CHANGELOG

v0.5.0

• Added Docker (instructions at the bottom of the page)

v0.4.0

• Added CHANGELOG section
• Added hash rate calculations (reports ~10s)
• Fixed bug that caused new jobs to periodically not register properly
• Added command line arguments for nloops and difficulty to fine tune miner (see options section)

Requirements

• Git
• CMake
• Python 3.12
• Cuda 12.6

Install instructions

git clone https://github.com/theeldermillenial/tuna-py 
cd tuna-py 
git submodule init 
git submodule update 
cmake . make 
pip install -e . 

Running the miner

1. Edit sample.env file
   • comment out SEED= -> #SEED=
   • change ADDRESS to your own mainnet wallet address
   • change STRATUM_HOST to 66.228.34.31
2. Rename sample.env to .env
   • mv sample.env .env
3. Run tuna miner
   • python -m tuna

Options

Options to help tune hash rate performance.

--nloops 4096

This is the number of hash loops the CUDA miner runs. The default is 4096, which will cause the miner to hash for ~2s on a GTX 1080ti. I recommend tuning this number so that the card hashes for ~1-2 seconds.

How do you tune the card? In .env, set TUNA_LOG=DEBUG and run with different values for --nloops. If the default (4096) only causes your card to run for 1 second, increase to 8192 and try again.

Why 2s? That seems to give roughly the best hash rate given the overhead. The GPU hasher is designed to find and hold up to 40 nonces, so it will report back up to 40 nonces after those 2s.

What if your card starts returning more than 40 nonces? Then set the --difficulty higher (see below).

--difficulty 8

This is the hash difficulty (leading zeros) required to submit to Stratum. This is not the new block mining difficulty. The difficulty should be 7 or higher, and defaults to 8. Ideally this is set so that every 1-3 hash rounds returns 1 nonce. Submitting higher difficulty hashes gives you more hash power on Stratum. I wouldn't recommend setting this higher than the current hash difficulty (at the time of writing, the difficulty is 10).

Docker

I have built a docker container to make it easier to get started, since there is a lot to download, install and configure. Make sure you have Docker and the Nvidia cuda runtime for Docker (if you're on Linux, on Windows it comes packaged with Docker Desktop).

Testing

By default the container has my mining address stored in it. You can donate and test if it works, you can just run it with all defaults:

docker run --gpus 0 eldermillenial/tuna-py:0.5.0

NOTE: The --gpus 0 indicates that the container should run and use the first GPU on your machine. You can run this container multiple times with different GPUs selected.

This should give you an output like this after a few minutes:

31-Aug-24 17:31:55 - tuna     - INFO     - tuna-py v0.5.0 by Elder Millenial
31-Aug-24 17:31:55 - tuna     - INFO     - Address: addr1q9dfupytkpdzqrkmp664vgjneelgh0yvwkqkx9dccyyw5r96h2p5jcgwnv4tw5tq3yzd2dmh3sgcgfyta3tv8x3vdq8qsc8jza
31-Aug-24 17:31:55 - tuna     - INFO     - Stratum Target: 66.228.34.31:3643
31-Aug-24 17:31:55 - tuna     - INFO     - Stratum Worker: HOME
31-Aug-24 17:31:55 - tuna     - INFO     - Submit Difficulty: 8
31-Aug-24 17:31:55 - tuna     - INFO     - Number of CUDA Loops: 4096
31-Aug-24 17:31:56 - tuna     - INFO     - New job: 00007f2a, (0.000 Mh/s, submissions=0, time=1.000s),
31-Aug-24 17:31:56 - tuna     - INFO     - Difficulty: 7
31-Aug-24 17:32:00 - tuna     - INFO     - Submitting nonce: 20000300f77320e050150000, hash=00000000488ce19bc39962a3b312e2669d3c94102a24017737e10b7ccee36743, address=addr1q9dfupytkpdzqrkmp664vgjneelgh0yvwkqkx9dccyyw5r96h2p5jcgwnv4tw5tq3yzd2dmh3sgcgfyta3tv8x3vdq8qsc8jza, worker=HOME
31-Aug-24 17:32:02 - tuna     - INFO     - Submitting nonce: c97700006117ab649a1a0000, hash=00000000477d86da8110ca98eeae62ab98a93146f1f9ea246ab00c2b213ef800, address=addr1q9dfupytkpdzqrkmp664vgjneelgh0yvwkqkx9dccyyw5r96h2p5jcgwnv4tw5tq3yzd2dmh3sgcgfyta3tv8x3vdq8qsc8jza, worker=HOME
31-Aug-24 17:32:06 - tuna     - INFO     - 420.913 Mh/s
31-Aug-24 17:32:09 - tuna     - INFO     - Submitting nonce: 54bc030033ec87d0a0030000, hash=00000000748881a3fcf0656d75a8f9871dc9061c00d149c0f8344ee0b88999d6, address=addr1q9dfupytkpdzqrkmp664vgjneelgh0yvwkqkx9dccyyw5r96h2p5jcgwnv4tw5tq3yzd2dmh3sgcgfyta3tv8x3vdq8qsc8jza, worker=HOME
31-Aug-24 17:32:11 - tuna     - INFO     - Submitting nonce: cf9f0100891a9008221a0000, hash=00000000da0385fda846010b69aa74f34bf00cabc23f037c2151c3aab0295a32, address=addr1q9dfupytkpdzqrkmp664vgjneelgh0yvwkqkx9dccyyw5r96h2p5jcgwnv4tw5tq3yzd2dmh3sgcgfyta3tv8x3vdq8qsc8jza, worker=HOME

Configuration

There are two types of parameters you can configure:

1. Environment Variables
2. Tool parameters

The environment variables allow you to set your mining address and worker name just like you would with the environment file. To set environment variables, use -e KEY=VALUE. For example, to set the address and worker name, you would do:

docker run --gpus 0 \
   -e ADDRESS=addr1q9dfupytkpdzqrkmp664vgjneelgh0yvwkqkx9dccyyw5r96h2p5jcgwnv4tw5tq3yzd2dmh3sgcgfyta3tv8x3vdq8qsc8jza \
   -e STRATUM_WORKER=HOME \
   eldermillenial/tuna-py:0.5.0

For tool parameters like --nloops, you can just add them to the end of the docker command:

docker run --gpus 0 eldermillenial/tuna-py:0.5.0 --nloops 4096 --difficulty 8