This is a new implementation of a chinilla plotter which is designed as a processing pipeline, similar to how GPUs work, only the "cores" are normal software CPU threads.
As a result this plotter is able to fully max out any storage device's bandwidth, simply by increasing the number of "cores", ie. threads.
Join the Discord for support: https://discord.gg/pQwkebKnPB
For <poolkey> and <farmerkey> see output of `chinilla keys show`.
To plot for pools, specify <contract> address via -c instead of <poolkey>, see `chinilla plotnft show`.
<tmpdir> needs about 220 GiB space, it will handle about 25% of all writes. (Examples: './', '/mnt/tmp/')
<tmpdir2> needs about 110 GiB space and ideally is a RAM drive, it will handle about 75% of all writes.
Combined (tmpdir + tmpdir2) peak disk usage is less than 256 GiB.
In case of <count> != 1, you may press Ctrl-C for graceful termination after current plot is finished,
or double press Ctrl-C to terminate immediately.
Usage:
chinilla_plot [OPTION...]
-k, --size arg K size (default = 32, k <= 32)
-x, --port arg Network port (default = 8444, chives = 9699)
-n, --count arg Number of plots to create (default = 1, -1 = infinite)
-r, --threads arg Number of threads (default = 4)
-u, --buckets arg Number of buckets (default = 256)
-v, --buckets3 arg Number of buckets for phase 3+4 (default = buckets)
-t, --tmpdir arg Temporary directory, needs ~220 GiB (default = $PWD)
-2, --tmpdir2 arg Temporary directory 2, needs ~110 GiB [RAM] (default = <tmpdir>)
-d, --finaldir arg Final directory (default = <tmpdir>)
-w, --waitforcopy Wait for copy to start next plot
-p, --poolkey arg Pool Public Key (48 bytes)
-c, --contract arg Pool Contract Address (62 chars)
-f, --farmerkey arg Farmer Public Key (48 bytes)
-G, --tmptoggle Alternate tmpdir/tmpdir2 (default = false)
-D, --directout Create plot directly in finaldir (default = false)
-K, --rmulti2 arg Thread multiplier for P2 (default = 1)
--help Print help
Make sure to crank up <threads> if you have plenty of cores, the default is 4.
Depending on the phase more threads will be launched, the setting is just a multiplier.
RAM usage depends on <threads> and <buckets>.
With the new default of 256 buckets it's about 0.5 GB per thread at most.
-G option will alternate the temp dirs used while plotting to give each one, tmpdir and tmpdir2, equal usage. The first plot creation will use tmpdir and tmpdir2 as expected. The next run, if -n equals 2 or more, will swap the order to tmpdir2 and tmpdir. The next run swaps again to tmpdir and tmpdir2. This will occur until the number of plots created is reached or until stopped.
sudo mount -t tmpfs -o size=110G tmpfs /mnt/ram/
Note: 128 GiB System RAM minimum required for RAM disk.
HCX: hcx1w5c2vv5ak08pczeph7tp5xmkl5762pdf3pyjkg9z4ks4ed55j3psgay0zh
XFX: xfx1succfn2z3uwmq50ukztjanrvs9kw294mqn4lv22rk6tzmcu7e2xsyxyaa5
XCC: xcc16j65y35fs8u289nq6krcyehsmp5eqd4we493rxf36pg7eymcqrqqltsrat
ETH-ERC20: 0x97057cdf529867838d2a1f7f23ba62456764e0cd
LTC: MNUnszsX2srv5EJpu9YYHAXb19MqUpuBjD
BTC: 15GSE5ymStxXMvJ58hyosEVm4FXFxUyJZg
On a dual Xeon® E5-2650v2@2.60GHz R720 with 256GB RAM and a 3x800GB SATA SSD RAID0, using a 110G tmpfs for <tmpdir2>:
Click to expand
Number of Threads: 16
Number of Buckets: 2^8 (256)
Working Directory: /mnt/tmp3/chinilla/tmp/
Working Directory 2: /mnt/tmp3/chinilla/tmp/ram/
[P1] Table 1 took 17.2488 sec
[P1] Table 2 took 145.011 sec, found 4294911201 matches
[P1] Table 3 took 170.86 sec, found 4294940789 matches
[P1] Table 4 took 203.713 sec, found 4294874801 matches
[P1] Table 5 took 201.346 sec, found 4294830453 matches
[P1] Table 6 took 195.928 sec, found 4294681297 matches
[P1] Table 7 took 158.053 sec, found 4294486972 matches
Phase 1 took 1092.2 sec
[P2] max_table_size = 4294967296
[P2] Table 7 scan took 15.5542 sec
[P2] Table 7 rewrite took 37.7806 sec, dropped 0 entries (0 %)
[P2] Table 6 scan took 46.7014 sec
[P2] Table 6 rewrite took 65.7315 sec, dropped 581295425 entries (13.5352 %)
[P2] Table 5 scan took 45.4663 sec
[P2] Table 5 rewrite took 61.9683 sec, dropped 761999997 entries (17.7423 %)
[P2] Table 4 scan took 44.8217 sec
[P2] Table 4 rewrite took 61.36 sec, dropped 828847725 entries (19.2985 %)
[P2] Table 3 scan took 44.9121 sec
[P2] Table 3 rewrite took 61.5872 sec, dropped 855110820 entries (19.9097 %)
[P2] Table 2 scan took 43.641 sec
[P2] Table 2 rewrite took 59.6939 sec, dropped 865543167 entries (20.1528 %)
Phase 2 took 620.488 sec
Wrote plot header with 268 bytes
[P3-1] Table 2 took 73.1018 sec, wrote 3429368034 right entries
[P3-2] Table 2 took 42.3999 sec, wrote 3429368034 left entries, 3429368034 final
[P3-1] Table 3 took 68.9318 sec, wrote 3439829969 right entries
[P3-2] Table 3 took 43.8179 sec, wrote 3439829969 left entries, 3439829969 final
[P3-1] Table 4 took 71.3236 sec, wrote 3466027076 right entries
[P3-2] Table 4 took 46.2887 sec, wrote 3466027076 left entries, 3466027076 final
[P3-1] Table 5 took 70.6369 sec, wrote 3532830456 right entries
[P3-2] Table 5 took 45.5857 sec, wrote 3532830456 left entries, 3532830456 final
[P3-1] Table 6 took 75.8534 sec, wrote 3713385872 right entries
[P3-2] Table 6 took 48.8266 sec, wrote 3713385872 left entries, 3713385872 final
[P3-1] Table 7 took 83.2586 sec, wrote 4294486972 right entries
[P3-2] Table 7 took 56.3803 sec, wrote 4294486972 left entries, 4294486972 final
Phase 3 took 733.323 sec, wrote 21875928379 entries to final plot
[P4] Starting to write C1 and C3 tables
[P4] Finished writing C1 and C3 tables
[P4] Writing C2 table
[P4] Finished writing C2 table
Phase 4 took 84.6697 sec, final plot size is 108828428322 bytes
Total plot creation time was 2530.76 sec
To make sure the plots are valid you can use the ProofOfSpace tool from chinillapos:
git clone https://github.com/Chinilla/chinillapos.git
cd chinillapos && mkdir build && cd build && cmake .. && make -j8
./ProofOfSpace check -f plot-k32-???.plot [num_iterations]cd chinilla-plotter
git checkout master
git pull
./make_devel.shI do have some history with GPU mining, back in 2014 I was the first to open source a XPM GPU miner, which was about 40x more efficient than the CPU miner. See my other repos.
As such, it's only a matter of time until I add OpenCL support to speed up the plotter even more, keeping most of the load off the CPUs.
- cmake (>=3.14)
Windows
Binaries built by stotiks can be found here: https://github.com/stotiks/chinilla-plotter/releases
Arch Linux
First, install dependencies from pacman:
sudo pacman -S cmake gmp gcc11Then, clone and compile the project:
# Checkout the source and install
git clone https://github.com/Chinilla/chinilla-plotter-madmax.git
cd chinilla-plotter
./make_devel.sh
./build/chinilla_plot --helpCentOS 7
git clone https://github.com/Chinilla/chinilla-plotter-madmax.git
cd chinilla-plotter
sudo yum install epel-release -y
sudo yum install cmake3 -y
ln /usr/bin/cmake3 /usr/bin/cmake
# Install a package with repository for your system:
# On CentOS, install package centos-release-scl available in CentOS repository:
sudo yum install centos-release-scl -y
# Install the collection:
sudo yum install devtoolset-7 -y
# Start using software collections:
scl enable devtoolset-7 bash
./make_devel.sh
./build/chinilla_plot --helpClear Linux
sudo swupd update
sudo swupd bundle-add c-basic git
echo PATH=$PATH:/usr/local/bin/ # for statically compiled cmake if not already in your PATH
# Checkout the source and install
cd ~/
git clone https://github.com/Chinilla/chinilla-plotter-madmax.git
cd ~/chinilla-plotter
./make_devel.sh
./build/chinilla_plot --helpUbuntu 20.04
sudo apt install -y cmake g++ git build-essential
# Checkout the source and install
git clone https://github.com/Chinilla/chinilla-plotter-madmax.git
cd chinilla-plotter
./make_devel.sh
./build/chinilla_plot --helpThe binaries will end up in build/, you can copy them elsewhere freely (on the same machine, or similar OS).
Debian 10 ("buster")
Make sure to add buster-backports to your sources.list otherwise the installation will fail because an older cmake version. See the debian backport documentation for reference.
# Install cmake 3.16 from buster-backports
sudo apt install -t buster-backports cmake
sudo apt install -y g++ git
# Checkout the source and install
git clone https://github.com/Chinilla/chinilla-plotter-madmax.git
cd chinilla-plotter
./make_devel.sh
./build/chinilla_plot --helpThe binaries will end up in build/, you can copy them elsewhere freely (on the same machine, or similar OS).
macOS
First you need to install the Brew package manager and Xcode OR Xcode Command Line Tools.
# Download Xcode Command Line Tools (skip if you already have Xcode)
xcode-select --install
# Now download chinilla-plotter's dependencies
brew install cmake git
brew link cmake
git clone https://github.com/Chinilla/chinilla-plotter-madmax.git
cd chinilla-plotter
./make_devel.sh
./build/chinilla_plot --helpIf a maximum open file limit error occurs (as default OS setting is 256, which is too low for default bucket size of 256), run this before starting the plotter
ulimit -n 3000
This file limit change will only affect the current session.
Running in a Docker container
In some setups and scenarios, it could be useful to run your plotter inside a Docker container. This could be potentially useful while running chinilla-plotter in Windows.
To do so, install Docker on your computer and them run the following command:
docker run \
-v <path-to-your-tmp-dir>:/mnt/harvester \
-v <path-to-your-final-dir>:/mnt/farm \
odelucca/chinilla-plotter \
-t /mnt/harvester/ \
-d /mnt/farm/ \
-p <pool-key> \
-f <farm-key> \
-r <number-of-CPU-cores>💡 You can provide any of the plotter arguments after the image name (
odelucca/chinilla-plotter)
In a Linux benchmark, we were able to find that running in Docker is only 5% slower than running in native OS.
For Windows users, you should check if your Docker configuration has any RAM or CPU limits. Since Docker runs inside HyperV, that could potentially constrain your hardware usage. In any case, you can set the RAM limits with the -m flag (after the docker run command).
While running in Windows, you may need to proper configure your Docker to allow multi CPUs. You can do so by following this article
In a nutshell, you could also pass the --cpus flag to your docker run command in order to achieve the same result.
So, for example, the following command...
docker run \
-v <path-to-your-tmp-dir>:/mnt/harvester \
-v <path-to-your-final-dir>:/mnt/farm \
-m 8G \
--cpus 8 \
odelucca/chinilla-plotter \
-t /mnt/harvester/ \
-d /mnt/farm/ \
-p <pool-key> \
-f <farm-key> \
-r 8...would run your plotter with 8 CPU cores and 8GB of RAM.
Simply build with docker build .
- Needs at least cmake 3.14 (because of bls-signatures)