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

[Omg9500]

So... what would you choose between Samsung and Micron memory? (Sapphire nitro+ 8gb).

Problem is usually that  you don't know what memory you get as you can't  check the cards before you buy. But it does not matter in my opinion. I have 14 RX580 8GN cards with Hynix and Micron Memory and all are giving around 30 mh/s per card with modded BIOS.

I got modded VBIOS for my cards from fellow miners over at anorak.tech

[Omg9500]

Hey, new miner also. Mining for 3 weeks allready have 11 sapphire nitro special edition 8gb. Stable 30.2 MH without modding bios. Dont have power meter, but seems my cards taken lot of electricify on globalpowertune 5. Trying to set 3. Stable for few hours now. How much mh i can reach if i flash bios? I'm useing core 1250, mem 2290 on ethos

You need a powermeter to determine your total consumption on the wall. It is not expensive.

I got 13 RX580 with mobo, SSD, monitor, etc, power draw from the wall around 1700W so I suspect 1 card will be around 110W on the wall... Without modded VBios this can jump up to 180 Watt and your cards will get HOT!

[pbfarmer]

So... what would you choose between Samsung and Micron memory? (Sapphire nitro+ 8gb).

Problem is usually that  you don't know what memory you get as you can't  check the cards before you buy. But it does not matter in my opinion. I have 14 RX580 8GN cards with Hynix and Micron Memory and all are giving around 30 mh/s per card with modded BIOS.

I got modded VBIOS for my cards from fellow miners over at anorak.tech

I've seen Newegg listing the mem mfg lately.

[Bigdrago]

So... what would you choose between Samsung and Micron memory? (Sapphire nitro+ 8gb).

Problem is usually that  you don't know what memory you get as you can't  check the cards before you buy. But it does not matter in my opinion. I have 14 RX580 8GN cards with Hynix and Micron Memory and all are giving around 30 mh/s per card with modded BIOS.

I got modded VBIOS for my cards from fellow miners over at anorak.tech

Thanks.
I could choose from the supplier. Hynix was the worst.

Have u tried adjusting gpu clock to 1250? That gave me much better result. (Dual mining)

[pinamalina]

I have configured 40+ GPUs achieving 30.0-31.9 on cards with 140-150W on wall per card on ETH only mining.
I use my own approach in dealing with the wiring/bios/modding/running Claymore and setting the undervolting/underclocking values for each card individually. Each card is different so it needs to be tuned individually.

As as see many newbies challenged with the setup I am offering my assistance to anyone willing to match my results.

If you are up for a test drive, I would be willing to help you tune your system directly to a similar results to mine (see below). Done this few times using Skype/TeamViewer - works great.
Send a request via PM.

Next is my current status report from Claymore 11.0 (in total running 1.25GH/s stable):

Rig number #1: 13 cards all ASUS ROG STRIX TOP or OC Edition 8GB with Hynix and some with Micron memory.
ETH - Total Speed: 406.426 Mh/s, Total Shares: 2693(198+209+191+216+211+196+228+203+229+175+202+232+203), Rejected: 0(0+0+0+0+0+0+0+0+0+0+0+0+0), Time: 07:48
ETH: GPU0 31.390 Mh/s, GPU1 31.593 Mh/s, GPU2 30.882 Mh/s, GPU3 31.742 Mh/s, GPU4 31.531 Mh/s, GPU5 31.208 Mh/s, GPU6 31.385 Mh/s, GPU7 30.805 Mh/s, GPU8 31.374 Mh/s, GPU9 31.385 Mh/s, GPU10 31.385 Mh/s, GPU11 31.048 Mh/s, GPU12 30.697 Mh/s
Incorrect ETH shares: none
1 minute average ETH total speed: 405.764 Mh/s
GPU0 t=49C fan=0%, GPU1 t=52C fan=33%, GPU2 t=51C fan=32%, GPU3 t=52C fan=0%, GPU4 t=51C fan=0%, GPU5 t=51C fan=0%, GPU6 t=51C fan=31%, GPU7 t=51C fan=31%, GPU8 t=51C fan=0%, GPU9 t=50C fan=31%, GPU10 t=51C fan=0%, GPU11 t=53C fan=32%, GPU12 t=54C fan=27%

Rig number #2: 13 cards: 9x - ASUS ROG STRIX TOP or OC Edition 8GB with Hynix and some with Micron memory, 4x Gigabyte Nvidia 1070 8Gb (GPU9-12)
ETH - Total Speed: 399.940 Mh/s, Total Shares: 7300(564+614+570+560+563+528+555+548+516+598+544+555+585), Rejected: 1(0+1+0+0+0+0+0+0+0+0+0+0+0), Time: 21:37
ETH: GPU0 30.641 Mh/s, GPU1 30.595 Mh/s, GPU2 30.793 Mh/s, GPU3 30.613 Mh/s, GPU4 30.753 Mh/s, GPU5 30.554 Mh/s, GPU6 30.600 Mh/s, GPU7 30.649 Mh/s, GPU8 29.419 Mh/s, GPU9 31.323 Mh/s, GPU10 31.323 Mh/s, GPU11 31.354 Mh/s, GPU12 31.323 Mh/s
Incorrect ETH shares: none
1 minute average ETH total speed: 399.767 Mh/s
GPU0 t=53C fan=33%, GPU1 t=51C fan=31%, GPU2 t=50C fan=29%, GPU3 t=51C fan=31%, GPU4 t=50C fan=31%, GPU5 t=52C fan=38%, GPU6 t=51C fan=32%, GPU7 t=50C fan=30%, GPU8 t=50C fan=0%; GPU9 t=63C fan=49%, GPU10 t=66C fan=54%, GPU11 t=61C fan=45%, GPU12 t=63C fan=48%

Rig number #3: 13 cards all ASUS ROG STRIX TOP or OC Edition 8GB with Hynix and some with Micron memory.

ETH - Total Speed: 411.677 Mh/s, Total Shares: 11652(895+845+871+858+942+892+896+924+904+942+891+887+905), Rejected: 1(0+0+0+0+0+0+0+0+0+0+0+0+1), Time: 33:36
ETH: GPU0 31.817 Mh/s, GPU1 30.646 Mh/s, GPU2 31.593 Mh/s, GPU3 31.901 Mh/s, GPU4 31.941 Mh/s, GPU5 31.939 Mh/s, GPU6 30.881 Mh/s, GPU7 31.951 Mh/s, GPU8 31.960 Mh/s, GPU9 31.951 Mh/s, GPU10 31.951 Mh/s, GPU11 31.388 Mh/s, GPU12 31.757 Mh/s
Incorrect ETH shares: none
1 minute average ETH total speed: 411.193 Mh/s
GPU0 t=52C fan=33%, GPU1 t=51C fan=34%, GPU2 t=49C fan=0%, GPU3 t=53C fan=0%, GPU4 t=51C fan=0%, GPU5 t=50C fan=29%, GPU6 t=50C fan=0%, GPU7 t=50C fan=0%, GPU8 t=52C fan=0%, GPU9 t=49C fan=0%, GPU10 t=50C fan=0%, GPU11 t=54C fan=30%, GPU12 t=52C fan=30%

BTW: One of my ongoing projects is creating autotune tool to individually tune the -dcri option in Claymore. Will share results/tool when complete.

[zeef]

Hi guys,

I have an rx580 4gb gigabyte aorus with hynix memory, running on windows 10.

I made the oneclickpatch 1.6.7 two months ago.

0 Problems.

She is hashing in 860 - 880 range with this settings in afterburner:

950mv core
-10 powerlimit
coreclock 1150
memoryclock 2020mhz

Everything ok.

But today when i unintentionally click on reset functionatlity in afterburner, the card get 960hs.  

So i think let me try to increase the freq of memory from 1750mhz to 1800 etc..

But changing up the memory clock doesnt increase anything..

And if i restart the amd gpu miner 11.2 with the default setting, the hash is not so high like was before restart,

It only increase if i toggle between my setting to default.

Could be a bug in miner?  (i tested in amd gpu miner and in cast xmr) Never happend to me.

So now i start with my settings, toggle to default and get more hash.. but...doesnt make anysense

Regards

[RYXES]

Hi guys,

I have an rx580 4gb gigabyte aorus with hynix memory.

I made the oneclickpatch 1.6.7 two months ago.

0 Problems.

She is hashing in 860 - 880 range with this settings in afterburner:

950mv core
-10 powerlimit
coreclock 1150
memoryclock 2020mhz

Everything ok.

But today when i unintentionally click on reset functionatlity in afterburner, the card get 960hs.  

So i think let me try to increase the freq of memory from 1750mhz to 1800 etc..

But changing up the memory clock doesnt increase anything..

And if i restart the amd gpu miner 11.2 with the default setting, the hash is not so high like was before restart,

It only increase if i toggle between my setting to default.

Could be a bug in miner?  (i tested in amd gpu miner and in cast xmr) Never happend to me.

So now i start with my settings, toggle to default and get more hash.. but...doesnt make anysense

Regards

I prefer v10.0 personally.

Are you with Windows OS?

if so check that it has updated, if so that is probably why, its killed 90% of peoples mining rigs on Windows 10. just need to do a PC Reset to an earlier build. Should work.

[zeef]

Hi guys,

I have an rx580 4gb gigabyte aorus with hynix memory.

I made the oneclickpatch 1.6.7 two months ago.

0 Problems.

She is hashing in 860 - 880 range with this settings in afterburner:

950mv core
-10 powerlimit
coreclock 1150
memoryclock 2020mhz

Everything ok.

But today when i unintentionally click on reset functionatlity in afterburner, the card get 960hs.  

So i think let me try to increase the freq of memory from 1750mhz to 1800 etc..

But changing up the memory clock doesnt increase anything..

And if i restart the amd gpu miner 11.2 with the default setting, the hash is not so high like was before restart,

It only increase if i toggle between my setting to default.

Could be a bug in miner?  (i tested in amd gpu miner and in cast xmr) Never happend to me.

So now i start with my settings, toggle to default and get more hash.. but...doesnt make anysense

Regards

I prefer v10.0 personally.

Are you with Windows OS?

if so check that it has updated, if so that is probably why, its killed 90% of peoples mining rigs on Windows 10. just need to do a PC Reset to an earlier build. Should work.

Hi,

Windows 10,

I blocked all the automatic updates. All my rigs run fine on windows 10 and in linux.

My question is just curiosity, if someone had have this bug too

I think its corellated with overclocking only...

What are the advantages from an older version 10.0 ?

Regards

[etsRX]

Hi,

can someone explain me please how to U calculate the wattage of gpu in GPU-Z.
Like, do U have to add: GPU only Power Draw and VDDC Power Draw and VDDCI Power Draw, or what?

GPU-Z version 2.8.0

[zeef]

Hi guys,

I have an rx580 4gb gigabyte aorus with hynix memory.

I made the oneclickpatch 1.6.7 two months ago.

0 Problems.

She is hashing in 860 - 880 range with this settings in afterburner:

950mv core
-10 powerlimit
coreclock 1150
memoryclock 2020mhz

Everything ok.

But today when i unintentionally click on reset functionatlity in afterburner, the card get 960hs.  

So i think let me try to increase the freq of memory from 1750mhz to 1800 etc..

But changing up the memory clock doesnt increase anything..

And if i restart the amd gpu miner 11.2 with the default setting, the hash is not so high like was before restart,

It only increase if i toggle between my setting to default.

Could be a bug in miner?  (i tested in amd gpu miner and in cast xmr) Never happend to me.

So now i start with my settings, toggle to default and get more hash.. but...doesnt make anysense

Regards

I prefer v10.0 personally.

Are you with Windows OS?

if so check that it has updated, if so that is probably why, its killed 90% of peoples mining rigs on Windows 10. just need to do a PC Reset to an earlier build. Should work.

Hi,

Windows 10,

I blocked all the automatic updates. All my rigs run fine on windows 10 and in linux.

My question is just curiosity, if someone had have this bug too

I think its corellated with overclocking only...

What are the advantages from an older version 10.0 ?

Regards

Hi guys,

i just uploaded an video of my problem on yotube:

https://www.youtube.com/watch?v=Yj_dDd3-TZY

Regards

[Rough_R]

Hi,
I recently got few of sapphire rx 580 nitro+ 4gb Samsung memory,
moded bios with Polaris 1.67, applied UBERMIX 3.1,
moded TDP from 145 to 120W, and Max Power Limit from 175 to 140W.
Got about 29,5mh/s at ETH.
What can mod next to rise MH/s and lower W from wall ? 
in Afterburner lowered Power limit -10 and control fan.
When rise mem to +100 got about 30.5mh/s but becomes unstable.

Regards

[Bigdrago]

Hi,
I recently got few of sapphire rx 580 nitro+ 4gb Samsung memory,
moded bios with Polaris 1.67, applied UBERMIX 3.1,
moded TDP from 145 to 120W, and Max Power Limit from 175 to 140W.
Got about 29,5mh/s at ETH.
What can mod next to rise MH/s and lower W from wall ? 
in Afterburner lowered Power limit -10 and control fan.
When rise mem to +100 got about 30.5mh/s but becomes unstable.

Regards

Samsung micron?

[Rough_R]

Hi,
I recently got few of sapphire rx 580 nitro+ 4gb Samsung memory,
moded bios with Polaris 1.67, applied UBERMIX 3.1,
moded TDP from 145 to 120W, and Max Power Limit from 175 to 140W.
Got about 29,5mh/s at ETH.
What can mod next to rise MH/s and lower W from wall ? 
in Afterburner lowered Power limit -10 and control fan.
When rise mem to +100 got about 30.5mh/s but becomes unstable.

Regards

Samsung micron?

sapphire rx 580 nitro+ 4gb Samsung memory

[beatifullive]

So my Sapphire Nitro+ RX580 Limited Edition cards came in today, so I ran some benchmarks.

I first tried to find the highest error free memory clock for each memory strap and chose one that performs the best.
I followed this up by downclocking the core until the mining performance was noticeably affected if I'd gone further.
I then proceeded to reduce voltages as much as possible.
....

hello Apneal, How did you test the memory straps? MSI Afterburner?

[onionpealer]

So my Sapphire Nitro+ RX580 Limited Edition cards came in today, so I ran some benchmarks.

I first tried to find the highest error free memory clock for each memory strap and chose one that performs the best.
I followed this up by downclocking the core until the mining performance was noticeably affected if I'd gone further.
I then proceeded to reduce voltages as much as possible.
....

hello Apneal, How did you test the memory straps? MSI Afterburner?

Wooh there little dawgie ! The original post was from almost a year ago, but anyway..
I used Afterburner to quickly test Memory Clock and Voltage settings to tweak my cards, you would also need HWinfo to check for "GPU memory errors", no errors are best, lots of errors are very bad.
Some people like to use OverdriveNtool instead of Afterburner.

Link to OverdriveNtool.
https://forums.guru3d.com/threads/overdriventool-tool-for-amd-gpus.416116/

Link to HWinfo.
https://www.hwinfo.com/download.php

[big al]

From testing my msi rx580s with 8gb micron I found they consistently wanted a multiplier of 1.81 between core and memory clocks.  This worked from 1900 to 2270 on the memory. Go lower on core and you lose hash. Go higher and you don't gain anything.  Plus you're just wasting voltage.  I think the multiplier theory will work across all rx580s but it will be a different number for different memory.

Keep this in mind when fine tuning cards.  It will help save time.

Another quirk I've found is the higher of the chip and ram voltage is used for both. At least that's what HW reports. It is also about 0.025 lower than what you put in overdriventool. Or is it just my cards?

When they crash they usually reset to the stock bios core and ram speed. Keep that in mind when flashing the BIOS. But they always go to 950 mv on the ram.

[beatifullive]

So my Sapphire Nitro+ RX580 Limited Edition cards came in today, so I ran some benchmarks.

I first tried to find the highest error free memory clock for each memory strap and chose one that performs the best.
I followed this up by downclocking the core until the mining performance was noticeably affected if I'd gone further.
I then proceeded to reduce voltages as much as possible.
....

hello Apneal, How did you test the memory straps? MSI Afterburner?

Wooh there little dawgie ! The original post was from almost a year ago, but anyway..
I used Afterburner to quickly test Memory Clock and Voltage settings to tweak my cards, you would also need HWinfo to check for "GPU memory errors", no errors are best, lots of errors are very bad.
Some people like to use OverdriveNtool instead of Afterburner.

Link to OverdriveNtool.
https://forums.guru3d.com/threads/overdriventool-tool-for-amd-gpus.416116/

Link to HWinfo.
https://www.hwinfo.com/download.php

Thanks! I test it in changing the config of claymore's start file...
I am confused by the mem errors reported by HWinfo, meanwhile there is no reject from the pool.
Using 1xSapphire Rx480 8G nitro+,with uber-mix timing 3.1.
29.7m @ -cclock 1200 -cvddc 950 -mclock 2020 -mvddc 950.
runned 24hours with no reject but 1000 in total mem errors reported by HWinfo.

[RYXES]

Hi guys,

I have an rx580 4gb gigabyte aorus with hynix memory.

I made the oneclickpatch 1.6.7 two months ago.

0 Problems.

She is hashing in 860 - 880 range with this settings in afterburner:

950mv core
-10 powerlimit
coreclock 1150
memoryclock 2020mhz

Everything ok.

But today when i unintentionally click on reset functionatlity in afterburner, the card get 960hs.  

So i think let me try to increase the freq of memory from 1750mhz to 1800 etc..

But changing up the memory clock doesnt increase anything..

And if i restart the amd gpu miner 11.2 with the default setting, the hash is not so high like was before restart,

It only increase if i toggle between my setting to default.

Could be a bug in miner?  (i tested in amd gpu miner and in cast xmr) Never happend to me.

So now i start with my settings, toggle to default and get more hash.. but...doesnt make anysense

Regards

I prefer v10.0 personally.

Are you with Windows OS?

if so check that it has updated, if so that is probably why, its killed 90% of peoples mining rigs on Windows 10. just need to do a PC Reset to an earlier build. Should work.

Hi,

Windows 10,

I blocked all the automatic updates. All my rigs run fine on windows 10 and in linux.

My question is just curiosity, if someone had have this bug too

I think its corellated with overclocking only...

What are the advantages from an older version 10.0 ?

Regards

There isn't many differences in v10.0 and v11.4 for Eth mining only, except for stale shares. I find v10.0 has waay less stale shares than v11.4 for Eth mining only.

[drale]

Hi just one question;

I followed instructions here
https://1stminingrig.com/best-bios-rom-sapphire-nitro-rx-580-8gb-oc-hynix-memory-30-mhs/
so
I leaved hardware in silent BIOS and modified with V1.
Core Clock 2130 MHz
Memory Clock 2175 MHz
Fan speed 69%.

So it mine ethereum on
31.18 Mh/s.

Should I change it from silent mod BIOS to turbo?

[deksa]

Anyone using ASUS Radeon RX 580 8GB Dual-fan OC Edition GDDR5?
I got a pretty good deal for this model but the thing is, I found different opinions someone says it is decent, someone says it has a bad cooling.
Any first hand experiences here? Someone mining with such? What are the temperature, clocks, and what algo are you mining whit it? Do you also have any other model to compare the Asus with it?
Thanks!