cuSPARSE

cuSPARSE Library - Generic APIs Examples

Description

This folder demonstrates cuSPARSE Generic APIs usage.

cuSPARSE Generic APIs Documentation

cuSPARSE Samples

• Vector - Vector Operations
• Matrix - Vector Operations
• Matrix - Matrix Operations
• Conversion
• Legacy APIs
• Optimizations
• Sparse Iterative Methods

---

Vector - Vector Operations

• cusparseAxpby

  The sample demonstrates Sparse Vector - Dense Vector scaling and sum

• cusparseGather

  The sample demonstrates Dense Vector to Sparse Vector element gathering

• cusparseRot

  The sample demonstrates Sparse Vector - Dense Vector Givens rotation

• cusparseScatter

  The sample demonstrates Sparse Vector to Dense Vector element scattering

• cusparseSpVV

  The sample demonstrates Sparse Vector - Dense Vector dot product

Matrix - Vector Operations

• cusparseSpMV CSR

  The sample demonstrates Sparse Matrix - Dense Vector multiplication, where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseSpMV COO

  The sample demonstrates Sparse Matrix - Dense Vector multiplication, where the sparse matrix is represented in COO (Coordinate) storage format

• cusparseSpSV CSR

  The sample demonstrates Sparse triangular solver with single right-hand side, where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseSpSV COO

  The sample demonstrates Sparse triangular solver with single right-hand side, where the sparse matrix is represented in COO (Coordinate) storage format

Matrix - Matrix Operations

• cusparseSpMM CSR

  The sample demonstrates Sparse Matrix - Dense Matrix multiplication = Dense Matrix, where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseSpMM CSR - Batched

  The sample demonstrates Batched Sparse Matrix - Dense Matrix multiplication = Dense Matrix , where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseSpMMOp CSR

  The sample demonstrates Sparse Matrix - Dense Matrix multiplication = Dense Matrix with Custom Operators, where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseSpMM COO

  The sample demonstrates Sparse Matrix - Dense Matrix multiplication = Dense Matrix, where the sparse matrix is represented in COO (Coordinate) storage format

• cusparseSpMM COO - Batched

  The sample demonstrates Batched Sparse Matrix - Dense Matrix multiplication = Dense Matrix, where the sparse matrix is represented in COO (Coordinate) storage format

• cusparseSpMM SDDMM

  The sample demonstrates Dense Matrix - Dense Matrix multiplication = Sparse Matrix, where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseSpMM Blocked-ELL

  The sample demonstrates Sparse Matrix - Dense Matrix multiplication = Dense Matrix, where the sparse matrix is represented in Blocked-ELL storage format

• cusparseSpMM SDDMM - Batched The sample demonstrates Batched Dense matrix - Dense matrix multiplication = Sparse matrix, where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseSpGEMM

  The sample demonstrates Sparse Matrix - Sparse Matrix multiplication = Sparse Matrix, where all operands are sparse matrices represented in CSR (Compressed Sparse Row) storage format

• cusparseSpGEMM - Memory Optmization

  The sample demonstrates Sparse Matrix - Sparse Matrix multiplication = Sparse Matrix, where all operands are sparse matrices represented in CSR (Compressed Sparse Row) storage format and the memory usage is optimzed

• cusparseSpGEMM_reuse

  The sample demonstrates Sparse Matrix - Sparse Matrix multiplication = Sparse Matrix, where all operands are sparse matrices represented in CSR (Compressed Sparse Row) storage format and the structure of the output matrix can be reused multiple times

• cusparseSpSM CSR

  The sample demonstrates Sparse triangular solver with multiple right-hand sides, where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseSpSM COO

  The sample demonstrates Sparse triangular solver with multiple right-hand sides, where the sparse matrix is represented in COO (Coordinate) storage format

Conversion

• cusparseSparseToDense

  The sample demonstrates Sparse Matrix to Dense Matrix conversion, where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseDenseToSparse

  The sample demonstrates Dense Matrix to Sparse Matrix conversion, where the sparse matrix is represented in CSR (Compressed Sparse Row) storage format

• cusparseDenseToSparse (Blocked-ELL)

  The sample demonstrates Dense Matrix to Sparse Matrix conversion, where the sparse matrix is represented in Blocked-Ellpack storage format

Legacy APIs

• cusparseXcoosortByRow

  The sample demonstrates how to sort a COO format matrix

• cusparseSgpsvInterleavedBatch

  The sample demonstrates how to solves two penta-diagonal systems with NOT interleaved format

Optimizations

• CUDA Graph Capture

  The sample demonstrates how to optimize Sparse vector - Dense vector dot product (cusparseSpVV) by exploiting CUDA Graph Capture functionality

• Hardware Memory Compression

  The sample demonstrates how to optimize Sparse vector - Dense vector scaling and sum (cusparseAxpby) by exploiting NVIDIA Ampere architecture Hardware Memory Compression

Sparse Iterative Methods

• Preconditioned CG

  The sample describes how to use the cuSPARSE and cuBLAS libraries to implement the Incomplete-Cholesky preconditioned iterative Conjugate Gradient (CG)

• Preconditioned BiCGStab

  The sample describes how to use the cuSPARSE and cuBLAS libraries to implement the Incomplete-LU preconditioned iterative Biconjugate Gradient Stabilized Method (BiCGStab)