# libdasp

**Repository Path**: ssslab/libdasp

## Basic Information

- **Project Name**: libdasp
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: AGPL-3.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2026-02-16
- **Last Updated**: 2026-02-16

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# libDASP

This library is based on the original DASP implementation (https://github.com/SuperScientificSoftwareLaboratory/DASP) and provides a cleaner API for external integration and reuse.

## Installation

To better reproduce experiment results, we suggest an NVIDIA GPU with compute capability 8.0. DASP evaluation requires the CUDA GPU driver, the nvcc CUDA compiler, and the cuSPARSE library, all of them are included with the CUDA Toolkit. 

## Build and Installation

Build the shared library only (default):

```
cmake -S . -B build
cmake --build build
```

Install `libdasp.so` and public headers:

```
cmake --install build --prefix build/install
```

Build test executables:

```
cmake -S . -B build -DBUILD_TESTS=ON
cmake --build build
```

## Execution

Our test programs currently support input files encoded using the matrix market format. All matrix market datasets used in this evaluation are publicly available from the SuiteSparse Matrix Collection.

1. Run code on matrix data. Running the program requires one parameter: matrix path.

`./spmv_double matrix.mtx`

## Tests (CTest)

When building with `-DBUILD_TESTS=ON`, tests can be run via CTest. Provide a matrix path using `MATRIX_PATH`.

```
cmake -S . -B build -DBUILD_TESTS=ON -DMATRIX_PATH=/path/to/matrix.mtx
cmake --build build
ctest --test-dir build
```