CandidateVectorSearch 1.7.2
Searching for peptide candidates using sparse matrix + matrix/vector multiplication.
|
This is a rough guide for compiling CandidateVectorSearch on Ubuntu, which should generally work on most Linux-based systems.
The following dependencies are required for compilation:
sudo apt-get install -y gcc g++
sudo apt-get install -y dotnet-sdk-6.0
sudo apt-get install -y dotnet-runtime-6.0
Dockerfile
to see a full list!Alternatively, a Docker image with all necessary build tools is also available via Dockerfile
:
docker build . -f Dockerfile -t cvs
docker run -it cvs
To build the C++ library used for matrix multiplication the following steps are required:
VectorSearch
directory: cd VectorSearch
g++ -shared -I eigen-3.4.0 -fPIC -fopenmp -O3 -o VectorSearch.dll dllmainUnix.cpp
To build the C# testing application the following steps need to be carried out:
DataLoader
directory: cd DataLoader
dotnet publish DataLoader.csproj --runtime ubuntu.22.04-x64 --self-contained --configuration Release
dotnet publish DataLoader.csproj --runtime linux-x64 --self-contained --configuration Release
]VectorSearch.dll
to the build directory of DataLoader
../DataLoader EigenSIntB 10000
.VectorSearchCUDA
yourself!We provide compiled binaries in DataLoaderExecutable/ubuntu22.04_x64
and in the Releases
tab. Please again note that we only supply the compiled binaries for CPU-based search. If you want to run GPU-based searches please compile VectorSearchCUDA
yourself! Running the compiled binaries requires dotnet-runtime-6.0
and g++
.