Skip to content

Latest commit

 

History

History
38 lines (28 loc) · 1.18 KB

README.md

File metadata and controls

38 lines (28 loc) · 1.18 KB

Using llama to accelerate your build

llama is a CLI for outsourcing computation to AWS Lambda.

You can use llama to accelerate your CDDA builds. To help you set that up, this directory contains a suitable image for compiling CDDA on.

After bootstrapping llama, you can use for example

llama update-function --create --timeout=120s --memory=4096 --build=tools/llama/gcc-focal gcc

This will configure llama to use the gcc-focal image here. Note that it's important to have larger-than-default timeout and memory settings.

Troubleshooting

If, when building, you get errors about failed includes of SDL2 headers, then you can work around these via a hack I discovered, by setting a carefully chosen CXXFLAGS. Your build command might look something like this:

CXXFLAGS=-I/x/../../../../usr/include/SDL2 CXX=llamac++ make PCH=0 TILES=1 -j100

Or, if using cmake, something like this:

cmake \
    -DTILES=ON \
    -DSOUND=ON \
    -DCMAKE_C_COMPILER=llamacc \
    -DCMAKE_CXX_COMPILER=llamac++ \
    -DCMAKE_CXX_FLAGS="-I/x/../../../../usr/include/SDL2" \
    ..

(/x has to be a directory that doesn't exist on your system)