Fluid Corpus Manipulation: SuperCollider Objects Library

This repository hosts code for generating the SC objects and documentation resources for the Fluid Corpus Manipulation Project. Much of the actual code that does the exciting stuff lives in this repository's principal dependency, the Fluid Corpus Manipulation Library. What lives here are:

• A wrapper from our code to the SC API, that allows us to generate SC objects from a generic class.
• Stubs for producing SC objects for each 'client' in the Fluid Corpus Manipulation Library.
• CMake code for managing dependencies, building and packaging.

I'm in a Hurry...

...and you already have a development environment set up, understand CMake, and have the SC source available?

Cool:

cmake -DSC_PATH=<location of your SC source> ..
make 

You can either symbolically link the release-packaging directory into your SC extensions folder, or use CMake to generate a 'clean package' with make install.

Pre-requisites

• CMake >= 3.11
• A C++ 14 compliant compiler for Mac or Windows (via XCode tools on Mac, and Visual Studio 17 >= 15.9 on Windows)

Dependencies

• SC Source Code: this is the only dependency we don't (optionally) manage for you, so there must be a version available to point to when you run, using the CMake Variable SC_PATH (see below). It can live anywhere on your file system.

These will be downloaded and configured automatically, unless you pass CMake a source code location on disk for each (see below):

• Fluid Corpus Manipulation Library
• Eigen (3.3.5)
• HISSTools Library

Unless you are simultaneously committing changes to these dependencies (or are seriously worried about disk space), there is nothing to be gained by pointing to external copies, and the easiest thing to is let CMake handle it all.

Building

Simplest possible build:

1. Download the SuperCollider source (>= 3.10.0) if you haven't already
2. Clone this repo

git clone <whereis this>

3. Change to the directory for this repo, make a build directory, and run CMake, passing in the location for the Max SDK

mkdir -p build && cd build 
cmake -DSC_PATH=<location of your SC source> ..

At this point, CMake will set up your tool chain (i.e. look for compilers), and download all the dependencies.

An alternative to setting up / running CMake directly on the command line is to install the CMake GUI, or use to use the curses gui ccmake.

With CMake you have a choice of which build system you use.

• The default on Mac is Unix Makefiles, but you can use Xcode by passing -GXcode to CMake when you first run it.
• The default on Windows is the latest version of Visual Studio installed. However, Visual Studio can open CMake files directly as projects, which has some upsides. When used this way, CMake variables have to be set via a JSON file that MSVC will use to configure CMake.

When using make, then

make

will compile all objects ('targets', in CMake-ish). Alternatively, in Xcode or Visual Studio, running 'build' will (by default) build all objects. Note that these IDEs allow you to build both for debugging and for release, whereas with Makefiles, you need to re-run CMake and pass a CMAKE_CONFIG variable for different build types.

make install 

Will assemble a clean package in release-packaging/FluidCorpusManipulation.

Structure:

The top-level folders mostly correspond to those you would find in a typical Max package. Some more will appear when you compile (such as externals and possibly docs).

The main action is in source:

source
├── include
├── src
└── scripts

• include contains the header files for the FluCoMa-SC wrapper
• src contains a subfolder for each object to be generated; each of these subfolders has a .cpp stub and a CMakeLists.txt. The .cpp file name needs to match its parent folder.
• scripts contains CMake scripts (most significantly, target_post.cmake, which sets behaviours for all the objects).

Development: Manual Dependencies

If you're making changes to the Fluid Corpus Manipulation Library and testing against Max (and, hopefully, our other deployment environments), then it makes sense to have the source for this cloned somewhere else on your disk, so that you can commit and push changes, and ensure that they work in all environments. This would be the case, e.g., if you're developing a new client.

To bypass the automatic cloning of the Fluid Corpus Manipulation Library, we pass in the cache variable FLUID_PATH

cmake -DSC_PATH=<location of your SC source> -DFLUID_PATH=<location of Fluid Corpus Manipulation Library> ..

Note

1. You don't need to run CMake on the Fluid Corpus Manipulation Library first (well, you can, but it doesn't make any difference!). CMake's FetchContent facility will grab the targets from there, but won't look in its CMakeCache, so there should never be a conflict between the state of a build tree at FLUID_PATH and your build-tree here.
2. It is up to you to make sure the commits you have checked out in each repository make sense together. We use tags against release versions on the master branch, so – at the very least – these should line up (unless you're tracking down some regression bug or similar). In general, there is no guarantee, or likelihood, that mismatched tags will build or run, as architectural changes can, do, will happen...

Dependencies of dependencies!

The same steps and considerations apply to manually managing the dependencies of the Fluid Corpus Manipulation Library itself. If these aren't explicitly passed whilst running CMake against this build tree, then CMake will download them itself against the tags / commits we develop against. Nevertheless, if you are in the process of making changes to these libraries and running against this (which is much less likely than above), then the CMake variables of interest are:

• EIGEN_PATH pointing to the location of Eigen on disk
• HISS_PATH pointing to the location of the HISSTools Library

To find out which branches / tags / commits of these we use, look in the top level CMakeLists.txt of the Fluid Corpus Manipulation Library for the FetchContent_Declare statements for each dependency.

Vector instructions and CPU architecture.

Most types of CPU in common use support a range of specialised instructions for processing vectors of numbers at once, which can offer substantial speed gains. For instance, on Intel / AMD chips there have been a succession of such instruction sets (MMX, SSE, AVX and so on).

If you find objects causing 'illegal instruction' segmentation faults, it is likely that our default vector instruction flag isn't supported by your CPU.

I'm in a hurry...

And only building for your own machine? You want to enable the maximal set of CPU features for your machine without worrying? Using Clang / GCC?

Pass -DFLUID_ARCH=-mnative when you run CMake. This tells the compiler to figure out what instruction sets your personal CPU supports, and enable them. This implies a possible performance gain in return for a portability loss.

More Detail

By default, on Intel / AMD, we enable AVX instructions (-mavx on clang/GCC; /ARCH:AVX on MSVC). These should work on all relatively recent CPUs (from, say, 2012 onwards). On Arm (with the Bela in mind), we use the following architecture flags:

-march=armv7-a -mtune=cortex-a8 -mfloat-abi=hard -mfpu=neon

If your needs are different, then these defaults can be overridden by passing the desired compiler flags to CMake via the FLUID_ARCH cache variable. Note that this variable expects you to pass arguments in a form suitable for your compiler, and currently performs no checking or validation.

• Clang and GCC options are generally the same, see https://clang.llvm.org/docs/ClangCommandLineReference.html
• See MSVC options here https://docs.microsoft.com/en-us/cpp/build/reference/arch-minimum-cpu-architecture?view=vs-2019

If your Intel / AMD chip is too old to support AVX, it probably still supports SSE. On Mac OS and Linux, sysctl -a | grep cpu.feat can be used to produce a list of the various features your CPU supports.

This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 725899).