@Patrick Are you able to comment on the difference in hashrate (if any) between running a miner inside a Docker container vs outside? Specifically I am thinking of your dwarfpool Monero image vs say xmr-stak-cpu. This might not be an apples to apples comparison, but still. Just curious to know if there is a price to be paid for the convenience of being able to deploy an image via a swarm...
New Dwarfpool Monero Mining Docker Image
- Thread starter Patrick
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Wolf's and stak are different miners. Stak will be better if you have a low number of machines. Once you get to dozens of different configurations (and using them part time) you need to use something that can easily scale out and losing low single digits percentage is usually worth it.
If I had one machine mining 24x7, I would use stak. If I wanted to learn docker and scale an operation, then moving to a container solution helps. Containerizing Wolf's loses less than 0.5% over bare metal which is a testing variation.
If I had one machine mining 24x7, I would use stak. If I wanted to learn docker and scale an operation, then moving to a container solution helps. Containerizing Wolf's loses less than 0.5% over bare metal which is a testing variation.
Hi, re-installing to the latest docker version fixed this problem for me. Good luck.
@Patrick what you think about it? GPU & CPU BENCHMARKS FOR MONERO MINING!
this claymore cpu miner look better than usual wolf's miner...
this claymore cpu miner look better than usual wolf's miner...
In my tests minergate has almost the same hashrate/s than wolf's,
These test by this guy achieve a better h/s with claymore windows thans minergate/wolf's....
"i7-2600k Win7 Ultimate 64-bit:
- 150 H/s using MinerGate
- 250 H/s using Claymore Cryptonote CPU miner v3.5 (4 threads used)
i7-6800k Win7 Pro 64-bit:
- 260 H/s using MinerGate
- 360 H/s using Claymore Cryptonote CPU miner v3.5 (7 threads used)"
Ahh, thanks. There might have been some issues that were fixed after the website maintenance last night. Seems to work now. Hash speed checks out on the webpage.
Trying the -e threads option with the nproc image gave a huge boost! Using 11 threads (for a dual X5660 DL380 G7) , has rate went up to 420 H/s compared to 220 H/s on latest image. Thanks for adding that Patrick!
To reduce power costs, I should probably pull most of the RAM out of the box. Really don't need all 144GB of ram for a test ESXi environment and mining on the side. Would switching to L5640 CPUs make that much of a difference?
Trying the -e threads option with the nproc image gave a huge boost! Using 11 threads (for a dual X5660 DL380 G7) , has rate went up to 420 H/s compared to 220 H/s on latest image. Thanks for adding that Patrick!
To reduce power costs, I should probably pull most of the RAM out of the box. Really don't need all 144GB of ram for a test ESXi environment and mining on the side. Would switching to L5640 CPUs make that much of a difference?
@Patrick when you next update the image, can you please consider adding the ability to set a fixed difficulty. I note that for my OCP nodes, difficulty starts off at 100K or so, and gradually decreases to 20K. It would be nice to be able to set that as the starting value instead.
Yes, Windows 10 x64, Ryzen 5 1600, 16 GB Ram, latest version of docker.
Thanks for your reply.
Ah maybe that is the reason. Docker on Windows is running in a VM... likely that is screwing up the thread counts. These miners work best on bare metal. Docker on Windows uses a Hyper-V VM to run Linux and then run containers atop of that.
@Amir - try using https://forums.servethehome.com/index.php?threads/docker-xmrig-cryptonight-universal.18579/
The old dwarfpool image uses a github clone to pull source code before compiling. To avoid these issues, we are cloning code into the Docker image itself now so there is no external network dependency for source code download.
The old dwarfpool image uses a github clone to pull source code before compiling. To avoid these issues, we are cloning code into the Docker image itself now so there is no external network dependency for source code download.