A Multi GPU Miner is Available for Ubuntu!


#61

The mining executable does not seem to work properly. I have no insight in what might be wrong. But the Cuda miner seem to abort on line 369?! I see your card has only 6GB memory, you could try to add -E 1 to extra_args in the epoch.yaml and see if the error goes away.


#62

Thank you for your reply, I joined this configuration, but still the same as the previous error. It’s unlucky, I try to reinstall CUDA again.


#63

gpu miner works with docker container?


#64

Yes, it should. Have a look at this guide:

Best,
Vlad


#65

Can we please get a clear answer on this:

How is this

But it is more safe to generate one in the Airgap Wallet. With this method your private keys don’t preside on the rig.

connected to this

---

keys:
 dir: keys
 peer_password: “secret”
 [...]
mining:
 beneficiary: “ak_blablabla”

-> Do I only need to ensure the beneficiary (AirGap wallet public key ) is set and can I ignore the keys and secret settings ?


#66

yes you can if you use airgap generated address


#67

Hey I tried this procedure everything worked well.

  1. I installed all dependencies
  2. I installed Drivers & Cuda
  3. I build the node
    4.I build the miner
  4. I configured the node
  5. I launched the node and run miner
    Everything worked perfect…But…the node got synced. But my miner shows wiered when I checked with nvidia-smi. I have been mining since last 10 hours and still the same issue. Here is the screen shot of nvidia-smi output.

I am using 3 1060 3GB GPU’s. You can see the values of usage or not upto. I don’t know what went wrong they are not all working. Even in the log files i get this error

2018-11-30 12:03:05.397 [error] <0.17998.7>@aec_pow_cuckoo:wait_for_result:406 ERROR: GPUassert: out of memory mean.cu 377

2018-11-30 12:03:05.403 [debug] <0.17998.7>@aec_pow_cuckoo:parse_generation_result:479 GeForce GTX 1060 3GB with 3016MB @ 192 bits x 4004MHz
2018-11-30 12:03:05.403 [debug] <0.17998.7>@aec_pow_cuckoo:parse_generation_result:479 Looking for 42-cycle on cuckoo30(“Ek25rhFTSQtK8OBdGjOzAyjfqzdZtWL7O1EGKx58b14=R27HW2B0Pk4=”,0) with 50% edges, 64*64 buckets, 176 trims, and 64 thread blocks.
2018-11-30 12:03:05.446 [error] <0.17999.7>@aec_pow_cuckoo:wait_for_result:406 ERROR: GPUassert: out of memory mean.cu 377

2018-11-30 12:03:05.446 [error] <0.18000.7>@aec_pow_cuckoo:wait_for_result:406 ERROR: GPUassert: out of memory mean.cu 377

2018-11-30 12:03:05.450 [debug] <0.17999.7>@aec_pow_cuckoo:parse_generation_result:479 GeForce GTX 1060 3GB with 3019MB @ 192 bits x 4004MHz
2018-11-30 12:03:05.453 [debug] <0.18000.7>@aec_pow_cuckoo:parse_generation_result:479 GeForce GTX 1060 3GB with 3019MB @ 192 bits x 4004MHz
2018-11-30 12:03:05.453 [debug] <0.17999.7>@aec_pow_cuckoo:parse_generation_result:479 Looking for 42-cycle on cuckoo30(“Ek25rhFTSQtK8OBdGjOzAyjfqzdZtWL7O1EGKx58b14=SG7HW2B0Pk4=”,0) with 50% edges, 6464 buckets, 176 trims, and 64 thread blocks.
2018-11-30 12:03:05.453 [debug] <0.18000.7>@aec_pow_cuckoo:parse_generation_result:479 Looking for 42-cycle on cuckoo30(“Ek25rhFTSQtK8OBdGjOzAyjfqzdZtWL7O1EGKx58b14=SW7HW2B0Pk4=”,0) with 50% edges, 64
64 buckets, 176 trims, and 64 thread blocks.
2018-11-30 12:03:05.498 [error] <0.17998.7>@aec_pow_cuckoo:wait_for_result:421 OS process died: {status,2}
2018-11-30 12:03:05.537 [error] <0.18000.7>@aec_pow_cuckoo:wait_for_result:421 OS process died: {status,2}
2018-11-30 12:03:05.576 [error] <0.17999.7>@aec_pow_cuckoo:wait_for_result:421 OS process died: {status,2}

Can someone help me to fix this issue. I followed the epoch.yaml file given by chris and I validated the file too. I dont know what went wrong.


#68

it seems out of memory? Maybe it doesn’t work on 3gb cards. I don’t know. Maybe dev can clarify.

sidenote, its using 7gb on my 1080ti


#69

So, we cannot mine with 3 GB cards. That is ridiculous. I am mining ethereum on them. Why can’t aeternity? May be do I need to change in epoch.yaml by giving so arguments in extra_args

But this dev’s they don’t respond if they find difficulty in answering the issue.


#70

try to add -E 1 to extra_args


#71

I did and restarted the node still same issue extra_args: “-E 1”. But still same issue.


#72

Here is my epoch.yaml file


#73

try “-E 3” , maybe that works, it will still require 3.4gb but maybe your lucky :stuck_out_tongue:


#74

no luck…:sob: I think the dev’s should answer this issue… and I know they wont reply.


#75

I am a member of the core team, but not a GPU expert I am afraid… However the efficient algorithm for solving the Cuckoo PoW algorithm (the so called mean miner) does unfortunately require quite a bit of memory. That is unfortunate, but as far as I have understood it something that is hard to improve and still keep the miner efficient.


#76

Then what is the use of mining? How come your network grow. Most of the mining operations run with 4 GB GPU’s. If your algorithm take such huge memory then its not worth mining.


#77

I have a same problem before.
This is my log after I addedd " -E 1 to extra_args

Is it working properly now?
(sorry for my bad screencapture)



#78

yes this looks good!


#79

Thx! I totally followed your intructor.
You are the best.


#80

I am trying to mining by following the instruction on https://medium.com/@TwenteMining/deploy-ae-mainnet-cuda-multigpu-miner-23989ad8b1bb

Everything seems good. But I am not sure if I am mining.

How can I check it?