For Developers

This page contains tips and tricks for developers, further resources, along with information on how to set up your build environment on your platform.

Before building Nimbus for the first time, make sure to install the prerequisites.

Code style

The code follows the Status Nim Style Guide.

Branch lifecycle

The git repository has 3 main branches, stable, testing and unstable as well as feature and bugfix branches.


The unstable branch contains features and bugfixes that are actively being tested and worked on.

  • Features and bugfixes are generally pushed to individual branches, each with their own pull request against the unstable branch.
  • Once the branch has been reviewed and passed CI, the developer or reviewer merges the branch to unstable.
  • The unstable branch is regularly deployed to the Nimbus prater and pyrmont fleets where additional testing happens.


The testing branch contains features and bugfixes that have gone through CI and initial testing on the unstable branch and are ready to be included in the next release.

  • After testing a bugfix or feature on unstable, the features and fixes that are planned for the next release get merged to the testing branch either by the release manager or team members.
  • The testing branch is regularly deployed to the Nimbus prater and pyrmont fleet as well as a smaller mainnet fleet.
  • The branch should remain release-ready at most times.


The stable branch tracks the latest released version of Nimbus and is suitable for mainnet staking.

Build system


mingw32-make # this first invocation will update the Git submodules

You can now follow the instructions in this this book by replacing make with mingw32-make (you should run mingw32 regardless of whether you're running 32-bit or 64-bit architecture):

mingw32-make test # run the test suite

Linux, macOS

After cloning the repo:

# Build nimbus_beacon_node and all the tools, using 4 parallel Make jobs
make -j4

# Run tests
make test

# Update to latest version
git pull
make update


Nimbus comes with a build environment similar to Python venv - this helps ensure that the correct version of Nim is used and that all dependencies can be found.

./ bash # start a new interactive shell with the right env vars set
which nim
nim --version # Nimbus is tested and supported on 1.0.2 at the moment

# or without starting a new interactive shell:
./ which nim
./ nim --version

# Start Visual Studio code with environment
./ code

Makefile tips and tricks for developers

  • build all those tools known to the Makefile:
# $(nproc) corresponds to the number of cores you have
make -j $(nproc)
  • build a specific tool:
make state_sim
  • you can control the Makefile's verbosity with the V variable (defaults to 0):
make V=1 # verbose
make V=2 test # even more verbose
make LOG_LEVEL=DEBUG bench_bls_sig_agggregation # this is the default
make LOG_LEVEL=TRACE nimbus_beacon_node # log everything
  • pass arbitrary parameters to the Nim compiler:
make NIMFLAGS="-d:release"
  • you can freely combine those variables on the make command line:
make -j$(nproc) NIMFLAGS="-d:release" USE_MULTITAIL=yes eth2_network_simulation
make USE_LIBBACKTRACE=0 # expect the resulting binaries to be 2-3 times slower
  • disable -march=native because you want to run the binary on a different machine than the one you're building it on:
make NIMFLAGS="-d:disableMarchNative" nimbus_beacon_node
  • disable link-time optimisation (LTO):
make NIMFLAGS="-d:disableLTO" nimbus_beacon_node
  • build a static binary
make NIMFLAGS="--passL:-static" nimbus_beacon_node
  • publish a book using mdBook from sources in "docs/" to GitHub pages:
make publish-book
  • create a binary distribution
make dist
  • test the binaries
make dist-test

Multi-client interop scripts

This repository contains a set of scripts used by the client implementation teams to test interop between the clients (in certain simplified scenarios). It mostly helps us find and debug issues.

Stress-testing the client by limiting the CPU power

make prater CPU_LIMIT=20

The limiting is provided by the cpulimit utility, available on Linux and macOS. The specified value is a percentage of a single CPU core. Usually 1 - 100, but can be higher on multi-core CPUs.

Build and run the local beacon chain simulation

The beacon chain simulation runs several beacon nodes on the local machine, attaches several local validators to each, and builds a beacon chain between them.

To run the simulation:

make update
make eth2_network_simulation

To clean the previous run's data:

make clean_eth2_network_simulation_all

To change the number of validators and nodes:

# Clear data files from your last run and start the simulation with a new genesis block:
make VALIDATORS=192 NODES=6 USER_NODES=1 eth2_network_simulation

If you’d like to see the nodes running on separated sub-terminals inside one big window, install Multitail (if you're on a Mac, follow the instructions here), then:

USE_MULTITAIL="yes" make eth2_network_simulation

You’ll get something like this (click for full size):

You can find out more about the beacon node simulation here.

Build and run the local state transition simulation

This simulation is primarily designed for researchers, but we'll cover it briefly here in case you're curious :)

The state transition simulation quickly runs the beacon chain state transition function in isolation and outputs JSON snapshots of the state (directly to the nimbus-eth2 directory). It runs without networking and blocks are processed without slot time delays.

# build the state simulator, then display its help ("-d:release" speeds it
# up substantially, allowing the simulation of longer runs in reasonable time)
make NIMFLAGS="-d:release" state_sim
build/state_sim --help

Use the output of the help command to pass desired values to the simulator - experiment with changing the number of slots, validators, , etc. to get different results.

The most important options are:

  • slots : the number of slots to run the simulation for (default 192)
  • validators: the number of validators (default 6400)
  • attesterRatio: the expected fraction of attesters that actually do their work for every slot (default 0.73)
  • json_interval: how often JSON snapshots of the state are outputted (default every 32 slots -- or once per epoch)

For example, to run the state simulator for 384 slots, with 20,000 validators, and an average of 66% of attesters doing their work every slot, while outputting snapshots of the state twice per epoch, run:

build/state_sim --slots=384 --validators=20000 --attesterRatio=0.66 --json_interval=16