Ethereum DAG Hashrate Drop FIX – AMD Official Mining Drivers Are Now Out
|- Like
- Digg
- Del
- Tumblr
- VKontakte
- Buffer
- Love This
- Odnoklassniki
- Meneame
- Blogger
- Amazon
- Yahoo Mail
- Gmail
- AOL
- Newsvine
- HackerNews
- Evernote
- MySpace
- Mail.ru
- Viadeo
- Line
- Comments
- Yummly
- SMS
- Viber
- Telegram
- Subscribe
- Skype
- Facebook Messenger
- Kakao
- LiveJournal
- Yammer
- Edgar
- Fintel
- Mix
- Instapaper
- Copy Link
AMD Official Mining Drivers
Apparently AMD wants to avoid shortages of their new Vega cards at any cost. Consequently, the company is doing their best to help miners to keep their older GPUs functional for mining. In order to do that, AMD issued an official driver release aimed to improve GPU mining hashrate.
AMD has been working on this driver for a while now. In fact, it became available for download about two weeks ago, though it was an unofficial beta release. Today, AMD posted the same driver on their official webpage.
Radeon Software Crimson ReLive Edition Beta for Blockchain Compute
I have tested this new driver on my rigs and I can confirm that they it does improve hashrate, no more hashrate drop, everything is back to normal. I suspect that the driver reverse the DAG epoch issue, rather than making GPUs mine faster.
I have tried the driver on RX 400 and RX 500 cards, though according to AMD webpage, the Crimson ReLive Edition Beta is compatible with a wide variety of GPUs, starting from the Radeon HD 7×00 series and up to the Vega edition.
FIRST CHECK MY TEST!
Sapphire Nitro+ RX 470 4GB 5x GPU Mining Rig
Ethereum Mining
Previous Drivers
- 27.100 Mh/s per GPU
New Beta Drivers
- 29. 5 Mh/s +2.5 Mh/sper GPU
The rig went from 135 Mh/s to 147 Mh/s on ethereum only mode.
Let’s see how it goes in dual mode
Ethereum Dual Mining Siacoin
Previous Drivers
- Ethereum: 27.400 Mh/s per GPU
- Siacoin: 460 Mh/s per GPU
New Beta Drivers
- Ethereum: 29.500 Mh/s per GPU
- Siacoin: 410 Mh/s per GPU
Ethereum Dual Mining Decred
Previous Drivers
- Ethereum: 27.500 Mh/s per GPU
- Decred: 635 Mh/s per GPU
New Beta Drivers
- Ethereum: 29.500 Mh/s per GPU
- Decred: 680 Mh/s per GPU
Yes, i can confirm that everything is now back to normal! YAY, no more DAG epoch hashrate drop :), also i see a slightly less power consumption. Note that i used the exact same clocks, dcri, undervolt with both driver versions.
Sapphire Nitro+ RX 580 4GB 4x GPU Mining Rig
Finally i was able to test the RX 500 series (this should work for rx 570 and 580) since many of you were complaining about high power usage, undervolting and high temps, i had the same problems.. no matter what i tried Beta MSI Afterburner, Trixx, WattTool.. no luck.. in the end i tested with WattMan from AMD Driver Settings, after some tweaks i was able to undervolt them to 130w – 140w on DUAL MINING. This was a quick test but i will add soon more pictures with results.
This is a 4 GPU rig and before undervolting and downclocking, power usage was around 900+w (forgot to take a picture, will add it soon).
- Clocks: 1130 / 2200
- Core voltage: 850
- Memory Voltage: 850
- GPU temps: 50c with 75% fan speed
- Ethereum Hashrate: 30.5 Mh/s
- Decred Hashrate: 900 Mh/s
- Total power usage: 650w (100w system)
- Power usage per 1 GPU: 130w – 140w
Mission accomplished! Still room for improvements 🙂 the rig is running stable with not much hashrate fluctuations.
Sapphire Nitro+ RX 580 4GB 8x GPU Mining Rig
Here is guys a 8x Sapphire Nitro+ RX 580 4GB GPUs minig rig with the same drivers and overclock / undervolt procedure via Wattman, the rig is running stable. Check screenshots!
From 1840w to 1230w
141 MH/s
How to fix Ethereum / Ethash DAG epoch hashrate drop!
1 DownloadClaymore released a new updated version fixing issues with voltage/clocks management for latest AMD blockchain drivers
- VERY IMPORTANT: Download latest claymore miner: https://mega.nz/#F!O4YA2JgD!n2b4iSHQDruEsYUvTQP5_w
- Download Beta Driver for Blockchain Compute http://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Crimson-ReLive-Edition-Beta-for-Blockchain-Compute-Release-Notes.aspx
- Download ATImdag Patcher
https://www.monitortests.com/atikmdag-patcher-1.4.5.ziphttps://www.monitortests.com/atikmdag-patcher-1.4.6.zip -
Download BETA MSI Afterburner http://forums.guru3d.com/showpost.php?p=5459908&postcount=637 - Download Display Driver Uninstaller http://www.guru3d.com/files-details/display-driver-uninstaller-download.html
- Uninstall drivers with DDU in safe mode
- Install the freshly-downloaded driver from setup
- Before restart install the atikmdag patcher
- Install Beta MSI Afterburner
- Restart and tweak your rig
VERY IMPORTANT To overclock and undervolt the GPUs you will have to use .bat file commands. For example i used this commands for Sapphire Nitro+ RX 580 8GB: -cclock 1150 -mclock 2200 -cvddc 850 -mvddc 850, you can increased -mclock to 2250 and also if system crashes lower voltages from 850 to 875 or 900.
Enjoy and happy mining! 🙂 share your results..
In Conclusion
This is indeed great news for those of you who were reluctant about installing the unofficial driver release. Now you can download it from the official AMD page and enjoy the increase in hashrate.
So, before this, at the actual DAG – #138 – your 470 4gb pulled out 27.4? Everybody complains the hashrate dropped to 25 and something, but you still had 27.4?
But it was never to 25, i mean not yet..
My Asus are at 26, 470 4gb. Of course, Asus cannot be compared with Sapphire in mining.
My Sapphire RX 470 Nitro (non-plus) 4GB (Hynix) actual hashrate: GPU0 28.232 Mh/s, GPU1 28.310 Mh/s (Amorak bios, undervolt -100mV and memory at 2050, driver 16.9.1)
any news about Linux drivers updates?
no idea :/
Does the freshly-downloaded driver need to be installed in safe mode as well?
Nope
great job …thank you
Ciprian V. could you tell me something about how its possible to undervolt from 1000mv to 860 – 880mv because of much energie usage?
yep 🙂
I would also appreciate a decent answer. I tried using AB before and I cannot undervolt more than -100. I’m guessing you’re using Watt tool, or you have the undervolt in the bios?
I think you might be able to mod the afterburner config file/ini, going to look into it over the weekend as I want to go past -100 too. For some strange reason using polaris to mod the voltage has done absolutely nothing – clearly I’m doing something wrong
Had a look in the afterburner config file, haven’t found a way of doing it yet
This may work by it haven’t tried it http://www.overclock.net/t/1543428/msi-afterburn-need-to-extend-te-maximum-voltage-offset-maximum-100-needs-to-be-163
Hey, sorry for the delay, actually WattMan is the only one who can properly undervolt the cards :/
Hey there, yeah i went ahead and experimented with Trixx and it worked for me. I posted about it probably 3 days ago in this thread too. Thanks!
I tested with Trixx and didn’t worked…
Wasn’t able to replicate your results with my Sapphire RX 580’s Nitro+. Power usage is through the roof.
Um.. it should not happen that.. hmm
What about the 7gpu moboz z170a where I had to do your combined drivers trick?
Use this instead 🙂
I’ll try it tomorrow.
I have 6 RX 580’s I bought the rig pre installed and want to run this fix but I am worried about the drivers only picking up 1/2 cards, I can’t afford too much downtime of the rig, can anyone confirm if this will update drivers for all 6 cards or if I will have to manually install the drivers on each card 1 by 1? Thanks for your advice.
It did not work for me, only 5 cards recognized. Got an increase from 28,5 Mh/s to 30,5 Mh/s though
Same problem only 5 GPU working now.
This completely destabilized my cards. Hash rate bounces from 16-27
what gpus do you have ?
4 x RX570s and 2 x RX580s, all Sapphires. Used the same voltage and clock settings that I was getting 150+ total Mh/s. I haven’t tried it on my Aorus setup yet, will probably give it a go tonight
Same in my case, with a 6 RX 580s (3 Gigabyte, 3 Sapphire) setup, on which I was getting 26.7 – 28.5 MH/s before, I was getting an extremely unstable 11 – 29.8 MH/s afterwards (funnily, only the Sapphire ones were sometimes getting 29.8 MH/s, the Gigabyte cards were never stepping above 22, even with OC, but mostly sitting at 11-12). My best hash rate was 143 after the driver update, while I’m normally getting 158-160 MH/s.
It seems like the drivers are still unstable, as I’ve also seen some comments on the Claymore thread on Bitcoin Talk. I’ve rolled back to the old combined drivers, and waiting for a new release.
Yea I might try one more time, but will likely roll back today. Hopefully the fully supported driver will come soon
compatible with clay miner ?
yes
super cool.. from 32.7 to 3.05 .. seems to be perfect 😀
Lol, this beta driver still need improvement
yes .. its stil beta status 😉 we have to wait
Updated driver on my 4x msi rx 580 armor 8gb modded bios, patched , hashrate increased from 28.6 to 30.4, power also increased from 82 to 101, temperature increase from 61 to 68 degree C, also hash not stable goes from 30.4 to 14. Settings used (-cclock 1225 -cvddc 900 -mclock 2175 -mvddc 900) with claymore miner 9.8. Also tried different settings still unstable. Anyone else to share their experience with the new driver on the 580 series ?
My screen goes blank
When you install amd driver uncheck HDMI audio. Try that and leave a feedback
Already done that, trying 5times but still mu screen goes blank. Another trick? My gpu still original setting from msi. Imnot OC or do anything in bios
I use windows 7
Did it work before or is this the 1st time you are trying? Also try plug the gpu in the closest pcie-x16 slot to the cpu on the motherboard
Yes. Already do that. Still blank
Windows 7 is the problem
Getting same, starts at 30MH/s and drops to 14MH/s. Your GPU load also hanging around 20% and mem clock jumping between 2150 and 1000MHz?
So I was powering some risers from a psu and the gpu from 2nd psu, managed to get it stable by powering risers and gpu from same psu
I’ve never managed to cvddc working, did it simply work for you or what did you do to get it to undervolt?
So now I run the rig with the steps shown above with new driver using afterburner beta 16 core voltage -96 power limit -20p memory clock 2175 -mem voltage 1250
yeah thats what I’m doing, I’ve managed to get my mem clock to 2200 and gpu down to 1000, currently undervolting by -100 but looking to go further if I can get away with it. Has reducing the voltage on your memory done much for temperatures and wattage?
Temperature 69 celsius, 500w for 3 cards measured from wall
Yes running rock solid on all three cards, just trying to under volt to -120, I’m 69 at 400w. EDIting my bios for voltage in Polaris just seems to get ignored by the cards for some strange reason..
Are you running stable with those settings?
For all my 6 amd gpus rigs in windows I used your guide with mixed drivers. Now will be ok with the new drivers for 6 gpus? Thank you
yep
is not working with 6 cards for me, one of the card does not get the driver recognized, it shows on device manager but with the ! inside the a yellow triangle
I used MSI armor RX 580. My screen goes to blank. So maybe this beta driver arnt compatible? Or something was wrong with my card?
Try again uninstalling, and connect only one GPU, then install it again.
I test on 1 gpu brother
Uninstalling using windows or DDU? I still use 1 card. My system windows 7. Is it the windows 7?
I’ve found a workaround for that – do “custom install” instead of “express install” and uncheck “hdmi audio driver”. You won’t be getting a black screen anymore.
Ic, tks brother.
Already tried ur suggestion, but its still black screen.
Also did you set the bios using pcie from onboard ?
Imnot oc or do anything. Factory standard
Windows 7 are the problem
Guys….if u have the RX580 8G than is not working good with this “beta driver”….we have a lot of discuss about this in the forums claymore.
that is really shit 🙁 after install the driver my comsum go up from 100W to 140W each card and cant undervolt it….
So u can still mining? I cant at all. My screen goes blank
Same Problem with me.
Same here, My hashrate increased from 28.6 to 30.5 but as you said consumption increased to 140+w. The issue is most probably lack of undervolting. I cant undervolt with msi ab. undervolting issue is related to heat and consumption accordingly. Digging into find a solution for undervolting :/
Why you can’t undervolt with AB? Did you read both article:
https://linustechtips.com/main/topic/89360-cant-adjust-gpu-voltage-msi-afterburner/?do=findComment&comment=1207583 and this one
https://rog.asus.com/forum/archive/index.php/t-81916.html
they are undervolted from wattman anyway. But the issue is not directly related to undervolting. I am revising what I said above.
which models do you use ?
1x Sapphire rx580 nitro+ 8gb, 4x powercolor rx580 red devil 8gb, 2x powercolor rx570 4gb
I will be testing it with rx 580 soon.
Please test is asap, my watt jumps up to 150w++
I’m using nitro+ 8gb hynix
My major issue on rx580s is that the temperature increased from 67ish celcius to constant 78 celcius which I do not feel comfy to run my gpus with. 570s are running just fine with hashrate increase as well. any idea why?
Please share your results.
I wanted to update my RIG (7 x Asus ROG-STRIX-RX580-O8G-GAMING) today but after reading the comments I am not sure.
Ive updated mine ive gone from 28.2 to 30.5, i cant undervolt so when dual mining SIA my cards went from 105w to 135w, ive managed to undervolt and overclock in trixx. Also i wasnt getting any fan control from claymore so ive set the fans in trixx, mine are running at 55C
In Trixx i have GPU Clock 1130 Mem Clock 2150 -120mV fan profile 40% at 40C climbing to 100% at 80C. In trixx this is per card.
ETH GPU0 30.556 GPU1 30.532 GPU2 30.555 GPU3 30.553 GPU4 30.034
SC 540.258 549.062 524.578 528.078 550.395
What version of trixx are u using? All of the tweaks is grayed out in my trixx
for me only works with 5 cards the other card never gets the driver recognized
My consumption went up 15% but I get 0.25 ETH more per month. Still better profit. Try undervolting in wattman or AF 4.4.0. beta 15 also worked for me
I have 2 RX 580 pulse and have the same power consumption issue as many here, I assume you do too. Can you write up another tutorial for undervolting in bios please if you have done it? I am able to undervolt both, core and memory in WattMan but can’t get it as stable as before. It does not crash, but voltage jumps 97-101. Basically, step 1 – start claymore, step 2 – set clocks in AB, step 3 – set voltage in Wattman.
AB clashes with Wattman, I found it causes instability issues. unistall AB and only use Wattman. I am testing 8 GPU RX580 8gb MSI armour OC in Windows 10 steady hashrate of 244 ~ 30.5 MHs per card with 1130w power draw.
Test dcri 5 for only mode it is just unbelievable for me
i have follow the procedure but now the overclock have no effect to te performance, i can overclock my rx 570 psi armor oc with che afterburn but the hash rate doesn’t increase can some one help me ? thanks guys
Try restarting claymore after overclock
i do but nothing and the pc crash if the overclock is over 1950 mh
Has anyone tried these driver on RX 470 4GB Sapphire? Thanx
I think the only solution right now is to modify the bios with polaris (set up voltage on right Mhz) and flash the cards. Or waiting for Claymore’s update to be possible to set it up in .bat file. I have 2 tipe of cards. 6x SapphireRX580nitro+ limited and 1x SapphireRX580nitro+ active. As we all know the limited ones are stronger and better. The limited editon cards work fine but they consume much power = 1060 mv on 1200 mhz and one piece of rx580 active works 960 mv also on 1200 mhz with same flashed bios as limited ones.
Which dual mining option do you guys prefer? ETH+SIA or ETH+DCR?
If i put DCR no change on ETH but if i config with SIA so ETH 30mhs down to 25 🙁
Its not DCR fix, its Ether ONLY FIX!!,
DCR is better now. It’s hitting 910MH/s at 1170/2050 (Elpida memory). Before the fix, I was hitting 860MH/s.
My hashrate it’s around 29.700 , but my power consume went from 800W to 1050W… Im using AB with -96 mV … Anything I can do more ?
Set powerlimit to -20%
i run with 6 MSI rx 570 i have a total off 155 m/hs steady working
iam afraid that when install update only 5 will be working (before to get 6 working was quite a husle )
If install new driver one has to do this 1 card after the other or all are going at same time?
Have an MSI RX580 Armor 8GB. Was getting 28.3MH/s. Followed the steps exactly.
When starting Claymore it starts great at 30MH/s but then eventually starts dropping to 19MH/s and then 12MH/s. AB is showing the Memclock jumping between 2150 and 1000 the whole time and the GPU load seems to stay on 16%. Not really working for me.
Updated with rx 500 series fix here :
https://1stminingrig.com/ethereum-dag-hashrate-drop-fix-amd-official-mining-drivers-are-now-out/#Sapphire_Nitro_RX_580_4GB_4x_GPU_Mining_Rig
In claymore 9.8 dont use -cclock -mclock etc. use afterburner set clocks then user -20% power limit.
Hi all!
After testing all day Im found one solution for undervolting with new beta driver which one works:) Maybe its a little bit primitive but working for sure for those who has rx5xx series cards. First I share my hardware. I have 7 pieces Sapphire rx580 Nitro+ 8gb limited edition in one rig. Here are what Im do:
– I have moded bios that one that Cirpian V. share with us in his blog.
– First Im remove the old one driver from the system with DDU in safe mode.
– Than I put the new one beta driver http://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Crimson-ReLive-Edition-Beta-for-Blockchain-Compute-Release-Notes.aspx
– Than Im install the new beta 4.4.0. 16 AB.
As many people than I have problem with undervolting. All my cards consume much more energie than before (1030-1080 mv / card with settings cc1200 mc2225) and I have also problem with heating. After all that Im do next:
– Im open WattMan in AMD’s software than I click on Gaming / Global Settings / Global WattMan and chose the first card. Than Im going to change to MANUAL on voltage control and than APPLY in right top corner of the screen. Than again change it to AUTOMATIC and than apply. To do that is really Important. Than you can check in AB the voltage low down to 830-840mv. You have to do this with every single card. Also important to do this all when claymore was runing!
The only problem with this that if you shut down claymore ,or you restart your rig WattMat forget all setting and you have to repeat all theese setup. Try it, it WORK’s.
If you have any questions with my setting do not hesitate to ask. If you try it and it’s work than buy me a beer :)))))
ETH:
0x82e50e3fd23936e3f7191da46c752b9c2e23a208
https://uploads.disquscdn.com/images/9ecc67edb27aa9f06ee5f15180c48d43fb3aa3fa89b170d242c557e5683b99c7.jpg
I can confirm it works! I’ll send a tip 🙂
Like I’ve posted before, Afterburner Beta 16 seems to undervolt ONLY one card in the rig, all the other cards draw a full power, I hope there will be a new Afterburner version soon that will fix this
Hi,
and your are mining with the 7 GPU ?
Because for me impossible to use more than 5.
If 6 one GPU go down to 0.000mh/s and crash the system…
Yes I’am. I have MSI Z170A pro carbon motherboard. Could you send a screenshoot of your AB?
I don’t use AB. Just flash Vbios and install driver + ATImdag Patcher + Enable 4G
With 5Gpu perfect.
With 6 or more crash and freeze
I try several risers and several RX580 GPU : Nitro+, ARMOR, Pulse. Same problem.
I’m using Z270A-Pro
install the pixel patcher and enable 4g in your mobo bios
do you speak about ATImdag Patcher ?
If yes i done. 4G too… 🙁
on a 6 gpu rig (h110) mobo i built
I installed the beta drivers (with 6 cards installed it would only see 5, until) you patch and restart n enable 4g in bios
(or get black screen on bootup)
Do you see 6 cards in device manager ?
Yes the 6 cards are in device manager with no problem.
Just on is crashing to 0.000mhs
1. did you try changing the riser ?
2. do all cards show same driver version is GPUz ?
3. what bios are you using ?
4. is that card crashing at 1100 & 2100 mem ?
5. what mobo / gpus are you using ?
6. what os ? win10 win7 ?
when my gpu crashed to 00.00 its bcz i undervolted too much (with high cclock )
i close the miner wait for wattman to recover and relaunch
So it was a riser 🙂
Thanks a lot !
Maybe you can help me one more time 🙂
I have 3x MSI ARMOR 8Go. 2 are working well, 1 is beginning with 30,1, next 28, 27, 26,24,… until 0 and crash.
Same BIOS as the others.
I test directly on a motherbord so no problem of riser, alim…
Any idee ?
reflash
flash with original bios and try if it makes the same prob
reflash
Yes i Fash with origin Bios, try another bios…
Maybe i flash 10 or 12 time but same problem 🙁
when you flashed original bios it does the same thing ?
No. With original VBIOS it’s OK.
But with 2 others vbios that are working on another ARMOR it’s not working.
if it does so on different risers / pcie slots
could be that there s an electrical loss
or vrm/chips not being cooled properly
or fail to write rom properly on this card ?
or bad silicon
imho
if rma is out of question
sell to gamer at reduced price ?
if you find the solution let me know
I’ve followed all the instructions very carefully and all my 3 rigs (RX570 and two RX580, all 7 GPU’s) draw about 200W more per rig. Afterburner does work in a way that it reduces the power draw but still I can’t come nowhere near as on the old drivers. I think the cause of it is that Afterburner Beta 16, at least in my case, doesn’t undervolt all 7 cards in a rig.
I’ve looked int GPU-Z and yes, first card have lower mV and responds to whatever settings I use in Afterburner, but another 6 cards draw a full power from my PSU hence overall I’m 200W worse than before.
Seems like we need another way of undervolting it, Claymore doesn’t do the job for me and rig is very unstable.
Updated with rx 500 series fix: https://1stminingrig.com/ethereum-dag-hashrate-drop-fix-amd-official-mining-drivers-are-now-out/#Sapphire_Nitro_RX_580_4GB_4x_GPU_Mining_Rig
To everyone having difficulties to undervolt RX 500s check my new test with solution here: https://1stminingrig.com/ethereum-dag-hashrate-drop-fix-amd-official-mining-drivers-are-now-out/#Sapphire_Nitro_RX_580_4GB_4x_GPU_Mining_Rig
Looks like your power usage jumps around and is not stable. In my case, this causes claymore to restart every few hours and reset the settings, once again causing gpus to use more power and run hot.
Actually is not, it is very stable.. i guess that is the gpu with display attached, the rig is running stable for about 6 hours now, stable hashrate and power usage. check the screenshot i just took, the power went down right after i connected via teamviewer to check it. but before that you can see an almost straight line 🙂 https://uploads.disquscdn.com/images/ec8f61883188836dacb382b494d18b696d02eaeda1952d0ad79e650817689135.png
Ok, thanks. I will try tweaking mine a bit more, not sure why it restarts claymore and resets to stock settings every few hours yet. I really don’t feel like flashing bios with undervolt settings. I have my core clock at 1200 and memory at 2250, wonder if I get a different result at 1130/2200.
Figured it out, I have a monitor hooked up to one of the cards. When it turns off to power save, claymore resets the settings for some reason and power consumption shoots to 140-150W per card, temperature to over 75C. Disabled power save and now running stable at 31+ Mh per card at 1200/2150, 860mv gpu/850mv memory with 98-99W per gpu.
Interesting. I use 6x Rx580 Dual OC Asus cards and will definitely give this a try tomorrow. But i’m wondering, are you using custom straps or you just did the 1750>2000 strap? Reason is that I’m using a custom one and I might not be able to boost my mem clock all the way to 2200 like in this settings. My settings on the cards are -cclock 1260 -mclock 2050 and 960mV both core and mem voltages.
Hello asian_dapper. I also use 6x RX580 DUAL OC Asus 8GB. can you find me on skype – ivaylo_panev so we can share some thoughts?
Hi ivaylo. I’m not big on skype, you can email me instead @ [email protected]. Sure, we can compare notes!
Hi Ciprian,
Firstly thanks for the guides. Very helpful indeed!
I just wanted to confirm to you that I can run your settings on my Sapphire Nitro+ RX 570 8GB OC cards. So your 580 settings seem to be fine for 570 as well.
It bumped up my MH/S from ~24 to ~28 immediately. I have not done any BIOS updates either so this is fine for now.
The key point is to hard wire the options you mentioned in the last part of this article. This ensures that whenever claymore reboots, it will use the same settings.
The only diff is I am using -cvddc and -mvddc of 875 rather than 850. I
am also setting the fans at ~50% which keeps the cards down to a cool
temp of 50c (using wattman).
I must also tell you that these settings let my cards run at an awesome ~80w according to GPU-Z (my watt meter is on the way so I cant confirm what it is at the wall).
Additionally, I havent tried any other clock settings yet, perhaps I can get some more juice out of them with some other combi.
A friend mentioned to me that under powering can damage the cards so any pointers u have seen on how u get whats a good number or its just trial and error? Thanks in advance.
“after some tweaks”
Could you explain these tweaks a little bit? 🙂
Check the images 🙂 what needs to be set via WattMan
Can confirm that this works. However, I noticed something strange, after I unplugged the display cable from one of the cards (while mining). The power usage jumped from 100w to 120w on GPU-z. I fixed it by restarting the rig (without the display cable plugged) and reconnected through teamviewer and started the miner. Everything worked after.
Now to figure out why 1 card is at 120w and the others at 100w…
It did not help. I have set just like you did on wattman. Running eth only at 147w and 77c. My cards are powercolor rx580 8gb. I have set it as you did and made no difference at all
btw I have reverted back to old stable driver 17.7.1 which was running below 70c degrees and 28.5mh each before. I have deleted the latest blockchain driver reverted back to it and heat is still up. What do you guys think happened?
ir not work on my sapphire rx 580 nitro + 8gb, it always hang when i try to indervolt on wattman, it wor for 10 minutes and hang again, i hope you can explain the step to video or more detail, thanks before
its work on my rx 580 pulse
I’m testing this Beta drivers, I don’t see any major improvements on the client side, adjusted to 1150 @ 850mv core, 2250 @ 900mv. -10% power limit. Getting 28 Mh/s ETH, 920 Mh/s DCR. Sapphire Rx580 8GB Nitro +.
All tuned with Wattman. Wonder how you got 30, which dcri are you using inside claymore?
I use -dcri 30. You can use the + and – keys while mining to adjust. Wait for about a minute or so each adjustment, since it will take time to stabilize hashrate each time. Each card, even in same models, can give better or worse eth hashrate at different dcri values, and with epoch changes in my experience. It also changes if you change your secondary coin. My decred performance is around 100-150 Mh/s less than yours per card on average.
Hello All,
Thx Ciprian V for your mod
I’ve been doing some tests and have my rig stable for a few days now
I got a Nitro+ Special (11265-21-20G)
Running your => Sapphire.RX580.8192._(ORIGINAL_1411)_modded_beta_256kb.rom
(got it from comments in LE bios page)
MSI AB settings (beta 16) + new amd beta drivers(*)
(To be known as profile 2)
-96cv
+15%
1296cc
1250mc
31.85eth + 1020 dcr (dcri 32)
Hottest GPU at 56 deg / coolest at 44 deg
113-115 Watts per GPU (in AB & GPUz)(No reading at the wall yet)
(reset – profile 1)
Save a profile in AB to
-96cv
0
1100
2100
this profile with above bios should consume less than 110w
(while mining eth+dcr@dcri32)
(giving me 28.9 eth + 924 dcr @ 105-108w)
————-Monitor your watt intake——————-
If one of your GPUs goes to 170w change to preset “profile 1” for a few seconds then go back “profile 2”
(*) After installing Crimson beta drivers, make sure you patch after install.
(atikmdag-patcher-1.4.5)
I also went to
Device manager => update 580 drivers => let me choose drivers Drivers amd betaDisplayWT6A_INF choose “C7316522”
Driver date shows 2017-07-27
Driver version shows 22.19.659.1
Side notes
– disable monitor sleep,
– don’t have browser or any other window open
Any ideas to make it better?
plz test and share your results ( i used claymore 9.8 )
spent a few night tuning 🙂 tips welcome
BTC 1Mz6D4D6voksrLe6qmZ9JLrJ2jtqQKX9M9
ETH 0x621d0aed1630e7d1a57096891b1741f870637ea1
Hi Ciprian V. Really great tutorials 🙂
I was wondering if you had a good bios mod and OC settings for Sapphire RX 570 8GB Hynix, Because i can’t get them running great :/ Thanks alot 🙂
Hi Ciprian V, It is working with my 580 rig thanks ! 😀 But I am having trouble findind a good bios for my Sapphire Rx570 8GB Hynix :/ Can you help me please?
post your memory model and i might be able to help.
@ciprianpt:disqus
I’m using the same tuning as u, 1130/2200 , 850 volts, around 90w/gpu @ AB,
stable at ETH single mine but not at dual mine ETH &SIA, wattage just shot up to 120w/gpu and crashed also, can u help me on this? i’m using rx580 nitro+ 8gb, hynix ram
https://uploads.disquscdn.com/images/310e2b65ad1b419f70c9f191a92691b9968ab69f9f8d1d2805e19ba1481c7c38.png
Issit because i’m not using the correct rom?
Well, I can’t undervolt for very long. Wattman keeps losing the undervolt. How are you getting it to stay? Until Claymore updates his miner, this is looking like an unstable no-go. Zero value in mining at 140watts gpu-only when I was getting 89watts with the regular drives and claymore’s settings. I bet half the momos excited about this have no idea about power draw and undervolt. If anyone has stable undervolt for multiple days, please post, and outline how that was accomplished. Your post just says to use Wattman. Not enough. Doesn’t stick. Thanks. Sapphire Nitro+ 570/580 cards.
I did undervolt and overclock in polaris bios because wattman keeps loosing it. Bios works fine
I don’t see how that’s possible. PolarisBiosEditor (Jaschaknack) doesn’t display GPU voltages properly. You only get 300Mhz/750mV, then it’s all unreadable for the other states (600/65282, 900/65283, 1145/65284, 1215/65285, 1257/65286, 1300/65287, 1411/65288). Mem looks like it might be easier since it’s apparently still in integer (300/1000, 1000/1000, 1750/1000). I tried setting all GPU mV states to the lowest 65282 but it crashed as soon as I started mining. I didn’t even bother just dropping in my mV value of 950, assuming it won’t work. Did you find a trick for converting those 65282..3…4 values into mV? Maybe using Jaschaknack just for the mem strap mod then opening that bios in a standard Polaris editor? I also tried lowering TDP (W) and TDC (A) in the Powertune section but it made zero difference.
From 900 do all same 65283 then mem 850 for all of them. My clock are 1200/2000 so what I’ve done after 900 I changed all of them to 1200-65283 then mem I left 300-850mv and 2000-850mv then all you need to do in after burner is fan speeds. Just replace numbers with your clocks and that’s it. Post if it works for you
Dude, thank you! I got it working on one card that I’m using as a test case to see if it’ll stay stable. But just having the ability to set the clocks without having to configure them in AB is nice, and it appears to be the same TDB now that I had with claymore settings. The only downside I can see is my full rigs are very much tuned per card; some are 890mV, others are 940mV for stability, and if my only option is “65283” — whatever that is — it may not be the most efficient. But it does appear to work. I half wonder if that’s the value that appears in Wattman for the 900 Mhz state. I forget what it is, but at least it would be some indication. Maybe 2…3…4 is the minimum step in mV, like 50mV or something. Either way, this worked — so huge thanks. No other solution came close.
Just check in wattman what’s say there for 900MHz because 65283 it’s mv but rigid. Find what it says for 900MHz then put the numbers in Mem. Or try undervolt in afterburner aswell it did work for me aswell. No worries 🙂
You need to create a profile in wattman for the etherminer exe and create the wattman settings in there. If you do this under global settings when one of the cards crash, it will reset back to default.
@ciprianpt:disqus
below is my set-up and I was wondering if I could use 1200w under 7GPU, or should I add another PSU to support the wattage.
Appreciate your guide on this Grin
CORSAIR – 1200w
MSI M5 7 PCIE
4x – RX 570
3x – RX 580
MSI AB 4.4.0 beta 16
Hello, i have put 850 mV in Wattman, but GPU-z shows 0.9500 mV.
What i put to the settings is not reflected to cards. Also, AB is not able to make any change to voltage even if i have put fixes for voltage change.
Have you done anything else to the cards?
i have this strong fluctuation on ethermine, is this normal or if not how can i fix it? https://uploads.disquscdn.com/images/7a4e917905eb7cddd611e9eb8070bbff986fa6b2774bfce88bca05967f414a95.png
everything installs fine, however rigs freeze after couple of minutes. no clue as power draw is fine and no errors in logs. anyone experiencing the same issue?
all 3 systems just straight up freeze after a random amount of minutes
maybe because too much OC
ive ran with these settings for 3 months so it should be fine. I’ve managed to get them working after trying things like keeping the dummy in and running starting the system with it, uninstalling afterburner and trying watt man oc. its been working for a while now but it feels that it’s going to drop sooner or later
after install driver… Pixel Fillrate, Texture Fillrate and Bandwidth going to N/A, Memory and GPU Clock to 0, Is it normal?
have you applied the patch ?
Yes sir, but GPU & Memory still 0 in GPU-Z and Pixel, Texture Bandwidth N/A, my GPU is Sapphire Nitro+ RX 580 8GB
I have run into this same problem as well..
Running from 48h with no problem with this mining driver. But with only 5 GPU. If 6 one is running 0.000mhs and crash system.
https://uploads.disquscdn.com/images/f84207bbec4405bb553a532756faae0b56e7581ddd551b727a5b1a0b12fcc8f9.jpg
On my Sapphire RX480 8gb reference I am getting 29.1mh (previously 27.5) on these drivers. My settings are (-cclock 1000 -cvddc 900 -mclock 2200 -mvddc 960). When I mine 3 together it falls to 28mh – runs at 68 degrees c. GPU-Z is saying each card is drawing 74w, total draw from the wall is 400w.
ETC + DCR
RX 480 8GB Nitro (Chipset Samsung)
https://uploads.disquscdn.com/images/c2eff42502331f8e59aabfdcac9d011fc7094082f07fc423d5064eed096e50ed.jpg
I connect to pc thru teamviewer, the resolution stucks and only showing at 640×480 Anyone knows how to fix this? https://uploads.disquscdn.com/images/64681a5d1264da5184bb5158d6d8eb669af6c06e936216901cd9a1ebcb9ef39b.png
Connect a Ghost HDMI plug on 1 card and restart. If you dont have ghost HDMI adapter then plug a monitor restart and then unplug the monitor.
Thanks alot for your advice!
Hi, if you are waiting for the new drivers on Linux (like many of us) please come by this thread https://community.amd.com/thread/219309 and drop a message..
Hi, I have problem when first start Claymore9.8, very good hashrate (RX580), 30.5 eth, ~900 dcr, but then in AB I see that the core is very “jumpy” (fluctuates between 1000-2150) and sometimes mem too. This cause the speeds to drop significantly. I notice my Asus Strix do not have this problem, only MSI Gaming X. I have to lower dcri to 12, only then the MSI core will be stable. But DCR is only 200 now. After few hours of running, ETH also drop down to 27. Do you have cclock, cvddc, mclock set in .bat? Ty.
Check your temperatures, they could be throttling clock speeds because your cards are getting too hot. You would have to increase case / environment
airflow to the card fans, make sure the air temperature is not hot but cool, and run the fans at 100%. You could also install a watercooler to the cards.
Great stuff. However, after installing these drivers, my rig crashed after three minutes of mining with the famous “thread stuck in device driver” error. I reinstalled every gpu (5x MSI rx470 8gb) seperately but to no avail. I have no idea what causes this. My virtual memory is 16Gb and I am running Claymore 9.3 on Windows 10. Any idea what could be the problem? And better yet, any idea what could be the solution? I mean, it was a great three minutes, but it would be nice to run this heap of crap for at least a few hours and I am not in the mood to go back to Ubuntu because that causes me even more problems.
Whoops. Spoke too soon. After half a dozen reboots and some Windows adjustments, things seem to be working pretty solid. Have not got around to modding but even at stock settings, things are looking pretty good. Especially fanspeeds and temperatures are way better than before.
What windows adjusment do u use? Mu screen goes blank when i installed this driver
I used the tweaking tutorial on this site. It turns off things like update and other unnecessary apps that could be problematic.
However, I cannot seem to get rid of the occasional crash with the aforementioned error. Uptime varies from 15 hours (yesterday) down to 20 minutes (4 times already today). I have no idea what the deal is with the blank screen, though. I found either the mobo or windows (or both) don’t like working without a monitor or a dummy plug but I have not encountered anything like the problem you are having. I wish I could be of help. I suppose uninstalling these drivers and installing a regular version of AMD drivers could help, but I’m not sure.
I did encounter a new problem, though. The whole rig freezes while mining. No errors in Claymore. No errors in Windows. It just freezes. This has happened twice in the last 40 minutes. Kinda annoying when it does that without any kind of error message.
I have 6 Sapphire nitro+ RX580 Special edition and I got the following results:
With the new amd beta driver I got to 185MH/s from 172MH/s eth only mining. My setting are 1130 core clock, 2210 memory clock, 850 voltages for both memory and core 880 watts power consumption. I used the AMD wattman for tweaking.
Would you mind sharing your Special Edition bios? Maybe link it with Google Drive, Dropbox or something? Thanks!
I used the one from here https://1stminingrig.com/amd-sapphire-nitro-rx-580-8gb-limited-edition-mining-rig-tutorial-190-mhs/ by the way I have a strange issue when a monitor is connected to one of the cards its hashrate drops to 26 from time to time when the cable is unplugged there are no drops.
Cards are unstable due driver, power spikes, mem up and down, ignore settings on WattMan, claymore wont able to change voltage, GPU-Z not working, Trixx fail completely, hash drop in half and goes back to original, this driver its VERY BAD, i know its BETA, but inject more problems than solution, i will switch back, and wait to a proper solution!
is not bad at all, you did something wrong. as you can see from my tests works great. today I installed the driver on a 8 gpu rig, and again works perfect. I’ll add the photos in the next hours.
Hi Ciprian , can you share the bios rom that youre using on the nitro + card w/ hynix ram. ty
Wattman did not work for me and it takes way too long to update the cards when tweaking. I ended up getting things to work in MSI afterburner. It is actually a better driver than 17.7.1. I am able to get back the lost Mhs and a bit more. I’m running around at least 30.5 and on most cards 31+. In GPUz it shows 110-120W per card. The driver just seems to have changed the way we used to OC this card. I’m going to post a short video on YouTube to help anyone else that was running into my same issue.
I just updated from 17.7.1 and I did get a slight increase on my 8x RX580 rig from 233 to 235 total Mhs but the power consumption went up 200 watts so I’m actually making almost $15 less with slightly more hash. I think I’m going to roll back unless someone has figured out this new driver. Before I was using MSI After burner with modded bios as found on the other forum on this site for RX 580 8GB Nitro+ with Hynix memory. I downloaded MSI afterburner 440 beta 16 with TRSS 700 beta 27. I’m not seeing the benefit it switching. A month ago on 17.7.1 I was getting about 30.5 Mhs per card but currently 29.5. I wanted to get that extra mhs back but this driver causes too much consumption
Have you checked my last test with rx 580 ? https://1stminingrig.com/ethereum-dag-hashrate-drop-fix-amd-official-mining-drivers-are-now-out/#Sapphire_Nitro_RX_580_4GB_4x_GPU_Mining_Rig
This definitely works. Does anybody know if there’s a fix for Lenix or right people running ethos??
Does anyone facing this problem? “Got incorrect temperature 511, ignore”
btw i am running rx 580 pulse 8gb hynix memory X 6.
Thanks for the post.
I have 5×580 Nitro+ LE and x2 580 Nitro+ on one rig. I entered the setting on Global Wattman and I get 1750 watts from the wall, if I lower the memory clock to 2100 I get 208MH/s @ 1650 watts.
Before the AMD mining driver I could get it to 193MH/s @ 1350 watts.
Not sure what I’m doing wrong. I have another rig with 7x ASUS 580 Strix OC running at 200MH/s drawing 1050 watts from the wall which I am very happy with.
Nice info definitely will try this asap
I had the memory clock floating between 2050 and 1000MHz when installed the driver. I did a biosmod to tricky the float, making p0 and p1 fixed at 2050. I undervolted the memory controller by 100mv.
After it, I had to make undervolt by BIOS, because AB only undervolts 100mv max, and the GPU accepts much more (around 250~300mv). All is working fine here after the workarounds. 30.4MH/s ETH + 910MH/s DCR at 100W on my Pulse 580 4GB Elpida.
how did you do this a am living same issue now with my rx 570 and rx 580 if use afterburner beta they only drop hashrate but if enter oc comand to claymore like -mclock 2100 they float between 2100 – 1000 – 300 sometimes
Hello everyone, I have a problem, i build a rig with this setup
MB: MSI Z270-A Pro
SSD: 60 GB
RAM: 4GB DDR4 Crucial
VGA 6xSapphire Radeon NITRO+ RX 580 8GB GDDR5 (Especial Edition)
SO: Win10 Pro
PSU: EVGA SuperNOVA 1000 G2 80+ GOLD
I Start mining ETH But the average Hashrate is 22H/s, When I understand that it should be 25H/s, I did not modify the BIOS, or OC, or anything like that, They are standard, this is normal? Or should I do something?
thanks
maybe you have something that uses gpu resources like animated graphics in the background?
Hello my friends Thank you for answering,
I only have the Windows background Basic black with blue
Since the BULK of cryptocoin GPU mining is done on a LINUX based OS, this driver is largely worthless at this point.
AMD dropped the ball BADLY by not making this available for LINUX first or at least making it available for LINUX as the same time as they released a WIndows version.
AMD Blockchain drivers are finally available for Linux:
https://steemit.com/ethereum/@walczakit/amd-blockchain-drivers-are-finally-available-for-linux
Thanks.. Its work for me.. Just following tutorial above. My card 3 Sapphire RX 580 Nitro+ 8Gb and 3 Powercolor Red Devil RX 580 8Gb.
Clocks: 1150 / 2150 (set on Global Wattman)
Core voltage: 975 (set on Global Wattman)
Memory Voltage: 975 (set on Global Wattman)
GPU temps: 53c – 70c
Ethereum Hashrate: 29.8 – 30.5 Mh/s
Decred Hashrate: 890 – 900 Mh/s
Total power usage: 980 – 990 watt
Power usage per 1 GPU: Sapphire 97-100w – Polor 120-128w
The problem is when you shut down claymore, or you restart your rig, WattMan will forget all setting and you have to repeat all setup.
https://uploads.disquscdn.com/images/6eadee122ac3903006e8892e7500383d5387fbd76151ce5ea92c8ea6593725ff.jpg https://uploads.disquscdn.com/images/252912443d0178d105d11d63cd55006214e665bde45eff968d3ab52fe2a08d1c.png https://uploads.disquscdn.com/images/7230f2e9decced928597385aabbd94b36df328614db3cb4d73a48dcd3d19149d.png
not really. you can actually create a profile for your miner in wattman so the settings will only go for that program specifically. in wattman choose games. on the top right choose add then search. now go to where you have your miner installed and select EthDcrMiner64.exe
now still in wattman you choose EthDcrminer64 and you put all the settings as you want. now open the miner. your settings will be applied only when you start the miner. and if you change some settings you must restart the miner for them to update
Thanks for your advise and I’ll try it..;)
I have one other problem…
It seems wattman can not store the settings and when computer restarts core clocks jump bask to 1260 and cards can not handle that and windows freezes…
Please help!!!
MSI AFTERBURNER dont work, just fan control.
For me settings are the same on restart..
This driver dont work on amd radeon r9 290,why?
Even I couldn’t get it to work on R9. Windows kept on restarting. Had any luck on R9?
https://uploads.disquscdn.com/images/6c301333c63124b7639391725067bb9cd5f75d0ef65a792767b7b5910c657c8d.png
Successfully updated to this blockchain driver and Claymore 9.8.
Contrary to some reviews, I was able to use Trixx to undervolt my cards, as embedding them even in BIOS doesn’t work.
Using 6x Asus Dual OC Rx 580 8GB cards.
Cool.. to be honest id didn’t bother to use Trixx at all.. but i will try with the 8x gpu rig.. thanks!
It’s very tricky as when you launch Trixx, it won’t show the clocks right away. You have to click “reset” button first then it’ll show. Then you’ll be able to undervolt, but changing clock settings still doesn’t work.
What is your -dcri. I do not understand how they are getting 31 mh/s eth + 900mh/s dcri with their saphire’s. Did you try to get that much hashes.. as I am using the same Asus dual OC 8GB cards and I am getting a little frustrated.
Hey, i am doing eth+dcr now and getting 29.3 eth and 733 dcr @ 118w gpu-z power draw. I think i have -dcri set at 25. I have to do trial and error again to bring down my power usage.
what about the -mclock and -cclock? can you send me the tunings you have and the program you use to tune your cards.
Does this driver works with modded BIOS??
Yes of course..
What is the max stable hashrate you re getting on a 580?
i tested this driver last night but it made my rx 570 4gs and 580 8gs non stable hasrate sometimes claymore shows my per and total hasrate like this did anyone have this issue or is it normal ? This is a test machine there are 3 msi gamingx rx570 4g Hynix memory and one msi gamingx rx 580 8gb Samsung memory cards on it and they are all strap level bios moded. I used the afterburner beta to overclock them 2000 to 570s 2180 to 580. Machine is stable but reported hashrate is not as seen on pictures ı’m not sure whats going on so want to ask if anyone had this. Thnx&Regards https://uploads.disquscdn.com/images/6a70bcf29f70bac4efa4df7168a9cf857a5d68bd9848ad967d77fac45292c5fd.png https://uploads.disquscdn.com/images/e019c15f6bcadbce5280201d2d4f446f53f840def49e21e1fa21cf4d0eb5f9ae.png
hi there. and thanks for this video. it helped me a lot already. i made some modifications on my bios just to tweak it and take the power consumption down. I have done it already in wattman so i know this will be my sweet spot i am getting 31.200MH/s at 1175mhz core and 2240mhz memory it’s consuming 94w at (CPUID and GPU-z). just please give it a ckeck and tell me if i’m in the right way please. Before i flash it. i don’t have another one if this card fail on me. the reason that i want to do this in bios is because i am buying more cards and if they are different i’ll need to do it anyway.as the wattman make the settings in general and not independently by card.so here’s my bios https://drive.google.com/drive/folders/0Bx3nX0h7azHEUWNxNzV3TFI4c1U?usp=sharing
careful the original bios is with the original timings (for the curious ones) you get 22.300MH/s
i have another bios only with the updated timings that give you 27.500Mh/s
and with the new drivers you’ll get up to the 31.200 after tweaks. any doubts just ask.
anyway see if i’ doing right in changing the max core frequency, and memory voltage to 950mV(not done yet)
So anyone can analyse the modded bios to see if I’m in the right direction?
After installing the new beta drivers, I can run claymore miner 9.8 for only a few minutes before my mining rig automatically restarts. Does anyone else have this problem? I have 6 XFX rx580 8gb gtr cards.
looks like one gpu is crashing ? maybe lowering oc for that one will fix it
https://uploads.disquscdn.com/images/8660e7242aef588b1a40bf524e53acbd32ab974de5a0a6846ca5cf501e039353.png
you can reduce power draw by lowering core clock to 1150 🙂 looks good
If i could push 40mhz eth then i dont mind the cards sucking 200w.
Thank you and Ti for the bios provided.
can you make a new(improved) 580 8gb hynix special ED bios.
would send u some ether
Hello , its possible share your bios ?
Tanks , good joob!
Hello All,
Thx Ciprian V for these & Ti for your Bios mod
I’ve been doing some tests and have my rig stable for a few days now
I got a Nitro+ Special (11265-21-20G)
Running your => Sapphire.RX580.8192._(ORIGINAL_1411)_modded_beta_256kb.rom
(got it from comments in LE bios page)
MSI AB settings (beta 16) + new amd beta drivers(*)
(To be known as profile
2 3
-96cv -90cv
+15% +15
1296cc 1298
1250mc 1250
31.85eth + 1020 dcr (dcri 32)
Hottest GPU at 56 deg / coolest at 44 deg
113-115 Watts per GPU (in AB & GPUz)(No reading at the wall yet)
(reset – profile 1)
Save a profile in AB to
-96cv
0
1100
2100
this profile with above bios should consume less than 110w
(while mining eth+dcr@dcri32)
(giving me 28.9 eth + 924 dcr @ 105-108w)
————-Monitor your watt intake——————-
If one of your GPUs goes to 170w change to preset “profile 1” for a few seconds then go back “profile 2”
(*) After installing Crimson beta drivers, make sure you patch after install.
(atikmdag-patcher-1.4.5)
Side notes
– disable monitor sleep,
– don’t have browser or any other window open
Any ideas to make it better?
plz test and share your results ( i used claymore 9.8 )
spent a few night tuning 🙂 tips welcome
BTC 1Mz6D4D6voksrLe6qmZ9JLrJ2jtqQKX9M9
ETH 0x621d0aed1630e7d1a57096891b1741f870637ea1
After installing this driver, after restarting the rig monitor shows the black screen.
Everytime I change voltage on wattman my gpu crash and restart the default setting , what can I do ?
Lower oc..
Hi everyone. I installed new beta driver. I think gpu0 not stable. My gpu0 card is Msi. Normally getting 28 or 29 mh but new getting 28, 26, 24, 28, 25, 28. always changing but other gpu 2 or gpu 1 cards are stable 28 or 29. Any idea?
Hi,
I have faced issues after upgrading/updating to the blockchain driver. I am suspecting 1 of my gpus might be bricked.
I have 4x identical powercolor rx 580 8gbs, they were running around 64C+- degrees. After the blockchain driver one of the gpus started functioning unstable.
After trying many things I have realized that power draw of this card is lowering down to 80w sometimes 40w and back to 110w as it should be. so the graph is up and down all the way. Normally gpu only power draw for each card is 110w, as the others run and draw constantly.
So there is something wrong with this cards, heat is higher than the others, power draw is unstable and it causes watchdog gpu hang error.
I have reverted to older driver as I used to run with but no change. I am afraid it might be a hardware issue.
What do you guys think might be the issue with this card?
Make sure you don’t undervolt too much.. that can also cause that high up and downs..
They are not too much undervolted. 4 cards are identical and only one of them is unstable :/
what bios did you used?
gpu cclock: 1200 memoryclock: 2150 tdp(w) 115
they are all set in bios so I do not undervolt from wattman,ab or such.
First off thanks for the article! Doing this now.
Also your Decred mining speed is way too low. I have 470 4GB cards pushing 1200MH/s and my 580 4GB crads do 1300MH/s while dual mining Decred and Ethereum.
what bios are you using for your rx 580 4gb cards? Also , what settings in claymore?
Dodgy drivers in my opinion. Got them running at the same hash rate as before.
The RX580 Nitro go 30.2 MHs each. thats +1,7 MHs per cards more. Nice.
though one card RX580 Armor 8GB OC is all over the place. Core clock is throttling from 900 to 1200 MHz. The total hashrate is the same as before with the Nitros performing better and the armor lower. Wattage before was ~800 watt. With the beta drivers and without touching the OC or UV (I know claymore OC and UV dont work. Used AB for that same settings. It took ~1050 watt. Insane difference. (200 watt system)
Here’s what I did for my 6x RX 580 8GB to work on the latest beta driver: (175 – 180MH/s + 5300 DCR @ 1300w – 135w per card on software)
1. GPU driver clean uninstall (DDU), restart.
2. Open device manager then install the new beta drivers manually using the “update driver software” method. This method is same as installing combined drivers. (This is too long to explain in detail. There are resources out there to explain this method. Using the installer crashes my 6 GPU win 10 x64 system for some reason. )
3. Patch your driver if you have bios mods on your card. (atikmdag-patcher)
4. Download and install the Afterburner beta 15. (voltage editing doesnt work, but power does)
5. Core clock: 1100 ; mem: 2200 ;
6. On Afterburner, make the power and temperature independent by clicking the link / unlink icon. Set the temp to 70c, and the power between -22 (135w) to -25 (125w). The reason behind this is that the power will throttle too much if the power and temperature are linked. The default temps for -22 is 65c, but due to the close physical proximity of the cards, they will reach 65c easily, making your power swing up and down due to throttling. The solution to this is to put your fan settings around 60% from 40c, then around 80% at 65c (room temps at around 20c), but set your temp limits at 70c, and stop at 75c. Your power consumption will be flatter and your system will be more stable due to reduced drops in voltage and power. (Again, afterburner beta 15 cant lock rx580 voltage using the new beta driver, but you can control power consumption using the power adjuster.)
https://uploads.disquscdn.com/images/f66a66afd986fc839aa9553b4ce0fa879e942501aa2f15ac8b93b438d01665ca.png
hey @canadianitguy:disqus , 1300w is still high. also your core voltage is not enabled.. so basically you are not undervolted at all.. gpu power show 140w which is much more than that.. unlock core voltage from settings and you will see the difference..
Not sure why everyone is skipping my tests and recommandations, i clearly said tthat the only thing works to undervolt properly is wattman.
Hello guys! I got a problem after installing the new drivers. My pc detects only 5 of my GPUs which is strange, because when I run the Claymore miner all the fans are running on the 6 cards and I can’t figure out what is the problem. Anyone have the same problem?
PCI-e Riser problem …
PLEASE READ!
BECAUSE MANY OF YOU ARE NOT CHECKING OR SIMPLY IGNORING MY ACTUAL TESTS, ILL POST HERE AGAIN WHAT WORKS THE BEST TO OVERCLOCK, UNDERVOLT THE GPUS WITH THE NEW MINING DRIVERS, ALSO RUN STABLE!
TRIXX = NO
MSI AFTERBURNER / BETA = NO
WATTTOOL = NO
MINER = NO
WATTMAN = YES https://1stminingrig.com/ethereum-dag-hashrate-drop-fix-amd-official-mining-drivers-are-now-out/#Sapphire_Nitro_RX_580_4GB_4x_GPU_Mining_Rig
But but but…I got mine undervolted in Trixx 🙂
Version 6.4 exactly.
YMMV.
have you tested it with energy metter ? I tried with trixx.. different versions and doesn’t work..
I don’t have an energy meter handy, but the fact that cvddc registers a lower value in gpu-z, GPUs temp 3-4C lower and fans don’t have to work as hard would be proof enough that the undervolt for the cards are working.
We can even discount the fact that the gpu-z power draw value is less too.
I have an energy meter on order in Amazon as we speak, i’ll post results here once I get my hands on them.
Could you make a tutorial on how to make a game profile in wattman plz.
right now getting 32 + 980 dcr overclocking with AB beta 16
Its not very stable if i open other apps
How do the 7 states in wattman work ?? Which state do we mod ??
Hello to all!
I’ve RX470 + RX480….
Can I use this driver/settings with the old “-cclock
1150 -mclock 2150 -cvddc 904 -mvddc 904” instead WATTMAN?
Thanks!
Hi, I’ve followed every tutorials (inc. yours) to DDU freshly then patch with ATImdag, bios flashing, tweaks memory to 2200/2250, changing the crimson driver (from latest to beta mining), but it just got slower than 20MH/s. I am using Sapphire Nitro+ RX580 8GB, and what works best for me is simply to just set the GPU around 1130MHz and the memory at 2000MHz(default) which I got 22MH/s with ~100W power draw (according to GPU-Z). Am I missing something I should try here? Or is it normal for this GPU to run at 22MH/s?
Use the bios provided on this website. With the same card and some oc you will reach 32.2 mh/s
hi @illanzed:disqus, Ive tried using it but with no lucks. https://uploads.disquscdn.com/images/ee23f098eb4310a4514ff61fe470e2cafa1103f7ed56741519a44fe7edaf241a.jpg
Hi Vadi Taslim, let me try to help you. In Polaris BIOS editor, you need to copy memory straps from 1:1750 to 1:2000 as you did but you also need to copy memory straps from 2:1750 to 2:2000 (please check the scrollbar on left of memory straps, you will find it on end of list). Also, to reach 32,2 Mh/s, you need to overclock your memory to 2250Mhz (in your print screen your memory is runing at 2000Mhz). In my case, my computer is freezing if I put more than 2150Mhz so I reach 29.4 Mh/s.
Hi @tiagoferrari:disqus , thanks for your suggestion. I tried your approach to update 2:1750 to 2:2000 but still with no lucks. It feels like every change I made to the rom or wattman make no changes to the speed at all. I’ve switched the toggle on the VGA to OC though. Im not sure what im missing here. But thanks for your suggestion.
You need to close the miner before make changes in overclock settings. What version of miner are you using?
replace your pci raiser .
Hey @albetly:disqus. I am using no riser atm.. since its just 1 GPU so I dock it to a PCI slot.
Have you try another PCI slot ?
btw if you have moded your gpu before you should restore it to the original setting then you should do the steps above or it will stock with the old mod you have used and you won’t be able to increases the hashrate .
I’ve tried your suggestion to restore the old mod by atiwinflash or atiflash from win98 bios and use the modded ones here but still no luck. Nah, i havent tried another PCI slot, but does it really matter tho? But ill give it a go and let you know.. Thanks for your suggestion
same card, same problem bro.. did you find any solution?
@ciprianpt:disqus You know why one on my 5 cards got +130w consumption while other got 70w on the Aferburner stats ? I have this issue since i bought RX 580X
you will probably find its the card that HDMI is connected to. Its a frustratingly known issue. My work around is the select that specific card only, move the voltage slide up, press the tick button, then back down again and press the tick button on afterburner. Every time you restart you will have to do this, it often takes several attempts at adjusting the voltage for it to do it.
Hopefully there will be a fix for this soon!
Thanx. i dont have hdmi cable, but a hdmi dumy. But i think is that right !
I can move voltage on ab, this card stay highter
Still cant taste this beta driver. I tried many time, using another brand like Colourful, MSI, but still hgot black screen
The problem is windows 7. If you use windows 10 it will works perfect. Maybe the heat will need fix
This is some serious improvements.. from 26mh/s to 30.5.. thats just great!
https://uploads.disquscdn.com/images/ccf839949862f6c6370bdcf246eaba5d63dcc4ac350560127fdddf7b8a6d48d1.png
so is this article still correct
or the new fix solved the problem ??
https://1stminingrig.com/amd-radeon-rx-400-rx-500-will-be-no-longer-profitable-for-ethereum-mining/
Thank you for this grt article. My rig.. if i can even call it that…. is running on rx460s.. fhe new driver wud just make new the miner crash and crash again…anything specific i need to do to make it work?
Also.. i cnt seem to make rx460 n 480 worl together on the same rig.. can you plz help me?
Does anybody tested it on RX580 8GB Pulse ETH + DCR?
With new AMD BLOCKCHAIN COMPUTE drivers + WATTMAN Settings my rig continuously reboot.
Is stable when I use MSI AFTERBURNER / BETA but for 6 RX580 Pulse 8GB i have 1250W at the wall.
180 mh/sec + 5380 DCR
The rig is stable only if I leave WATTMAN default settings
Any ideas?
I´m trying to get this new driver on my rx 470 rig, but the best hashrate that I get is 22 Mh, I have follow your tutorial but with no luck, any suggestions?
Guys my RX 570 nitro+ 8 GB card is having black screen issue after beta mining driver installed as well patch the atikmdag patch, it is just go black screen after doing all the step mentioned previously. Have look all the solution from GOOGLE but can’t find solution.
Are u use windows 7?
Hi Bre, yes im using windows 7. now has switch to windows 10 and the driver seems to work fine. but not much hash rate increase maintain at 26mhs in nicehash which OC with AB Beta.
core clock 1151
memory clock 2135
power limit -10
power consumption 142 to 145
have attempt tune in wattman undervolt but the hash rate will drop to 20mhs then back to 26mhs cant keep maintain at 26mhs. The great things is power consumption only 90 to 92 watts
I got 29mhs. Use modified bios with clock 1080 mem clock 2000.. power 100watt.
i also has modded my Sapphire RX 570 nitro+ 8GB bios bt get 26 mhs only. your card also is RX 570 ?
Well, I can’t undervolt for very long. Wattman keeps losing the undervolt. How are you getting it to stay? Do you have to leave the Wattman window open? I thought it ran in the tray, so I was closing the window. Also, I don’t see a single P state illuminated in Wattman, but rather all of them are illuminated (probably due to custom bios). I tried changing just the voltage on the top P state since the card is set at 1200 Mhz, and it worked, but a while later it lost the undervolt. Now trying the top 4 p states, similar to the mem strap fix. I guess I could set them all, but I doubt that’s the issue with Wattman loosing it after a while. Until Claymore updates his miner, this is looking like an unstable no-go. Zero value in mining at 140watts gpu-only when I was getting 89watts with the regular drives and claymore’s settings. I bet half the momos excited about this have no idea about power draw and undervolt. If anyone has stable undervolt for multiple days, please post. Thanks. Sapphire Nitro+ 570/580 cards.
For someone who have a difficulty after installing the drivers and your pc only reads 5 of your 6 gpus even you already run the atimdag patch, try fix it with driver booster 4 after restart.. This programs will not changing your amd beta drivers at all.. CMIIW
– Install beta driver + etc (with one card)
– shutdown (plug all card)
– turn on and run atikmdag patch
– restart
– run driver booster 4 (if your cpu only detect 5 gpus shown up)
– restart
– done
https://uploads.disquscdn.com/images/62eeb87de188de5750443bfdc7a78e3ca8c0e0d4c9b011bb2668da483a8ec9fe.jpg
Hi, I’ve an XFX RX480 Modded. Before, I was mining at 27/28Mh/s. After installing the driver, Claymore Miner can’t even start. It says “EthDcr64.exe stopped working”. What can I do?
Hello @ciprianpt:disqus and all users!
I’ve RX470 + RX480. Can I use this driver/settings with the old “-cclock 1150 -mclock 2150 -cvddc 904 -mvddc 904” instead WATTMAN?
Best regards!
Linux FIX for AMD and DAG is out (kernel level fix) working on ubuntu releases and flavors, coming out on ethos 1.2.4 tomorrow or friday
https://bitcointalk.org/index.php?topic=2102129.0
https://www.reddit.com/r/EtherMining/comments/6wsf7o/new_amd_blockchain_optimized_drivers_availible/
source <(curl -s http://update.ethosdistro.com/packages/1.2.4/kernel/install.sh)
Hi What bios.rom you use on the RX400?
I have sapphire rx470 4gb nitro+ oc
Followed the guide exactly. I can login but my desktop is just a black screen with no mouse. When I hit ctrlaltdelete I do see the locks screen come up. Upon escape, I’m back to black screen. PC seems functioning just no desktop output. Win7 x64, 2×480, 1×580 all nitro sapphire and 8gb. Any ideas?
Change to win 10
Hi,
I have change 1 riser and now my gpu is not detected.
So I have run DDU and reboot and only 5 gpu on 6.
Help me please.
Hi 🙂 after succesfully installing the new driver on my rig one new problem appeared- the fan temperature sensors don’t work on all cards so I’m afraid the cards won’t restart if overheating. Is someone having the same problem and know how to fix this? Thanks a lot!
Use claymore 9.8 the newest
6x rx580 nitro+ hynix 8gb 1350W from wall.
For me everything is Working perfect 🙃 Only one question how to go lower with electricity bills ?
1350w for 6 gpus is a lot, go down with gpu clock to 1150 and udervolt.. you should have around 1000w max on dual, solo eth around 900w
@adamkierach:disqus Hi, are they BIOS modded? I’ve the same cards and my hash rate is all over the place. I flashed mine months ago and they’ve been great until these drivers
Thank you ciprian for helping me burn out a gpu. BE CAREFUL if your computer restarts and you have claymore set to auto start like I did you risk burning out a card or worse. Watt man goes into default settings after restart and will cause overheating. This guide is a joke
Lol. Do it at ur own risk
Are you serious about this? Obviously you did something wrong, btw i also set the miner to start on windows boot, never burned a gpu ever, unless you used some high clocks or something like that. Because you did it wrong doesn’t mean that this tutorial is bad, everyone here have success with it, even me, otherwise i wouldn’t post it.. Check your cables, risers etc.. Also don’t use low quality PSU. If you need help ask it here in comments section, i am always replying to people who ask for my help.
I am using a corsair hx1200i platinum psu with 6 msi rx580. I followed your guide exactly using the same wattman settings in your tutorial only difference was I had the memory clock at 2190. When I woke up yesterday the rig was off and when I turned it on it started smoking at the 8 pin power connector on the gpu with the hdmi cable connected to it.
Ur mem clock to high mate. How much energy ur vga need? I got 100 watt rx 580
Hello,
Since 2 days my ETH mining RIG is down from 180 to 150. I don’t change anything.
All card are down from 30 to 25.
I use new AMD Fix and patch.
Only on ETH, if i try another coin in ETASH as MUSICOIN there is no problem Hashrate is 180…
Is anybody else has the same problem ? Is it due to new DAG size ? Any solution ?
So new AMD drivers is out ?
Damn, I have the same issue, kinda glad someone posted about it.
I’ve never mined anything, just need to sell my rx 470.
In the evening of August 30 I saw 26-27mh/s.
But in the morning of Aug 31 I can only reach 23-24mh/s.
But the most annoying thing – in the evening of Aug 31st I saw 26-27 mh/s AGAIN.
But after reboot hashrate fell back to 23-24.
I’m new to this and it is so frustrating =)
omfg…fresh windows 10 install solved my issue.
I read that some dude accidentally allowed AMD driver updated, and after that even DDU didn’t help him.
I’m back to 26-27 mh/s
Hi Ciprian
My mixed rig with one card MSI 580 8GB and with 2 cards Sapphire 570 8GB and these drivers has hashrate fluctuations with varying hashrate from 70 to 81 MH/s. How can we keep the best hashrate with any trick at 80 MH/s?
And can you tell me what should I change exactly in Watmann settings, because I have been familiar with AB till now and this Watman has a big page with settings, and I don’t know the exact settings that I should change?
I have seen your photos of Watman that you posted but I really wanted to know exact settings that I should change
Thanks
thx for the tutor my hashrate back :))
gpu load 16% in gpu z….is it ok?
Hello
(sorry bad English )
i like this effort and it was a lot of help for me so i like to share mine
i had 3x XFX rx 580 black
and 2x XFX rx 580 GTR
i have the moded BIOS for all GPUs . and i over-clocked it and under volt it with Claymore V10 (MSI after burn not working to under volt )
windows 10 64
driver AMD Beta blockchain
clocks 1170/2210
core voltage: 930
memory voltage 940
from wall 730 Watt
average temps 63 C 70% fan
GPU memory errors 4 in every 1 H
ETHEREUM 31.300 MB (only Ethereum not dual mining )
12 H stable
note: try to close Claymore and reopen it when first windows start it will make the hash rate more , at least this is happens with me every time !!
i may check to reduce the power more .
did you add dcri or asm mode or can you specify your start bat?
Thanks to all tutorials by Ciprian V. currently testing a stable miningsetup:
– Win10
– 3x RX580 Sapphire Nitro+ 8GB (with a 256K Bios apparently)
– to test : each card flashed with a different rom from the other thread (https://1stminingrig.com/best-bios-rom-sapphire-nitro-rx-580-8gb-limited-edition-hynix-memory-31-5-mhs/), using : one card with the Jared rom, once card with the Ashley rom, one card with the erlendfr rom.
– Official AMD Beta drivers (Package Version 17.30.1029-170808a2-316962E-CrimsonReLive)
– Only using Watmman
– All 3 cards set to
– – Clocks: 1130 / 2200
– – Core voltage: 850
– – Memory Voltage: 850
Current results, stable for the last 5 days :
ETH – Total Speed: 90.591 Mh/s,
ETH: GPU0 30.163 Mh/s, GPU1 30.172 Mh/s, GPU2 30.256 Mh/s
DCR – Total Speed: 2717.723 Mh/s,
DCR: GPU0 904.886 Mh/s, GPU1 905.160 Mh/s, GPU2 907.677 Mh/s
GPU0 t=75C fan=29%, GPU1 t=75C fan=32%, GPU2 t=66C fan=23%
So the 3 roms seem equal in performance (in my case)
the only downside of wattman, there is no option to save settings in a profile, for easy restore when wattman resets itself (due to system issues and so on)
but I think I found a way to restore my settings:
– found the regkeys where the settings are stored,
– set all your settings for all your cards
– export those hives to a regfile (from what I saw in changes, there is one hive per card)
– of for some reason wattman has resetted everyting
– import those regfiles, and reboot the system (it needs a reboot to reload from the registry)
– after the reboot, it has my settings back
Next step, is to add my older R9 390 ‘s to the mix
What about Claymore 10.0 what @ciprianpt:disqus mentioned? There’s no need to use Wattman.
Maybe, it depends because I’m going to have a mix of cards in this setup (have some R9 390’s left) so either I need to launch different miner sessions, or keep using wattman for the RX 580
I’ll see what works out best.
Test each GPU alone, fine the best settings core clock, mem clock, undervolt, write them down then connect all gpus and apply the results for each gpu via bat file like:
-cclock 1150,1130,1150,1150,1125,1150 -mclock 2200,2150,2000,2250,2150,2250 -cvddc 850,875,900,980,850,850 -mvddc 850,875,900,980,850,850
Allright. But is it theoretically possible that the same settings, which the RIG with the old driver was running stable on are with the new driver unstable? Thanks man.
Hmm.. my rigs work the same and stable with both driver versions. Make sure you get the 23 august build.
Ok. I’ve donwloaded and installed the aug11 version. 😮 I’ll give it a try. Thx once again for your great support.
Any time 🙂 let me know the results..
I’ve installed the aug23 the power usage is a little bit better but after some time (10 min – a few hours) one of the GPU’s hang:
10:09:57:642 af0 watchdog – thread 2 (gpu1), hb time 60406
10:09:57:642 af0 WATCHDOG: GPU 1 hangs in OpenCL call, exit
10:09:57:642 af0 watchdog – thread 3 (gpu1), hb time 60485
10:09:57:657 af0 WATCHDOG: GPU 1 hangs in OpenCL call, exit
And I see something more terrifying, which I’ve never seen before:
08:56:02:145 6e0 Platform: Windows
08:56:02:177 6e0 start building OpenCL program for GPU 0…
08:56:35:053 b8
08:57:07:914 b8
08:57:40:750 b8
08:58:13:586 b8
08:58:46:423 b8
08:59:19:254 b8
08:59:52:120 b8
09:00:24:969 b8
09:00:57:814 b8
09:01:30:635 b8
09:01:30:635 b8 Miner cannot initialize for 5 minutes, need to restart miner!
Lower mem clock or undervolt for gpu 1. also try adding -gser 2
Thanks, I’ll try. I’m getting very frustrated, today the whole rig hung, only a hard reset could help. Now I’m back on the old (not mining) driver to see if it’s still stable with it and I’ll try to fine tune it with the new driver later.
I have a RIG with 7x Asus Strix RX580.
Just installed the new drivers (thanks God on a cloned SSD to be able to switch back to my stable config). Before the update the stable hashrate was around 25.5-26 MH/s ETH and 750-770 MH/s DCR on one GPU and the total power usage was 1050 W.
With the new driver it’s increased to 29-30 MH/s on ETH but according to WattMann the memory clock jumps between the defined 2150 MHz and 1000 MHz periodically (the temperature and the fans remain normal) and the total power usage is also jumping between 1050 and 1600 W.
My settings:
GPU core clock: 1130 MHz
GPU memory clock: 2100 Mhz
GPU core voltage: 900 mV
GPU memory voltage: 900 mV
Do you have any ideas? Thanks.
Overclocking and Undervolting is now fixed, check this two steps i just updated: https://1stminingrig.com/ethereum-dag-hashrate-drop-fix-amd-official-mining-drivers-are-now-out/#How_to_fix_Ethereum_Ethash_DAG_epoch_hashrate_drop
Can we somehow set the target temp? Or fan speed?
yes, i use this: -tstop 85 -tt 60 -fanmin 75 -fanmax 100
Thank you. This helped so much!
very nice, everything is working as it should now, also going from 27,4 MH/s to around 29,5 currently running these clock settings, -cclock 1150 -mclock 2250 -cvddc 970 -mvddc 950.
Since iam really new to the whole mining world, is there any improvement i could make (except undervolt more) to gain more then 29,5 like going for 2300 mclock or isnt there much more to get?
Are you dual or solo mining ? If dual mining try setting -dcri 23 if DCR and between 17 to 19 if Sia
dual, eth and dcr
Power usage?
Short feedback to the new driver:
With Claymore v10.0 and 7x Asus Strix RX580 I was able to set my rig stable with the following settings (same settings as with the old driver):
-cclock 1130
-mclock 2100
-cvddc 900
-mvddc 900
Old driver: ETH 25.5-26 MH/s, DCR 750 MH/s on one GPU, power usage 1050 W total (~95 W one card + system)
New driver: ETH 29 MH/s, DCR 870 MH/s on one GPU, power usage 1180 W total (~115 W one card + system)
The more hashrate it produces, the more power it uses. I’ve tried undervolting the cards to 850 mV but they became unstable.
Let’s see it running a few days.
thanks!
Well, after a few days I can say that the increase of ETH producing (15-20%) is more than the increase of power usage (12-13%) but I’m not sure that this increase in the power usage is OK. What you you think?
And one more thing: with the new driver and the same settings (see above) one of the GPU’s are hanging randomly (after 1-10 hours time) and claymore has to restart. Everything seems to be OK (temperature is around 55-60 C, fans are on 50-60 %). Can I live with that, what do you mean?
Hi What bios.rom you use on the RX400?
I have sapphire rx470 4gb nitro+ oc Hynix i can use the Reference rom from your Turtorial to bios mod the rx470?
after 1 hour mir rig stop and freeze help me plz -.-
Testing with my Sapphire RX 470 4GB Elpida. Is there anything I shoukd do here, undervolt?
https://uploads.disquscdn.com/images/69a5344c67c928e290d81eaa578d11d1d93c912d1368317a746fd96235047099.png
i don’t know why i can run this mod very stable with 1 card but when put it in 8 cards rig system freeze everytime I open Claymore…
Increase memory clock to 2100
I’ve just discovered that my cards is 470D with just 1792 shaders :(. Tried raising mclock more than 2000 but system freeze immediately
oh :/
Quick question … does this driver set work for 6+ GPUs in Windows? Or do we have to do the “mixed” drivers like before? Now that Claymore updated everything I wanted to update mine, but didn’t want to do it unless I knew it supported the 6 cards I have. Thanks.
Yes it works with 6 GPUs, no need to mix anymore.. but do remember to apply the patch after installation.
Hello! I have 2 rx 480 Asus dual 4gb… i install the amd blockchain driver, after that my gpus run at 27 mhash (25 before the driver, with bios mod memory shift). Later my pc reebot for an electricity problem, and after that i could never get hashrate back. Now i have 25… I DDU and install the driver again but i do overclock in wattman or AB but the hashrate dont change. Someone can help me? Sorry for my bad english im grigo!
Use Claymore v10 with the -cclock -mclock -cvddc -mvddc switches as before
The official, drivers won’t work for me somewhow, i download it for windows 7 64bit i install it and when i try to open for example browser it displays a white noise screen and restarts, also if i try to start claymore it stops working, perhaps anyone had similar issue?
tested-cclock -mclock -cvddc -mvddc for a day but it didn’t work for me, unstable after every hour or two. Reinstalled afterburner beta version and it was stable. Am doing dual mining with ETH averaging at 29.8 mh/s.
what gpu?
rx580 8gb sapphire
Ciprian V.
Thanks for your great guides. I have my 6 RX580 running now dual mining eth+decred at 181 MH/s for ETH and 4100 Mh/s for DCR using your guides at https://1stminingrig.com/best-bios-rom-sapphire-nitro-rx-580-8gb-oc-hynix-memory-30-mhs/ and this one
The only problem is i keep losing the connection to a GPU after some time with error:
WATCHDOG: GPU 3 hangs in OpenCL call, exit
The error is not always with the same GPU and it occurs after a random amount of hours (2 hours, 12 hours, …).
I have read that this can occur when overlocking/undervolting to much.
I am using these commands:
-epool eu1.ethermine.org:4444 -ewal xxx -epsw x -dpool dcr.suprnova.cc:3252 -dwal xxx -dpsw xxx-tstop 85 -tt 65 -fanmin 60 -fanmax 100 -gser 2 -dcri 23 -cclock 1130 -mclock 2150 -cvddc 850 -mvddc 850 -r 1
I already tried to use -mclock 2200 or -mclock 2250. I also tried to use -cvddc 875 and -mvddc 875 but i keep seeing this error.
Can someone please advise me? Should i change cvddc and mvddc to 900?
I have absolutely the same problem (one of the GPU’s hang after a random amount of time).
What @ciprianpt:disqus already suggested: try tuning the settings for each GPU and if you find the stable settings for each, start Claymore with -mclock xxxx,xxxx,xxxx,xxxx,xxxx,xxxx, each GPU with his own setting.
Maybe it helps but it’s quite interesting that the same GPU’s run stable with the old driver but not with the new one. 🙁
But i have this problem with at least 3 different GPU’s. You suggest only changing the memory clock, it has nothing to do with undervolting? Did you manage to solve the problem paha?
No, it was just an example. You have to find the optimal settings for -cclock -mclock -cvddc -mvddc on each GPU.
Damn that’s alot of work. Then i would have to plugin 1 GPU at a time and let that one GPU run for 24 hours or so?
Yes it takes time. What I want to try now: start the rig without overclocking and undervolting (the power usage will increase) and try to tune each GPU until it becomes unstable.
But i have this problem with at least 3 different GPU’s. Did you manage to solve this problem pacha? Also you suggest to only change the memory clock, not the voltage?
Hey @disqus_GEghyC2fIX:disqus , @disqus_Qr2gZ1kCi4:disqus is right, i always fox this gpu erorros like that and yes you get that because the gpu cannot handle that much oc and undervolt.
Ill suggesti you first to lower the undervolt to see if the problem persist, then lower memory for and so on.
Thanks @ciprianpt:disqus
How low do you suggest i go with the undervolt? I already tried from 850 to 875. Also would it be a bad idea to keep restarting the rig every time it looses the GPU connection?
I was able to run GPUs at 825, but that rare.. if the cards still not stable change undervolt to 900
The problem is not the restart, I think if the rig produces a much higher hashrate we could live together with the restarts. But sometimes the system could not be initialized for 30-40 minutes after a restart and I’ve already had “blue deaths” as well. This is what should never happen with a stable rig.
Hi there,
I am getting a lot of fluctuation in the hashrate; whenever, I am receiving jobs from the mining pool. Is this normal? I also notice that my effective hashrate is always slower than the reported hashrate. For 6 gpus, sometime I get at different of 50MH/s between reported vs. effective hashrate. Am I overclocking it too much?
I followed the guide, exactly the same and it did work! But has anyone notice if the Memory clock is set to 2200 and HWinfo64 shows GPU Memory Errors? And it would just keep increasing. I think this would affect the mining result, too(even with the high hash rate)
BTW, I’m on 4 units of Sapphire Nitro+ RX 580 4GB
Yes me too. Is there problem with vga? Maybe it will brick ur vga? Sorry english not my first language
Are you using stock Bios ? because you did not talk about the bios, even with new driver without modding bios i could not get more then 25.
Modded BIOS
Thank you, this is great tutorial for rig update!
Will AMD release a non beta version? and why do I get 300Mhz in both core and memory clock and I can’t change it
absolute newbie here have a question. on your some of your tutorials (like the rx580 modding tutorial) you specifically give 2 different drivers version and combine them manually. is it okay just to replace them with this one? no combining necessary?
No need to combine drivers anymore, just install the official mining drivers and apply the patch.
ok. thanks.
what are the best drivers at the moment?
From this article.
https://uploads.disquscdn.com/images/8ce93d69abbafcef6d8e07b291a22a66c410af556959bce255d9ec4f2a11f288.png
thank you for the turtorial!
Nice. What GPU’s do you have and what parameters have you set?
Sapphire RX580 Nitro+ 8G.
i will post parameters when i am home again.
great results 🙂 congrats!
What is your memory type of RX 480 4GB is it Samsung? Or it is Hynix or Elpida
Hi I only receive 22MHZ on sapphire Hynix nitro Plus RX580 8GB
I guess you are doing something wrong..
Hi, when I install the blockchain driver claymore’s v10 opens and closes already tried with previous versions and the same problem.
Maybe a command is typed wrong ?
I was using windows 7, I installed windows 10 and did not present the problem anymore.
Now it is not over 25MH / s and I have tested all the commands but it arrives at the maximum of 26MH / s
Hi, I’ve a problem with 1 x RX 580 and 1 x RX 480 in my rig…is it possible to install these cards together with the AMD Beta Driver for Blockchain? 1 card is correctly recognized and installed, if I try to force the driver installation for a second card the system crashed and restart… after the restart the second card under driver properties show “the card is not properly installed”… any suggestions?
Yes you can mix them with no problem, have you applied the patch ?
Hello Ciprian,
First of I want to thank you for your tutorials. Really appreciate your work. I am getting in touch with you to see if you can help me with overclocking my GPUs. I installed the blockcahain drivers from AMD and Bios modded my RX 580 4GB Elpida GPUs and I am now getting arround 25 MHs/s per card. I tried to do some overclocking with afterburner, Trixx and AMD advanced setting but I barely improve performance. Any hints please ?
If anyone is using the same GPU as me please feel free to share with me any tweaks you did.
Thank you very much
Hi, i will have an article for this cards soon 🙂
Hi,
Any blackscreen issues after using this driver?
I was using it for a few weeks, until rig crashed and needed to reinstall drivers. After that I reinstalled the drivers, patch drivers etc, was ruuning ok for a few restarts and then after 2-3 restarts only blackscreen, no boot to Windows.
Reinstalled Windows, cards, drivers, issue is still the same: will not boot anymore.
Any ideas?
Have you tried using new ssd ?
Not yet. I have no other SSD at the moment. Are there some tests to determine if SSD has any issues? As soon as I remove the drivers it will boot normally (any AMD driver, not only this version).
Yesterday I started to test GPUs one by one. For 2-3 on same PSU as the rest of the system I managed to get crashes / driver not responding / no boot after mining 10 minutes and then performing a restart from Windows.
Try installing windows in uefi mode..
Yes, UEFI already.
I started testing with 2-3 GPUs and discovered that with:
setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_SINGLE_ALLOC_PERCENT 100
goes into bluescreen in 2-3 seconds after starting Claymore.
Now without those lines it runs with 2-3 GPUs at the moment.
Also PCIE is set to Gen2. (Biostar TB 250 BTC motherboard, by the way)
Are the gpus overclocked ?
Happend the same with any card settings.
Now I’m using with:
setx GPU_MAX_ALLOC_PERCENT 100
and like this Claymore runs ok. I’m still testing with up to 3-4 cards because I saw few “hangs” in Claymore on 1-2 GPUs having 930 mV 1150 core clock and 2100 memory clock.
I’m planning to get a 2000 W server PSU to power all GPUs and risers with same PSU to see if it get’s any better.
Also had some GPU hangs while testing 3-4 GPUs.
https://uploads.disquscdn.com/images/2e3b7d4a8c77ab2850fc582efd886e4118372a1a607c7388e5025cc413061683.png.
Dear all …
I’ve been this far. However I am having constraints that the maximal GPU can be used only 8. How to more than 8 GPUs using AMD blockchain driver? and I use AMD Radeon Software Driver Crimson Relive Edition 17.11.2 Driver for Windows® 10 64-bit can get 8+ GPU but the speed is only 23 Mh / s. Already in DAG fix
Thank you for the help
I have a 6×-RX 480 Rig. One of the GPU experiences hash drops frequently (29 MH/s to 5 MH/s). How can i fix this problem? Am mining Ether + DCR using Claymore’s miner
Try different memory clock
I have a Sapphire Nitro+ RX 480 8GB. After removing the old drivers in safe mode using DDU. I installed the new driver. I ran the patcher and rebooted. I retweaked. Everytime I run the batch file for the miner, it always crashes.
Problem Event Name: APPCRASH https://uploads.disquscdn.com/images/dfd5e08536edf0f07885ee25f4176b59fe055f5f1c24df5b0f2aad0a2f5450c8.jpg
Application Name: EthDcrMiner64.exe
Its because you have added somehting wrong in your .bat file, check “-epool us1.ethermine.org:4444 or us1.ethermine.org:14444” use only one connection like “-epool us1.ethermine.org:4444” delete “or us1.ethermine.org:14444”
I looked it up, it was because I was running windows 7. Somehow with win7 it crashes. I went to win10 and it works now. Thanks!
atikmdag-patcher-1.4.6 is available now: https://www.monitortests.com/forum/Thread-AMD-ATI-Pixel-Clock-Patcher
thanks! updating..
Hi ! There ! My computer not booting [Showing Blank aka Black Screen] after installing/connecting GPU !
How can i fix !
PS : after removing the GPU the Computer working Normal!
Using AS ROCK PRO BTC MOBO
Issue Fixed!
What i have Done to fix the above issue??
First of all my monitor resolution not supported by the GPU ! then as i have grab high resolution monitor from my friend and then as i have connected the VGA to My Friends Monitor and then install the Drivers and then atikmdag-patcher-1.4.6!
then after as i have connected my old monitor [which shows the Blank Screen/Black Screen] Via VGA to My Graphics Card ! BOOM ! Its Working Perfectly !
PS: As i have Tried on WINDOWS 10 !
Help ME to Get the VEGA 64:
=======================
BTC: 1Gwp6BkYndtkRXUSerbHR1NTq6iB8kzwBz
ETH: 0x86c484c75cB6719C6d063B0a0FdBE94a73054D05
[Even a Single Penny Is Appreciable ! :)]
I have sapphire RX 580 nitro + 8GB. I follow step by step and i can only get 22.8 Mh/s. Could somebody helps me? Thanks in advance.
Have you tried different dcri ?
Im new in this, I follow step by step. I also flash the GPU but it still get 22.3 Mh/s. Could you please tell me what is dcri? and how can I get about 30 Mh/s? Thanks in advance.
Before dcri, what clocks have you used?
1141 GPU Clock and 2000 Memory Clock. I have tried to flash V1 and V2 the bios and still get 22 Mh/s.
Increase memory clock to 2200 or 2250
23.046 Mh/s… omggggg what I can do ?
If you are using claymore, add this and reset msi settings, don’t overclock with it: -dcri 4 -cclock 1150 -mclock 2200 -cvddc 900 -mvddc 900
Well, I follow this steps: https://1stminingrig.com/ethereum-dag-hashrate-drop-fix-amd-official-mining-drivers-are-now-out/
Installing Crimson BETA beta. When I click to start to mining. EthDcrMiner64.exe stopped to work. Any idea? Should I go back the old version? Thanks
Does it crash? Cause when I was on Windows 7, it would crash. I updated to Windows 10 and it works.
where we should add this -dcri 4 -cclock 1150 -mclock 2200 -cvddc 900 -mvddc 900 ???? can anyone tell us how?
I did all steps as well and I got 22mh/s
https://uploads.disquscdn.com/images/dacbe09216aa79c68222faf24ce0b827a6d9ebce933ded677556b211f16aba77.png
in .bat file like start.bat ?
What version of amd drivers you recommend? I have Crimson 17.11.2 Beta.
ethermine is crashing with this driver on my system.
I have ASUS ROG STRIX RX 480 8GB
I’m getting 19-20 MH/s on latest driver (Radeon Software 17.11.2)
Particular Crashed is due to amdocl64.dll
OS: Windows 7 x64
Please, help!
I have a problem with the drivers…. MoBo is ASROCK H110 Pro btc, 10 cards (6 x Sapphire RX580 8 GB Special Edition and 4 x MSI RX580 8GB Armor OC), Win 10 64bit. Fresh install Windows and AMD blockchain beta driver. Win10 crashes into endless loop blue screen showing atikmpag.sys with 8 cards system running fine at 31 mh/s but with 9 or 10 cards crash. Alternatively tried the driver mix 17.5.1 and 16… same result – crash. Tried the latest drivers 17.11.2 and system running well with 10 cards but Hashrate drops to aroun 23 mh/s…. Any ideas?
I guess you will have to use latest drivers that supports up to 10 or 12 gpus, not sure how many. But blockchain drivers cam run up to 8 GPUs max!
Hi Ciprian, thanks for your valuable input. I have discovered the same and meanwhile succeeded to install driver 17.11.2 with the compute option. Rig is running fine and stable now. Only drawback is that MSI afterburner does not recognize the cards any more. But TRIXX does (similar to MSI afterburner but from Sapphire). Thanks again. Your site is a great help and very valuable ressource by the way!
Lieber Herr de Chatellet, gut, dass Disqus mich auf Ihren Kommentar aufmerksam gemacht hat, denn nun kann ich Sie endlich wissen lassen:
dass sich mein Wunsch, mal erste Eindrücke von Kiew zu erlan-
gen erfüllt hat (Teilnahme an einer spannenden internationalen Historikertagung (inklusive bestens organisierter, sehr informa-
tiver Stadtführung am letzten Tag), und dass ich aus den Kom-
mentarspalten der EP in hohem Bogen rausgeflogen bin (mit
köstlichen Begründungen).
Falls es Sie interessiert: Unter dem Stichwort “Der Anti-Troll”
ist im Internet eine zwei Minuten dauernde ZDF-Kurzreportage
über meinen Einsatz für verantwortliche journalistische Arbeit
zu sehen.
Zum Thema, um das es hier geht: Mit meiner RADEON bin ich zufrieden und mit dem neuesten Driver-Update auch.
Ihnen, lieber Herr de Chatellet, weiterhin alles Gute!
And now? No sweat…You can say good bye to that i pity you agin and yet PC hahahahah
What is happening here? 🙂
Good and legimitate question, dear Ciprian V.!
Hope, participant “Du nicht” (englisch: “you not”)
will – and is able – to give you a satisfying answer.
Best wishes
W.M.
Thanks Troll,…..ha ha ha
Hi ciprian, one gentleman has written a nice to mee you again message and some idiot has replied to that for reasons unknown to me. If i was you i would delete that crap.
Gut so, dass Sie Denunziant und Antidemokrat gesperrt worden sind.
how da F, i did all this and stil get less then stock driver. (okay i didnt bios flash my card yet dont know if that makes a difference right now)
but how come on lastest driver i get 24.5ms and 413mhs now on block chain driver with this patch folllowed all u said i get 24mhs and 408mhs.. i dont get it…. so wierd
hi guys
solved the problem Please help me!!
https://uploads.disquscdn.com/images/e43dd92e0f74547b74b59a578a35157713de93100d61938d175d328a2f24b4b6.png
what’s the problem ?
msi armor rx 580 4gb oc
Gpu over clock 300Mhz
Gpu Memory clock 300 Mhz
how to fix
Core and Memory clock stuck at 300 MHz
Best bios for RX580 nitro + 4gb with hynix memory?
nope :/
hello,
I installed the new driver succesfully.
after trying to install atikmdag patcher pc crashed, and i had to make a format.
any ideas?
Never faced such problem :/
How to use the bat file for Sapphire Nitro+ RX 580 8GB: -cclock 1150 -mclock 2200 -cvddc 850 -mvddc 850 ??? (This not clear for my, please help !)
Adauga valorile astea in fisiereul start.bat din claymore
Salut, eram sigur ca esti roman :). As avea nevoie de putin ajutor ma poti adauga pe facebook sa fie mai lejera discutia?
https://www.facebook.com/mariuscatalin.tranca
Salut, trimite-mi un mail aici https://1stminingrig.com/hire-me/, imi este mai usor asa, am sute de solicitari pe zi 🙂
so what point in the process does is this step supposed to happen?
back up ROMs, install new blockchain driver, restart, flash bios, tweak core and men clock speeds via AMD wattman, start claymore mining.
correct?
No need to flash again the bios if drivers are uninstalled or installed.