Don't perform pivoting for symmetric Hermitian problems

exanauts · Dec 8, 2023 · 6e6196d · 6e6196d · amontoison · Dec 8, 2023
1 parent 5abc10e
commit 6e6196d
Show file tree

Hide file tree

Showing 4 changed files with 4 additions and 5 deletions.
diff --git a/README.md b/README.md
@@ -10,7 +10,7 @@
 ## Overview
 
 [CUDSS.jl](https://github.com/exanauts/CUDSS.jl) is a Julia interface to the NVIDIA [cuDSS](https://developer.nvidia.com/cudss) library.
-NVIDIA cuDSS provides three factorizations (LU, LDLᵀ, LLᵀ) for solving sparse linear systems on GPUs.
+NVIDIA cuDSS provides three factorizations (LDU, LDLᵀ, LLᵀ) for solving sparse linear systems on GPUs.
 
 ### Why CUDSS.jl?
 
@@ -93,8 +93,6 @@ using CUDA, CUDA.CUSPARSE
 using CUDSS
 using SparseArrays
 
-# Sparse an hermitian positive define linear system
-# with multiple right-hand sides
 T = ComplexF64
 n = 100
 p = 5

diff --git a/docs/src/index.md b/docs/src/index.md
@@ -3,7 +3,7 @@
 ## Overview
 
 [CUDSS.jl](https://github.com/exanauts/CUDSS.jl) is a Julia interface to the NVIDIA [cuDSS](https://developer.nvidia.com/cudss) library.
-NVIDIA cuDSS provides three factorizations (LU, LDLᵀ, LLᵀ) for solving sparse linear systems on GPUs.
+NVIDIA cuDSS provides three factorizations (LDU, LDLᵀ, LLᵀ) for solving sparse linear systems on GPUs.
 For more details on using cuDSS, refer to the official [cuDSS documentation](https://docs.nvidia.com/cuda/cudss/index.html).
 
 ## Installation

diff --git a/src/interfaces.jl b/src/interfaces.jl
@@ -58,7 +58,7 @@ The available configuration parameters are:
 - `"solve_mode"`: Potential modificator on the system matrix (transpose or adjoint);
 - `"ir_n_steps"`: Number of steps during the iterative refinement;
 - `"ir_tol"`: Iterative refinement tolerance;
-- `"pivot_type"`: Type of pivoting ('C', 'R' or 'N');
+- `"pivot_type"`: Type of pivoting (`'C'`, `'R'` or `'N'`);
 - `"pivot_threshold"`: Pivoting threshold which is used to determine if digonal element is subject to pivoting;
 - `"pivot_epsilon"`: Pivoting epsilon, absolute value to replace singular diagonal elements;
 - `"max_lu_nnz"`: Upper limit on the number of nonzero entries in LU factors for non-symmetric matrices.

diff --git a/test/test_cudss.jl b/test/test_cudss.jl
@@ -152,6 +152,7 @@ function cudss_execution()
         config = CudssConfig()
         data = CudssData()
         solver = CudssSolver(matrix, config, data)
+        (structure == 'H') && cudss_set(solver, "pivot_type", 'N')
 
         cudss("analysis", solver, X_gpu, B_gpu)
         cudss("factorization", solver, X_gpu, B_gpu)