AOCL-Sparse version 5.0

• New APIs added:
  • Level 3: sp2md, spmmd, syrk, syrkd, sypr, syprd
  • Sparse preconditioner: sorv
• Support for:
  • Symmetric and Hermitian matrices in the csrmm API
  • Strided dense vector in TRSM and TRSV
  • Symmetric Gauss Seidel
• Performance improvements:
  • Level 1: dot, gthr, sctr, and roti
  • Level 2: SpMV for complex general matrices
• Multi-Threading support
  • Level 3: Sp2M, SpMM, and SpAdd
• Improvements to Benchmarking framework
• Enhanced statistics, support for new APIs, and random matrix generation capability (Hermitian and diagonally dominant)
• APIs to support HPCG

AOCL-Sparse version 4.2

• Support for one-based indexing
• Support for complex datatypes
• Improved support for operations on matrices (transpose, conjugate transpose and none) and descriptor interpretations.
• New APIs added:

• Level 1 APIs: aoclsparse_?gthr, aoclsparse_?gthrz, aoclsparse_?gthrs, aoclsparse_?sctr, aoclsparse_?sctrs, aoclsparse_?dotci, aoclsparse_?dotui, aoclsparse_?doti, aoclsparse_?roti, and aoclsparse_?axpyi
• Level 2 API: aoclsparse_?dotmv
• Level 3 APIs: aoclsparse_?csrmm, aoclsparse_sp2m, aoclsparse_?add, and aoclsparse_?trsm
• Auxiliary APIs: aoclsparse_copy, aoclsparse_order_mat, aoclsparse_export_?csr, aoclsparse_export_?csc, and aoclsparse_convert_csr

• Enhancements for TRSV's kernel templates
• Framework enhancements and bug fixes
• Integration with AOCL-Utils for detecting supported ISA

AOCL-Sparse version 4.1

Highlights of AOCL-Sparse 4.1

• Dynamic dispatch support for SpMV on AMD “Zen” architectures
• APIs to support HPCG
• Test suite enhancements
• Reusable kernel templates to enable quick design and implementation of level 1 and level 2 kernels
• Integration with AOCL-BLAS and AOCL-LAPACK for level 1 APIs
• Minor bug fixes

AOCL-Sparse version 4.0

Highlights of AOCL-Sparse 4.0

• New Iterative Solver APIs: CG, GMRES, and Gauss-Seidel (pre-conditioner)
• AVX512 implementation for SPMV API
• Improved performance of the following:

• TRSV
• Multi-thread SPMV

AOCL-Sparse version 3.2

Highlights of improvements on AMD EPYCTM processor family CPUs

• New API for multiplying two Sparse matrices (aoclsparse_xcsr2m)
• New API aoclsparse_xilu_smoother that acts as a preconditioner to compute an update to the iterative solution x of Ax=b
• Improved performance of single thread SPMV routine that supports hint and optimize functions analyzing the sparsity pattern for better optimization
• Multi-thread support for SPMV routine

AOCL-Sparse version 3.1

Highlights of improvements on AMD EPYC^TM processor family CPUs

• New API for Sparse Matrix Dense Matrix Multipy (SPMM) giving Dense Matrix output.
  • Supports Single and Double Precision data types.
  • Supports General sparse matrices in CSR format with zero-based indexing and no transpose.
• New conversion routine: CSR to Dense which converts sparse matrix in CSR format to dense matrix.
• Test and sample usage examples for new APIs

AOCL-Sparse version 3.0

AOCL-Sparse version 3.0

Highlights of improvements on AMD EPYC^TMprocessor family CPUs

• Supports CSR, Ellpack, Diagonal, Blocked-CSR data formats for SPMV function
• New API, Sparse Triangular Solve(TRSV) for Single and Double Precision data types
• Supports General matrices with zero-based indexing and no transpose.
• New Sparse data format conversion routines:
  • CSR to Ellpack
  • CSR to Diagonal
  • CSR to Blocked-CSR
  • CSR to CSC

AOCL-Sparse version 2.2

AOCL-Sparse version 2.2

Highlights of improvements on AMD EPYC^TMprocessor family CPUs

• Includes Sparse Matrix Vector Multipy(SPMV) API for Single and Double Precision data types
• Supports CSR and Ellpack data formats for SPMV function
• Supports General matrices with zero-based indexing and no transpose.
• Test and sample usage examples