diff --git a/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Complex.cs b/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Complex.cs
index ace87e43..f7a83813 100644
--- a/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Complex.cs
+++ b/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Complex.cs
@@ -2174,6 +2174,7 @@ namespace MathNet.Numerics.Algorithms.LinearAlgebra
         /// right singular vectors.</param>
         /// <param name="work">The work array. Length should be at least <paramref name="rowsA"/>.</param>
         /// <remarks>This is equivalent to the GESVD LAPACK routine.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         public virtual void SingularValueDecomposition(bool computeVectors, Complex[] a, int rowsA, int columnsA, Complex[] s, Complex[] u, Complex[] vt, Complex[] work)
         {
             if (a == null)
@@ -2559,7 +2560,7 @@ namespace MathNet.Numerics.Algorithms.LinearAlgebra
                 // If too many iterations have been performed throw exception.
                 if (iter >= Maxiter)
                 {
-                    throw new ArgumentException(Resources.ConvergenceFailed);
+                    throw new NonConvergenceException();
                 }
 
                 // This section of the program inspects for negligible elements in the s and e arrays,  
diff --git a/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Complex32.cs b/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Complex32.cs
index 038ce62d..5cff5024 100644
--- a/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Complex32.cs
+++ b/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Complex32.cs
@@ -2172,6 +2172,7 @@ namespace MathNet.Numerics.Algorithms.LinearAlgebra
         /// right singular vectors.</param>
         /// <param name="work">The work array. Length should be at least <paramref name="rowsA"/>.</param>
         /// <remarks>This is equivalent to the GESVD LAPACK routine.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         public virtual void SingularValueDecomposition(bool computeVectors, Complex32[] a, int rowsA, int columnsA, Complex32[] s, Complex32[] u, Complex32[] vt, Complex32[] work)
         {
             if (a == null)
@@ -2557,7 +2558,7 @@ namespace MathNet.Numerics.Algorithms.LinearAlgebra
                 // If too many iterations have been performed throw exception.
                 if (iter >= Maxiter)
                 {
-                    throw new ArgumentException(Resources.ConvergenceFailed);
+                    throw new NonConvergenceException();
                 }
 
                 // This section of the program inspects for negligible elements in the s and e arrays,  
diff --git a/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Double.cs b/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Double.cs
index a6f28baa..423c51e3 100644
--- a/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Double.cs
+++ b/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Double.cs
@@ -2059,6 +2059,7 @@ namespace MathNet.Numerics.Algorithms.LinearAlgebra
         /// right singular vectors.</param>
         /// <param name="work">The work array. Length should be at least <paramref name="rowsA"/>.</param>
         /// <remarks>This is equivalent to the GESVD LAPACK routine.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         public virtual void SingularValueDecomposition(bool computeVectors, double[] a, int rowsA, int columnsA, double[] s, double[] u, double[] vt, double[] work)
         {
             if (a == null)
@@ -2446,7 +2447,7 @@ namespace MathNet.Numerics.Algorithms.LinearAlgebra
                 // If too many iterations have been performed throw exception.
                 if (iter >= Maxiter)
                 {
-                    throw new ArgumentException(Resources.ConvergenceFailed);
+                    throw new NonConvergenceException();
                 }
 
                 // This section of the program inspects for negligible elements in the s and e arrays,  
diff --git a/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Single.cs b/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Single.cs
index dc32bdd4..dbdbfce7 100644
--- a/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Single.cs
+++ b/src/Numerics/Algorithms/LinearAlgebra/ManagedLinearAlgebraProvider.Single.cs
@@ -2060,6 +2060,7 @@ namespace MathNet.Numerics.Algorithms.LinearAlgebra
         /// <param name="vt">If <paramref name="computeVectors"/> is <c>true</c>, on exit VT contains the transposed
         /// right singular vectors.</param>
         /// <param name="work">The work array. Length should be at least <paramref name="rowsA"/>.</param>
+        /// <exception cref="NonConvergenceException"></exception>
         public virtual void SingularValueDecomposition(bool computeVectors, float[] a, int rowsA, int columnsA, float[] s, float[] u, float[] vt, float[] work)
         {
             if (a == null)
@@ -2449,7 +2450,7 @@ namespace MathNet.Numerics.Algorithms.LinearAlgebra
                 // If too many iterations have been performed throw exception.
                 if (iter >= Maxiter)
                 {
-                    throw new ArgumentException(Resources.ConvergenceFailed);
+                    throw new NonConvergenceException();
                 }
 
                 // This section of the program inspects for negligible elements in the s and e arrays,  
diff --git a/src/Numerics/LinearAlgebra/Complex/Factorization/DenseEvd.cs b/src/Numerics/LinearAlgebra/Complex/Factorization/DenseEvd.cs
index 8a2c1b0f..9b68ab59 100644
--- a/src/Numerics/LinearAlgebra/Complex/Factorization/DenseEvd.cs
+++ b/src/Numerics/LinearAlgebra/Complex/Factorization/DenseEvd.cs
@@ -226,6 +226,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex.Factorization
         /// Bowdler, Martin, Reinsch, and Wilkinson, Handbook for
         /// Auto. Comp., Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         internal static void SymmetricDiagonalize(System.Numerics.Complex[] dataEv, double[] d, double[] e, int order)
         {
             const int Maxiter = 1000;
@@ -324,7 +325,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex.Factorization
                         // throw exception that Convergence Failed
                         if (iter >= Maxiter)
                         {
-                            throw new ArgumentException(Resources.ConvergenceFailed);
+                            throw new NonConvergenceException();
                         }
                     } while (Math.Abs(e[l]) > eps*tst1);
                 }
diff --git a/src/Numerics/LinearAlgebra/Complex/Factorization/UserEvd.cs b/src/Numerics/LinearAlgebra/Complex/Factorization/UserEvd.cs
index b8d6bed5..94369877 100644
--- a/src/Numerics/LinearAlgebra/Complex/Factorization/UserEvd.cs
+++ b/src/Numerics/LinearAlgebra/Complex/Factorization/UserEvd.cs
@@ -246,6 +246,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex.Factorization
         /// Bowdler, Martin, Reinsch, and Wilkinson, Handbook for
         /// Auto. Comp., Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         private void SymmetricDiagonalize(double[] d, double[] e, int order)
         {
             const int Maxiter = 1000;
@@ -344,7 +345,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex.Factorization
                         // throw exception that Convergence Failed
                         if (iter >= Maxiter)
                         {
-                            throw new ArgumentException(Resources.ConvergenceFailed);
+                            throw new NonConvergenceException();
                         }
                     }
                     while (Math.Abs(e[l]) > eps * tst1);
diff --git a/src/Numerics/LinearAlgebra/Complex/Factorization/UserSvd.cs b/src/Numerics/LinearAlgebra/Complex/Factorization/UserSvd.cs
index 0d5c1e1c..22ce6cd6 100644
--- a/src/Numerics/LinearAlgebra/Complex/Factorization/UserSvd.cs
+++ b/src/Numerics/LinearAlgebra/Complex/Factorization/UserSvd.cs
@@ -57,7 +57,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex.Factorization
         /// <param name="matrix">The matrix to factor.</param>
         /// <param name="computeVectors">Compute the singular U and VT vectors or not.</param>
         /// <exception cref="ArgumentNullException">If <paramref name="matrix"/> is <c>null</c>.</exception>
-        /// <exception cref="ArgumentException">If SVD algorithm failed to converge with matrix <paramref name="matrix"/>.</exception>
+        /// <exception cref="NonConvergenceException"></exception>
         public UserSvd(Matrix<Complex> matrix, bool computeVectors)
         {
             if (matrix == null)
@@ -356,7 +356,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex.Factorization
                 // throw exception that Convergence Failed
                 if (iter >= Maxiter)
                 {
-                     throw new ArgumentException(Resources.ConvergenceFailed);
+                    throw new NonConvergenceException();
                 }
 
                 // This section of the program inspects for negligible elements in the s and e arrays. On
diff --git a/src/Numerics/LinearAlgebra/Complex32/Factorization/DenseEvd.cs b/src/Numerics/LinearAlgebra/Complex32/Factorization/DenseEvd.cs
index 880a326a..96dea823 100644
--- a/src/Numerics/LinearAlgebra/Complex32/Factorization/DenseEvd.cs
+++ b/src/Numerics/LinearAlgebra/Complex32/Factorization/DenseEvd.cs
@@ -227,6 +227,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex32.Factorization
         /// Bowdler, Martin, Reinsch, and Wilkinson, Handbook for
         /// Auto. Comp., Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         internal static void SymmetricDiagonalize(Numerics.Complex32[] dataEv, float[] d, float[] e, int order)
         {
             const int Maxiter = 1000;
@@ -325,7 +326,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex32.Factorization
                         // throw exception that Convergence Failed
                         if (iter >= Maxiter)
                         {
-                            throw new ArgumentException(Resources.ConvergenceFailed);
+                            throw new NonConvergenceException();
                         }
                     } while (Math.Abs(e[l]) > eps*tst1);
                 }
diff --git a/src/Numerics/LinearAlgebra/Complex32/Factorization/UserEvd.cs b/src/Numerics/LinearAlgebra/Complex32/Factorization/UserEvd.cs
index d91773ac..30ebf305 100644
--- a/src/Numerics/LinearAlgebra/Complex32/Factorization/UserEvd.cs
+++ b/src/Numerics/LinearAlgebra/Complex32/Factorization/UserEvd.cs
@@ -250,6 +250,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex32.Factorization
         /// Bowdler, Martin, Reinsch, and Wilkinson, Handbook for
         /// Auto. Comp., Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         private void SymmetricDiagonalize(float[] d, float[] e, int order)
         {
             const int Maxiter = 1000;
@@ -348,7 +349,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex32.Factorization
                         // throw exception that Convergence Failed
                         if (iter >= Maxiter)
                         {
-                            throw new ArgumentException(Resources.ConvergenceFailed);
+                            throw new NonConvergenceException();
                         }
                     }
                     while (Math.Abs(e[l]) > eps * tst1);
diff --git a/src/Numerics/LinearAlgebra/Complex32/Factorization/UserSvd.cs b/src/Numerics/LinearAlgebra/Complex32/Factorization/UserSvd.cs
index 14b55a83..c9aa3ecd 100644
--- a/src/Numerics/LinearAlgebra/Complex32/Factorization/UserSvd.cs
+++ b/src/Numerics/LinearAlgebra/Complex32/Factorization/UserSvd.cs
@@ -57,7 +57,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex32.Factorization
         /// <param name="matrix">The matrix to factor.</param>
         /// <param name="computeVectors">Compute the singular U and VT vectors or not.</param>
         /// <exception cref="ArgumentNullException">If <paramref name="matrix"/> is <c>null</c>.</exception>
-        /// <exception cref="ArgumentException">If SVD algorithm failed to converge with matrix <paramref name="matrix"/>.</exception>
+        /// <exception cref="NonConvergenceException"></exception>
         public UserSvd(Matrix<Complex32> matrix, bool computeVectors)
         {
             if (matrix == null)
@@ -356,7 +356,7 @@ namespace MathNet.Numerics.LinearAlgebra.Complex32.Factorization
                 // throw exception that Convergence Failed
                 if (iter >= Maxiter)
                 {
-                     throw new ArgumentException(Resources.ConvergenceFailed);
+                    throw new NonConvergenceException();
                 }
 
                 // This section of the program inspects for negligible elements in the s and e arrays. On
diff --git a/src/Numerics/LinearAlgebra/Double/Factorization/DenseEvd.cs b/src/Numerics/LinearAlgebra/Double/Factorization/DenseEvd.cs
index 100e4292..6ef6cae2 100644
--- a/src/Numerics/LinearAlgebra/Double/Factorization/DenseEvd.cs
+++ b/src/Numerics/LinearAlgebra/Double/Factorization/DenseEvd.cs
@@ -256,6 +256,7 @@ namespace MathNet.Numerics.LinearAlgebra.Double.Factorization
         /// Bowdler, Martin, Reinsch, and Wilkinson, Handbook for
         /// Auto. Comp., Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         internal static void SymmetricDiagonalize(double[] a, double[] d, double[] e, int order)
         {
             const int maxiter = 1000;
@@ -354,7 +355,7 @@ namespace MathNet.Numerics.LinearAlgebra.Double.Factorization
                         // throw exception that Convergence Failed
                         if (iter >= maxiter)
                         {
-                            throw new ArgumentException(Resources.ConvergenceFailed);
+                            throw new NonConvergenceException();
                         }
                     } while (Math.Abs(e[l]) > eps*tst1);
                 }
@@ -528,6 +529,7 @@ namespace MathNet.Numerics.LinearAlgebra.Double.Factorization
         /// by Martin and Wilkinson, Handbook for Auto. Comp.,
         /// Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         internal static void NonsymmetricReduceHessenberToRealSchur(double[] a, double[] matrixH, double[] d, double[] e, int order)
         {
             // Initialize
@@ -732,7 +734,7 @@ namespace MathNet.Numerics.LinearAlgebra.Double.Factorization
                     iter = iter + 1;
                     if (iter >= 30*order)
                     {
-                        throw new ArgumentException(Resources.ConvergenceFailed);
+                        throw new NonConvergenceException();
                     }
 
                     // Look for two consecutive small sub-diagonal elements
diff --git a/src/Numerics/LinearAlgebra/Double/Factorization/UserEvd.cs b/src/Numerics/LinearAlgebra/Double/Factorization/UserEvd.cs
index 3f0a5315..a60af1ad 100644
--- a/src/Numerics/LinearAlgebra/Double/Factorization/UserEvd.cs
+++ b/src/Numerics/LinearAlgebra/Double/Factorization/UserEvd.cs
@@ -288,6 +288,7 @@ namespace MathNet.Numerics.LinearAlgebra.Double.Factorization
         /// Bowdler, Martin, Reinsch, and Wilkinson, Handbook for
         /// Auto. Comp., Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         private void SymmetricDiagonalize(double[] d, double[] e, int order)
         {
             const int Maxiter = 1000;
@@ -386,7 +387,7 @@ namespace MathNet.Numerics.LinearAlgebra.Double.Factorization
                         // throw exception that Convergence Failed
                         if (iter >= Maxiter)
                         {
-                            throw new ArgumentException(Resources.ConvergenceFailed);
+                            throw new NonConvergenceException();
                         }
                     }
                     while (Math.Abs(e[l]) > eps * tst1);
diff --git a/src/Numerics/LinearAlgebra/Double/Factorization/UserSvd.cs b/src/Numerics/LinearAlgebra/Double/Factorization/UserSvd.cs
index eb69bf8c..5f673d5f 100644
--- a/src/Numerics/LinearAlgebra/Double/Factorization/UserSvd.cs
+++ b/src/Numerics/LinearAlgebra/Double/Factorization/UserSvd.cs
@@ -56,7 +56,7 @@ namespace MathNet.Numerics.LinearAlgebra.Double.Factorization
         /// <param name="matrix">The matrix to factor.</param>
         /// <param name="computeVectors">Compute the singular U and VT vectors or not.</param>
         /// <exception cref="ArgumentNullException">If <paramref name="matrix"/> is <c>null</c>.</exception>
-        /// <exception cref="ArgumentException">If SVD algorithm failed to converge with matrix <paramref name="matrix"/>.</exception>
+        /// <exception cref="NonConvergenceException"></exception>
         public UserSvd(Matrix<double> matrix, bool computeVectors)
         {
             if (matrix == null)
@@ -341,7 +341,7 @@ namespace MathNet.Numerics.LinearAlgebra.Double.Factorization
                 // throw exception that Convergence Failed
                 if (iter >= Maxiter)
                 {
-                     throw new ArgumentException(Resources.ConvergenceFailed);
+                    throw new NonConvergenceException();
                 }
 
                 // This section of the program inspects for negligible elements in the s and e arrays. On
diff --git a/src/Numerics/LinearAlgebra/Single/Factorization/DenseEvd.cs b/src/Numerics/LinearAlgebra/Single/Factorization/DenseEvd.cs
index 3e86aa98..1ac111e9 100644
--- a/src/Numerics/LinearAlgebra/Single/Factorization/DenseEvd.cs
+++ b/src/Numerics/LinearAlgebra/Single/Factorization/DenseEvd.cs
@@ -256,6 +256,7 @@ namespace MathNet.Numerics.LinearAlgebra.Single.Factorization
         /// Bowdler, Martin, Reinsch, and Wilkinson, Handbook for
         /// Auto. Comp., Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         internal static void SymmetricDiagonalize(float[] a, float[] d, float[] e, int order)
         {
             const int Maxiter = 1000;
@@ -354,7 +355,7 @@ namespace MathNet.Numerics.LinearAlgebra.Single.Factorization
                         // throw exception that Convergence Failed
                         if (iter >= Maxiter)
                         {
-                            throw new ArgumentException(Resources.ConvergenceFailed);
+                            throw new NonConvergenceException();
                         }
                     } while (Math.Abs(e[l]) > eps*tst1);
                 }
@@ -528,6 +529,7 @@ namespace MathNet.Numerics.LinearAlgebra.Single.Factorization
         /// by Martin and Wilkinson, Handbook for Auto. Comp.,
         /// Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         internal static void NonsymmetricReduceHessenberToRealSchur(float[] a, float[] matrixH, float[] d, float[] e, int order)
         {
             // Initialize
@@ -732,7 +734,7 @@ namespace MathNet.Numerics.LinearAlgebra.Single.Factorization
                     iter = iter + 1;
                     if (iter >= 30*order)
                     {
-                        throw new ArgumentException(Resources.ConvergenceFailed);
+                        throw new NonConvergenceException();
                     }
 
                     // Look for two consecutive small sub-diagonal elements
diff --git a/src/Numerics/LinearAlgebra/Single/Factorization/UserEvd.cs b/src/Numerics/LinearAlgebra/Single/Factorization/UserEvd.cs
index 3053bae0..9246b8a9 100644
--- a/src/Numerics/LinearAlgebra/Single/Factorization/UserEvd.cs
+++ b/src/Numerics/LinearAlgebra/Single/Factorization/UserEvd.cs
@@ -289,6 +289,7 @@ namespace MathNet.Numerics.LinearAlgebra.Single.Factorization
         /// Bowdler, Martin, Reinsch, and Wilkinson, Handbook for
         /// Auto. Comp., Vol.ii-Linear Algebra, and the corresponding
         /// Fortran subroutine in EISPACK.</remarks>
+        /// <exception cref="NonConvergenceException"></exception>
         private void SymmetricDiagonalize(float[] d, float[] e, int order)
         {
             const int Maxiter = 1000;
@@ -387,7 +388,7 @@ namespace MathNet.Numerics.LinearAlgebra.Single.Factorization
                         // throw exception that Convergence Failed
                         if (iter >= Maxiter)
                         {
-                            throw new ArgumentException(Resources.ConvergenceFailed);
+                            throw new NonConvergenceException();
                         }
                     }
                     while (Math.Abs(e[l]) > eps * tst1);
diff --git a/src/Numerics/LinearAlgebra/Single/Factorization/UserSvd.cs b/src/Numerics/LinearAlgebra/Single/Factorization/UserSvd.cs
index 7dc2a343..c7dfafa9 100644
--- a/src/Numerics/LinearAlgebra/Single/Factorization/UserSvd.cs
+++ b/src/Numerics/LinearAlgebra/Single/Factorization/UserSvd.cs
@@ -56,7 +56,7 @@ namespace MathNet.Numerics.LinearAlgebra.Single.Factorization
         /// <param name="matrix">The matrix to factor.</param>
         /// <param name="computeVectors">Compute the singular U and VT vectors or not.</param>
         /// <exception cref="ArgumentNullException">If <paramref name="matrix"/> is <c>null</c>.</exception>
-        /// <exception cref="ArgumentException">If SVD algorithm failed to converge with matrix <paramref name="matrix"/>.</exception>
+        /// <exception cref="NonConvergenceException"></exception>
         public UserSvd(Matrix<float> matrix, bool computeVectors)
         {
             if (matrix == null)
@@ -341,7 +341,7 @@ namespace MathNet.Numerics.LinearAlgebra.Single.Factorization
                 // throw exception that Convergence Failed
                 if (iter >= Maxiter)
                 {
-                     throw new ArgumentException(Resources.ConvergenceFailed);
+                    throw new NonConvergenceException();
                 }
 
                 // This section of the program inspects for negligible elements in the s and e arrays. On