PhoenixMiner 6.2c: fastest Ethereum/Ethash miner with lowest devfee (Win/Linux)

[Ursul0]

In the case of mixed rigs with 8gb and 4gb cards, what will be the behavior? I seemed to have understood that if there is at least an 8GB card the rig should continue to work correctly, is it correct?

no. that's total crap. DAG must be loaded into the VRAM of the GPU completely, if it does not fit => it will not work.

This is where we are now:
(both have PCIe as main GPU, "Low res" has no display, "higher res" has HDMI dummy display )

[bategojko74]

In the case of mixed rigs with 8gb and 4gb cards, what will be the behavior? I seemed to have understood that if there is at least an 8GB card the rig should continue to work correctly, is it correct?

no. that's total crap. DAG must be loaded into the VRAM of the GPU completely, if it does not fit => it will not work.

This is where we are now:
(both have PCIe as main GPU, "Low res" has no display, "higher res" has HDMI dummy display )

I think it's time PhoenixMiner to update and after that all 4GB cards will be happy until 21 of December and all 8GB cards will be happy next 5-6 years

[delevic]

It's OS problem....

[bategojko74]

It's OS problem....

for the first GPU or your integrated video

[Congdm]

-rvram -1

Yes use -rvram -1   or NBminer maby it works better,i think maby Phoenixminer developer is dead !!!    

This worked for me too. That DAG epoch change caused 3 rigs to crash this morning. 4GB cards are getting tight now...

-rvram -1 WORKS PERFECTLY FOR 4GB AMD CARDS!!!

That was literally the first thing I said...

Get ready to move on Linux slowly. Windows reserves 300 + MB of VRAM. The current DAG size is 3.57gb. The Phoenix miner free vram check was set to 384mb by def.  When the dag + win reservation is taken away from 4gb vram, only 130MB of vram is left (that's why Phoenix made the problem). And with the command -rvram -1 you have time up to DAG 3.7gb or ~ 1.7.2020. After that, Linux....until 21.12.2020.

In the case of mixed rigs with 8gb and 4gb cards, what will be the behavior? I seemed to have understood that if there is at least an 8GB card the rig should continue to work correctly, is it correct?

I think the lead card (GPU0) is better with 8GB VRAM as it reserves more than the rest if you check GPU memory usage in MSI Afterburner, or you can try integrated graphics for the main display output but I had problems with not being able to control the GPU voltages after that - that was nearly 2 years ago that I last tried though.

It would be interesting to understand now if it is working because this would allow to keep the 4Gb cards having only 1 8Gb card

That is impossible. Only hardware intervention

[invicta]

-rvram -1

Yes use -rvram -1   or NBminer maby it works better,i think maby Phoenixminer developer is dead !!!    

This worked for me too. That DAG epoch change caused 3 rigs to crash this morning. 4GB cards are getting tight now...

-rvram -1 WORKS PERFECTLY FOR 4GB AMD CARDS!!!

That was literally the first thing I said...

Get ready to move on Linux slowly. Windows reserves 300 + MB of VRAM. The current DAG size is 3.57gb. The Phoenix miner free vram check was set to 384mb by def.  When the dag + win reservation is taken away from 4gb vram, only 130MB of vram is left (that's why Phoenix made the problem). And with the command -rvram -1 you have time up to DAG 3.7gb or ~ 1.7.2020. After that, Linux....until 21.12.2020.

In the case of mixed rigs with 8gb and 4gb cards, what will be the behavior? I seemed to have understood that if there is at least an 8GB card the rig should continue to work correctly, is it correct?

I think the lead card (GPU0) is better with 8GB VRAM as it reserves more than the rest if you check GPU memory usage in MSI Afterburner, or you can try integrated graphics for the main display output but I had problems with not being able to control the GPU voltages after that - that was nearly 2 years ago that I last tried though.

It would be interesting to understand now if it is working because this would allow to keep the 4Gb cards having only 1 8Gb card

That is impossible. Only hardware intervention

GPU0 needs more reserved VRAM for Windows, so by placing an 8GB card there, you don't need to worry about this. It is not to say an 8GB card first and the rest 4GB cards will work past 23rd December 2020, it is to say it'll work until then and not stop early.

In the case of mixed rigs with 8gb and 4gb cards, what will be the behavior? I seemed to have understood that if there is at least an 8GB card the rig should continue to work correctly, is it correct?

no. that's total crap. DAG must be loaded into the VRAM of the GPU completely, if it does not fit => it will not work.

This is where we are now:
(both have PCIe as main GPU, "Low res" has no display, "higher res" has HDMI dummy display )

Actually sometimes the GPU seems to use shared system memory and drop the hashrate down to 1.5-2MH/s - which is still basically useless as it's slow AF.

[delevic]

GPU0 needs more reserved VRAM for Windows, so by placing an 8GB card there, you don't need to worry about this. It is not to say an 8GB card first and the rest 4GB cards will work past 23rd December 2020, it is to say it'll work until then and not stop early.

To test your theory just add in .bat file -benchmark 353. In that epoch dag size is 3.76gb and miner in Windows fail to create dag. Maybe you will need to use Claymore's  miner  because Phoenix has some bug when you start -benchmark option.

my test in Claymore with IGFX primary gpu:

epoch 353 dag 3.76gb
Windows failed , Ubuntu works

epoch 378 3.95gb
Ubuntu failed

[delevic]

It's OS problem....

for the first GPU or your integrated video

In any variant windows will make reservation of vram on all GPUs.

[shostt58]

to phoenix - raveos adding amd drivers 20.10 but phoenix miner - opencl not support
rx470 and rx5700
need adding new driver or opencl
senks

[WARMANGALAXY]

when new releases a amd linux 20.10 drivers

[Ursul0]

when new releases a amd linux 20.10 drivers

what is this weird obsession with driver versions?
what hw do you have and what are the expectations?

[Ursul0]

to phoenix - raveos adding amd drivers 20.10 but phoenix miner - opencl not support
rx470 and rx5700
need adding new driver or opencl
senks

that's the versions I'm using for my polarises. 18.6.1 is the best. 19.9.1(or 2 whatever) - fucked up fan control

[bategojko74]

New epoch has come. For all with 4GB AMD cards who receive message not enough memory to allocate DAG please use -rvram -1
This will fix your problem.

[MireNS]

I can't fix Inccorect ETH share from GPU and geting around 50% stales on my RX 5700(not XT) in Phoenixminer 4.9c. I got stability and GPU temp around 60 C but this is anoing rly
Here is my setup file:
-pool ssl://eu1.ethermine.org:5555 -pool2 ssl://us1.ethermine.org:5555
-wal
-worker test1
-tt 62
-cvddc 750
-cclock 13000
-fanmin 0
-mclock 910
-amd
-acm
-mi 12
-gt 126
-mode 1
-clKernel 1
-clNew 1
-clf 0
-lidag 1
-openclLocalWork 128
-openclGlobalMultiplier 4096
pause

Pls some1 help me or I will kill this GPU
My print scr: https://ibb.co/SxkmtYf

Ummm... confirm your -cclock again?
I'm no expert but 13000 seems a little high. Maybe 1300 would be better.

And as the other said, ditch -openclLocalWork 128 -openclGlobalMultiplier 4096

Its 1300 dono how that 0 come there

"And as the other said, ditch -openclLocalWork 128 -openclGlobalMultiplier 4096"
Done this and its stabil on 50Mhs on 114w for now but need better drivers coz on last one 20.4 you get bad shars coz of driver

[ivcelmik]

Hi all,

Does anyone know why on ubuntu gpu0 has lower hasrate (when mine ethereum) instead of 31 like the others it has 15-16 Mhs and how can this be fixed.

Ty in advance

[bategojko74]

I can't fix Inccorect ETH share from GPU and geting around 50% stales on my RX 5700(not XT) in Phoenixminer 4.9c. I got stability and GPU temp around 60 C but this is anoing rly
Here is my setup file:
-pool ssl://eu1.ethermine.org:5555 -pool2 ssl://us1.ethermine.org:5555
-wal
-worker test1
-tt 62
-cvddc 750
-cclock 13000
-fanmin 0
-mclock 910
-amd
-acm
-mi 12
-gt 126
-mode 1
-clKernel 1
-clNew 1
-clf 0
-lidag 1
-openclLocalWork 128
-openclGlobalMultiplier 4096
pause

Pls some1 help me or I will kill this GPU
My print scr: https://ibb.co/SxkmtYf

Ummm... confirm your -cclock again?
I'm no expert but 13000 seems a little high. Maybe 1300 would be better.

And as the other said, ditch -openclLocalWork 128 -openclGlobalMultiplier 4096

Its 1300 dono how that 0 come there

"And as the other said, ditch -openclLocalWork 128 -openclGlobalMultiplier 4096"
Done this and its stabil on 50Mhs on 114w for now but need better drivers coz on last one 20.4 you get bad shars coz of driver

-openclLocalWork 128 is totally wrong and it creates all your wrong shares. PM can't work properly for openclLocalWork higher than 32 for NAVI cards and higher than 64 for Polaris cards

[MireNS]

I can't fix Inccorect ETH share from GPU and geting around 50% stales on my RX 5700(not XT) in Phoenixminer 4.9c. I got stability and GPU temp around 60 C but this is anoing rly
Here is my setup file:
-pool ssl://eu1.ethermine.org:5555 -pool2 ssl://us1.ethermine.org:5555
-wal
-worker test1
-tt 62
-cvddc 750
-cclock 13000
-fanmin 0
-mclock 910
-amd
-acm
-mi 12
-gt 126
-mode 1
-clKernel 1
-clNew 1
-clf 0
-lidag 1
-openclLocalWork 128
-openclGlobalMultiplier 4096
pause

Pls some1 help me or I will kill this GPU
My print scr: https://ibb.co/SxkmtYf

Ummm... confirm your -cclock again?
I'm no expert but 13000 seems a little high. Maybe 1300 would be better.

And as the other said, ditch -openclLocalWork 128 -openclGlobalMultiplier 4096

Its 1300 dono how that 0 come there

"And as the other said, ditch -openclLocalWork 128 -openclGlobalMultiplier 4096"
Done this and its stabil on 50Mhs on 114w for now but need better drivers coz on last one 20.4 you get bad shars coz of driver

-openclLocalWork 128 is totally wrong and it creates all your wrong shares. PM can't work properly for openclLocalWork higher than 32 for NAVI cards and higher than 64 for Polaris cards

This is now my setup and its work nice for now, 50MHs on 114W:
-tt 62
-cvddc 750
-cclock 1250
-fanmin 0
-mclock 910
-amd
-acm
-mi 12
-gt 126
-mode 1
-clKernel 1
-clNew 1
-clf 0
-lidag 1

[bategojko74]

This is now my setup and its work nice for now, 50MHs on 114W:
-tt 62
-cvddc 750
-cclock 1250
-fanmin 0
-mclock 910
-amd
-acm
-mi 12
-gt 126
-mode 1
-clKernel 1
-clNew 1
-clf 0
-lidag 1

I am not sure if your -gt 126 is result from auto-tune or copy-paste from some other config. You can remove it, start the miner and when the auto-tuning finishes to take the best gt value from the last log and put it back in your -gt setting.

Also apply this advices from SerpentXSF:

Asrock Challenger RX 5700 (Non-XT) - I got it 24/7 stable with the following settings in Phoenix mining ETH. I also share some tips for those like myself who experience AMD drivers crashing and resetting their clocks in Windows. Check it out below.

Config:   0.850 vcore | 1350 core | 1840 Mem | 56.97 Mhs | 118 watts

I can get to 58 Mhs + but it is not 24/7 stable with current drivers, I also modified my GPU in past content released on my channel.

Improved ETH Hashrate | Red Bios Editor Mem Mod
https://youtu.be/LwGUJjQgj38

Red Bios Editor - More Power Tool (Igor's Lab):
https://www.igorslab.de/red-bios-editor-bios-eintraege-anpassen-optimieren-und-noch-stabiler-uebertakten-navi-unlimited/

Force flash your GPU vbios, but MAKE SURE to always back up your original.
Commands to run:

amdvbflash -unlockrom 0

amdvbflash -p -f 0 NAMEOFYOURBIOS.rom

Updated - MorePowerTool - 1.2
https://www.igorslab.de/morepowertool-amd-radeon-rx-5700-und-rx-5700-xt-tweaking-und-uebertaktungssoftware/2/

[superlative1974]

Changes in version 4.9c (since 4.8c):

You can download PhoenixMiner 4.9c from here:

https://mega.nz/#F!2VskDJrI!lsQsz1CdDe8x5cH3L8QaBw (MEGA)

you can download this miner somewhere else, is it really so difficult for you, many do not have the ability to download because mega requires money, it’s sad to see.

[superlative1974]

I’ve been trying to download for free for 2 hours, but the speed is less than 9 kb per second. it's cruel
https://c.radikal.ru/c39/2004/02/51abf7050897.png