Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux)

[cyeedar]

Hi Claymore

Have you managed to replicate the issue with 12 gpu systems
issue: on version 9.8 if you have more than 3 amd cards in a rig of 11 or 12 gpus ethdcrminer hangs on Algorith ASM, if you start the instance with 3 amd cards and the rest nvidia then it starts mining. then you have to start another instance of ethdcrminer with the remaining cards, .
This issue was also experienced by a youtube channel Cryptomined where he had to have multiple instances on 9.8 running in order to use all the cards.
When all cards are selected after the cmd window hangs nothing but a hard reset solves the hang.
Hope that helps, if you need any more info or troubleshhooting with this regards you can let me know.
thanks
 

I don't have appropriate hardware to test 12GPUs configuration. You can try some different miner to check if it is more stable, because the reason can be not in miner. Also try "-gser 1" or "-gser 2" option.

Hi Claymore can you please tell me what's with checkread timeout that's been happening to some of my rigs inside log. That's usually the last line of message before the whole system reboots itself or hang in a loop required hard restart. Thanks!

[1714601000]

1714601000

[1714601000]

1714601000

[1714601000]

1714601000

[1714601000]

1714601000

[TechPark]

Hi Claymore

Have you managed to replicate the issue with 12 gpu systems
issue: on version 9.8 if you have more than 3 amd cards in a rig of 11 or 12 gpus ethdcrminer hangs on Algorith ASM, if you start the instance with 3 amd cards and the rest nvidia then it starts mining. then you have to start another instance of ethdcrminer with the remaining cards, .
This issue was also experienced by a youtube channel Cryptomined where he had to have multiple instances on 9.8 running in order to use all the cards.
When all cards are selected after the cmd window hangs nothing but a hard reset solves the hang.
Hope that helps, if you need any more info or troubleshhooting with this regards you can let me know.
thanks
 

Newbie with 12 cards setup? Did you go straight to uni after 5th grade?

Do you know any expert here that have a solution for this problems? We the group of noobs didn't found one ...

Well... If one is not sure what the one is doing the one asks for an advice BEFORE the action...
Why anyone here will spoon feed spoon feed a noob who is going to be his direct competition...

[notbatman]

Only MSI RX cards are availible in stock and those guys sold me a bunch of defective cards then took 5 weeks to RMA. So now my question is what's the fastest non-MSI GTX1060 card/brand?

[jackbox]

Only MSI RX cards are availible in stock and those guys sold me a bunch of defective cards then took 5 weeks to RMA. So now my question is what's the fastest non-MSI GTX1060 card/brand?

You want Samsung or Micron memory. Hynix memory sucks on these cards.

[c1oud]

I guess my rx 580s hashing 161mh/s dropped to 159mh/s today is due to dag epoch 140. It is getting lower and lower as the epoch goes up. I hope remedy for this is on its way.

We have been trying our best for the last 10 days. It that did not help, nothing else will 

same problem with RX480 rigs.... any ideas that might help prevent continues decreasing of hasing power? Seems like its not only gives less hashes it uses less energy as well but who cares about energy we need maximum output until difficulty is still suitable for these cards...

so may be there are new ideas on upgrading Polaris bios'es to work with new epochs? In past 2 months hashing power decreased by almost 9%

any suggestions are very appreciated...

[sikci]

Hello cryptomasters,
been watching this thread for a while and after succesful update, tought I should contribute to growing statistics around latest AMD beta blockchain driver on Win10.

Report:
Upgraded from amd 17.7.1 to AMD Beta blockchain driver-aug11 via DDU reinstall in safe mode and atikmdag patcher. Installed MSIafterburner 4.4.0 Beta16 + RTSSSetup700Beta28 (latest) afterwards. On previous amd driver I had all OC/UV parameters set in Claymore's bat file, after upgrade to beta driver I deleted them all and set everything in MSIafterburner. Undervolting and memory oc succesful.

The rig is pulling around 50W more electricity from the wall (have wattmeter on it) than last month average on old driver. That's from 900W to around 950W on dual mining ETH+SIA -dcri 16, Claymore 9.8. I think this is mostly due to parameter (in)precisions on api level and power limit in MAB still has some reserves...

My rig:
6x MSI RX570 GamingX 4GB (Elpida+Hynix). Basic bios mod applied (1500 and 2000 straps copied)
MB ASRock H81 Pro BTC R2.0
PSU Corsair HX1200i
CPU Intel Celeron G1840
8GB crucial ram
molex powered raisers
ssd 60gb
Claymore 9.8
OS: Windows 10 Pro in Test Mode, Build 15063.rs2_release.170317-1834

MSI Afterburner 4.4.0 Beta16 settings:
-core voltage (mV): -96
-power limit (%): -28
-temp. limit (unlinked): 72
-core clock (MHz): 1100
-memory clock (MHz): from 1922 to 1995 (depends on specific card)
-fan on auto.

All cards temperatures on average: 69C.

Verdict:
Beta blockchain-aug11 driver improved my average hashrate to around 160-162mhs from 156-158 on previous driver. The rig is stable 24h+ now and counting.

Cheers from cryptorookie and his first post :=

[ascaloon]

Hello!

Anybody has news about whats the diference between 11th aug and 23th aug relesed blockchain fix driver?


win10-64bit-crimson-relive-beta-blockchain-workloads-aug11.exe
Win10-64Bit-Crimson-ReLive-Beta-Blockchain-Workloads-Aug23.exe


Thanks.

[jackbox]

Hello!

Anybody has news about whats the diference between 11th aug and 23th aug relesed blockchain fix driver?


win10-64bit-crimson-relive-beta-blockchain-workloads-aug11.exe
Win10-64Bit-Crimson-ReLive-Beta-Blockchain-Workloads-Aug23.exe


Thanks.

VEGA support added. That's it. It still does not work properly.

[rednoW]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

[mizzack]

Finally figured out how to reproducibly get undervolts to (mostly) stick on the blockchain drivers with Afterburner and Wattman

1) Apply clocks/volts in Wattman. e.g. 1130/850.
2) Launch AB with no adjustments set. (+0mV)
4) Launch Claymore and wait for the miner to fully kick in. Observe your voltage in AB (that you set in Wattman) get completely ignored.
5) Change your Wattman clocks to, say, 1135Mhz and apply (then back to 1130 and apply). Observe the voltage in AB go back to a static, undervolted clock.
6) Note the ∆ in AB. E.g if I set Wattman to 850mV, it might be reading 937mV in AB.
7) Adjust AB voltage offset but don't bother applying it yet. It won't do anything. So, to go from 937mV actual we'd want -90 offset in AB.
8 ) Stop the miner.
9) Apply AB voltage offset.
10) Launch miner. You'll see your voltages jump up again, so do step #4 again. You should now be pretty close to your desired voltage.[/li][/list]

I've only seen it reset to high voltage occasionally when doing something like turning the monitor on/off. Step #4 resolves it.

With this I'm seeing ~5-10W per card improvement over 17.7.1 dualmining.

[Shnikes101]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

The card definitely prefers to stay cool. What's your one minute ETH avg.? Good to see you got Memory stable at 1100 - also at 1v? Beta mining drivers or 17.8.1?

[nadrojcote]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

That decred hashrate is impressive!

[rednoW]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

The card definitely prefers to stay cool. What's your one minute ETH avg.? Good to see you got Memory stable at 1100 - also at 1v? Beta mining drivers or 17.8.1?

HBM2 voltage cannot be changed with now available tools.
I'm using 17.8.2 drivers.
Beta mining drivers caused constant rebooting win7 64 on windows load.

[gegstore]

ethman don't work with 9.8 version
help me

It does work for me.
What is your bat/config file?
What is the error message?
Screenshots will be helpful...

no error message but only from other pc i don't see nothing loss connection
with 9.7 i see all

Provide screenshots and config files

i have resolve from myself, windows firewall don't have the miner in the whitelist

i don't can post the screenshot but i repeat with claymore 9.7 all work good

Well... if you do not provide any more information there is nothing to work with...
ETHMAN works for me and I have 9.8

[NavyCuda]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

Drop your clocks to 1137.  Same hash rate, much cooler running.

[rednoW]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

Drop your clocks to 1137.  Same hash rate, much cooler running.

Definetly not the case when dual-mining. Hashrate decreased significantly.
And the voltage stay the same.
Will get power meter and need some advice howto downvolt below 1v to test low-power modes.

[NavyCuda]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

Drop your clocks to 1137.  Same hash rate, much cooler running.

Definetly not the case when dual-mining. Hashrate decreased significantly.
And the voltage stay the same.
Will get power meter and need some advice howto downvolt below 1v to test low-power modes.

I was getting 42.1Mh/s and ~1100Mh/s DCR at 1137cc/1100HBM2.  Until it hard crashes... Which I think is related to the block chain driver and Vega FE.

[rednoW]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

Drop your clocks to 1137.  Same hash rate, much cooler running.

Definetly not the case when dual-mining. Hashrate decreased significantly.
And the voltage stay the same.
Will get power meter and need some advice howto downvolt below 1v to test low-power modes.

I was getting 42.1Mh/s and ~1100Mh/s DCR at 1137cc/1100HBM2.  Until it hard crashes... Which I think is related to the block chain driver and Vega FE.

What was the core voltage with this settings?
Pitty, I had to low memclock to 1080mhz because there were some incorrect ETH shares reported by miner.
Now seems ok ... and no single crash yet ...

[noviceminer]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

Drop your clocks to 1137.  Same hash rate, much cooler running.

Definetly not the case when dual-mining. Hashrate decreased significantly.
And the voltage stay the same.
Will get power meter and need some advice howto downvolt below 1v to test low-power modes.

I was getting 42.1Mh/s and ~1100Mh/s DCR at 1137cc/1100HBM2.  Until it hard crashes... Which I think is related to the block chain driver and Vega FE.

What was the core voltage with this settings?
Pitty, I had to low memclock to 1080mhz because there were some incorrect ETH shares reported by miner.
Now seems ok ... and no single crash yet ...

Have you tried ETH + SIA and other combinations ?

[NavyCuda]

RX Vega 64 Liquid 1560gpu/1100mem@1v gpu core
dualminer 9.8 -dcri 60
55-60 degrees with 3000 rpm

Drop your clocks to 1137.  Same hash rate, much cooler running.

Definetly not the case when dual-mining. Hashrate decreased significantly.
And the voltage stay the same.
Will get power meter and need some advice howto downvolt below 1v to test low-power modes.

I was getting 42.1Mh/s and ~1100Mh/s DCR at 1137cc/1100HBM2.  Until it hard crashes... Which I think is related to the block chain driver and Vega FE.

What was the core voltage with this settings?
Pitty, I had to low memclock to 1080mhz because there were some incorrect ETH shares reported by miner.
Now seems ok ... and no single crash yet ...

I set core voltage with claymore's miner, 1000mV.  I needed 1000 or my memory overclock would fall to 800MHz.  On vega fe beta drivers, 950Mv.