30 -31 Mh/s per card on my 8 x RX580 Sapphire Nitro+. Rig pulls about 1450 Watts. See the video below if you have been trying to get them to run at 30 Mh/s with improved power consumption. If you still have questions let me know.

https://www.youtube.com/watch?v=8clkIopKZ-Q&t=1s

[vasilykozlov]

30mh/s too now. tx mate

[FirstRigger]

Hi Guys,

thank you for your time and help.
After installed the fresh Win10, I downloaded the rom what Brockshad suggest and now its very good. The miner is dual mode (ETH and LBC), ETH is fix 29.7, LBC is 91Mh/s. Its really good. In only ETH mode is 30.4. 
Its under test in dual mode.

Core cl.: 1130, Mem cl.: 2150, the voltage 850-850.

I check GPU-Z, the "GPU only Power Draw" is 105W in dual mode. BUT from wall 216. What does it mean, why its so hight? What do you think? Maybe my computer need so lot of volt extra?


My other question, maybe you have a solution: before the fresh Win install I received a lot of "video scheduler internal error" and blue screen. I'm looking for solution, this is vor VGA driver, I must to update it. ...but this beta driver direct for [Suspicious link removed]puting is okay for me.
What do you think?

[AS10]

Is there any difference in Hashrate of mining ETH based coins like Ubiq and musicoin ( with low size dag ) and Original ETH? ( using RX 580 )

[Call_Me_Bambi]

Is there any difference in Hashrate of mining ETH based coins like Ubiq and musicoin ( with low size dag ) and Original ETH? ( using RX 580 )

I've mined various Ethash coins and the hashrate, as far as I could tell, did not change between the different coins.
Give it a go for yourself and see what happens, it should only take a few minutes to test it out.

[biobit]

Quick question. What is the use of bios modding the card, if you can set extra launch parameters in Claymore? Don't these override the former?

[Branko]

Quick question. What is the use of bios modding the card, if you can set extra launch parameters in Claymore? Don't these override the former?

Its not the same thing...you can change frequency in Claymore, but you can't change TIMINGS

[greatauror28]

Recently setup 8x Nitro+ Rx 580 4gb with various clocks ranging from 2025 up to 2100, but only has a single core clock of 1200 across. Mvddc and cvddc are set to 875mV. Right now it is pulling ~1320w at the wall and pushing a total of 244Mh/s eth and 6.1Gh/z dcr.

All's nice and shiny BUT it gives me random openCL errors across all cards!  

Sometimes after 4 hrs GPU1 will hang, sometimes after 5 hrs GPU6 will do the same. I can't put my finger on it why it is happening but I know it is not a hardware issue as I've changed, replaced risers, usb cables, x1 connector, pcie ports, pcie 8-pin cable - the whole laundry list of things you can change but still no avail. Note that when a GPU gives out openCL error and hangs, it's only a single GPU that hangs and not two or three or more.

I'm on Day 6 now and it is still restarting.

One thing to note too is that I almost have zero, as in ZERO memory errors across all GPUs. And when I say almost, I mean one or two GPUs have about 20-30 errors in a span of 3-4hrs, so it's nothing and I know it's not enough for them to crash it. Interesting fact too is that the GPU/s that has some memory errors on them aren't the ones crashing.

Could it be the 1200 mem clock? Could it be it needs more than 875mV of voltage? I'm almost out of options here. I need help.

Thoughts?

Win10 Pro
Claymore v10
Aug23 blockchain drivers
Recently DDU 'd
Paused Win updates
4GB Ram
16gb-20gb page file

[Call_Me_Bambi]

Recently setup 8x Nitro+ Rx 580 4gb with various clocks ranging from 2025 up to 2100, but only has a single core clock of 1200 across. Mvddc and cvddc are set to 875mV. Right now it is pulling ~1320w at the wall and pushing a total of 244Mh/s eth and 6.1Gh/z dcr.

All's nice and shiny BUT it gives me random openCL errors across all cards!  

Sometimes after 4 hrs GPU1 will hang, sometimes after 5 hrs GPU6 will do the same. I can't put my finger on it why it is happening but I know it is not a hardware issue as I've changed, replaced risers, usb cables, x1 connector, pcie ports, pcie 8-pin cable - the whole laundry list of things you can change but still no avail. Note that when a GPU gives out openCL error and hangs, it's only a single GPU that hangs and not two or three or more.

I'm on Day 6 now and it is still restarting.

One thing to note too is that I almost have zero, as in ZERO memory errors across all GPUs. And when I say almost, I mean one or two GPUs have about 20-30 errors in a span of 3-4hrs, so it's nothing and I know it's not enough for them to crash it. Interesting fact too is that the GPU/s that has some memory errors on them aren't the ones crashing.

Could it be the 1200 mem clock? Could it be it needs more than 875mV of voltage? I'm almost out of options here. I need help.

Thoughts?

Win10 Pro
Claymore v10
Aug23 blockchain drivers
Recently DDU 'd
Paused Win updates
4GB Ram
16gb-20gb page file

I think you are pushing your cards too hard. I found my RX580's were very touchy about overclocking/undervolting too much (they had modded BIOS's).
I also found that after overclocking the card(s) needed to be run for several hours to check for stability.

[greatauror28]

Recently setup 8x Nitro+ Rx 580 4gb with various clocks ranging from 2025 up to 2100, but only has a single core clock of 1200 across. Mvddc and cvddc are set to 875mV. Right now it is pulling ~1320w at the wall and pushing a total of 244Mh/s eth and 6.1Gh/z dcr.

All's nice and shiny BUT it gives me random openCL errors across all cards!  

Sometimes after 4 hrs GPU1 will hang, sometimes after 5 hrs GPU6 will do the same. I can't put my finger on it why it is happening but I know it is not a hardware issue as I've changed, replaced risers, usb cables, x1 connector, pcie ports, pcie 8-pin cable - the whole laundry list of things you can change but still no avail. Note that when a GPU gives out openCL error and hangs, it's only a single GPU that hangs and not two or three or more.

I'm on Day 6 now and it is still restarting.

One thing to note too is that I almost have zero, as in ZERO memory errors across all GPUs. And when I say almost, I mean one or two GPUs have about 20-30 errors in a span of 3-4hrs, so it's nothing and I know it's not enough for them to crash it. Interesting fact too is that the GPU/s that has some memory errors on them aren't the ones crashing.

Could it be the 1200 mem clock? Could it be it needs more than 875mV of voltage? I'm almost out of options here. I need help.

Thoughts?

Win10 Pro
Claymore v10
Aug23 blockchain drivers
Recently DDU 'd
Paused Win updates
4GB Ram
16gb-20gb page file

I think you are pushing your cards too hard. I found my RX580's were very touchy about overclocking/undervolting too much (they had modded BIOS's).
I also found that after overclocking the card(s) needed to be run for several hours to check for stability.

Nah, it turns out it’s a driver issue.

Updated to win10 FCU last night then installed 17.10.3 driver then boom! issue gone.

Been running 15+ hrs straight now without no hiccups or dropping cards.

After this stability run, i’d push them to 31 again.

[DevelopmentBank]

look the new cards are basically the same as the old ones

rx 470 = rx 570
rx 480 = rx 580

So what ever you can find at a good price get.

That is true - and it will MUCH easier to sell rx 580. So transition to new series is must

I beg to differ. The older RX 470/480 series cards consume LESS POWER. This is a known fact (based on specs on boxes) and one that i myself have tested at the wall with a power meter. The additional 1-2 hash rates are nice but if electricity is a big factor in your profit (mining in places where electricity is expensive), it will be best to stick with the old cards.

Resale value is a small matter. If you're mining you should be in this for the long term. Mine until your GPUs die.