| Products & Services | Industries | Academia | Support | User Community | Company |
| Download Product Updates | | | Get Pricing | | | Trial Software |
| Documentation → MATLAB |
| Contents | Index |
| Learn more about MATLAB |
New features and changes introduced in this version are:
MATLAB's new multithreading capability now includes:
MATLAB features significant performance improvements for:
Sparse matrix indexed assignment and deletion
mrdivide (/)
The factorization routine in qr produces an upper triangular matrix, R. Now this matrix always contains real and nonnegative diagonal elements. In previous releases, the diagonal of R could contain complex and negative elements.
R now contains only real, nonnegative diagonal elements. The QR factorization is not unique, so the answer is still correct.
Now subscripted reference into a sparse matrix always returns a sparse matrix. In previous versions of MATLAB, using a double scalar to index into a sparse matrix resulted in full scalar output.
Scalars extricated from sparse matrices are no longer full. If you previously used the output with a function that does not support sparse input, you now need to update your code. For a list of the functions that do not support sparse matrices, see Functions That Do Not Support Sparse Matrices in the MATLAB Mathematics documentation.
Functions that erroneously accepted N-D arrays and returned reshaped sparse 2-D outputs no longer accept full N-D inputs and now return an error. The functions affected are:
Also, binary functions that formerly accepted both full N-D inputs along with sparse 2-D inputs and warned about the reshape, now return an error. For example, ones(2,2,2).*sparse(3) formerly returned a sparse 2-by-4 matrix along with a warning about the reshape. This code now returns an error.
The computational geometry functions delaunay, convhull, griddata, voronoi, delaunay3, griddata3 no longer use the options arguments.
Use of the options arguments to delaunay, convhull, griddata, and voronoi now throws a warning.
The functions griddata3, delaunay3, dsearch, tsearch will be removed in a future release. Use of these functions now throws a warning.
| Function Name | Use This Method Instead |
|---|---|
| griddata3 | TriScatteredInterp |
| dsearch | DelaunayTri/nearestNeighbor |
| tsearch | DelaunayTri/pointLocation |
| delaunay3 | DelaunayTri |
Update your code to use the DelaunayTri and TriScatteredInterp computational geometry classes .
The lsqnonneg solver no longer uses the optional input x0 as a starting point.
If you give a starting point x0, MATLAB issues a warning. Also, lsqnonneg ignores x0, and instead uses a starting point of a vector of zeros.
Support for the erfcore function has been removed. Use of erfcore now results in the following error message:
ERFCORE will be removed in a future release. Use ERF, ERFC, ERFCX, ERFINV, or ERFCINV instead.
Replace erfcore with erf, erfc, erfcx, erfinv, or erfcinv. See the function reference pages for the individual error functions.
These warning messages for integer math and conversion have been removed:
MATLAB:intConvertNaN
MATLAB:intConvertNonIntVal
MATLAB:intConvertOverflow
MATLAB:intMathOverflow
The warning messages for integer math and conversion are no longer available. Remove these warning IDs from your code.
The intwarning function will be removed in a future release. Use of intwarning now throws a warning:
Warning: All four integer warnings are removed. INTWARNING will be removed in a future release.
The warnings previously thrown when you used intwarning on are now removed. Remove all instances of intwarning from your code.
The warning thrown when you use atan(i) and atan(-i) will be removed. Currently, the warning message is:
Warning: Singularity in ATAN. This warning will be removed in a future release.Consider using DBSTOP IF NANINF when debugging.
Remove instances of the warning ID MATLAB:atan:singularity from your code.
In previous releases, nextpow2 with a vector input v returned nextpow2(length(v)). Now nextpow2(v) returns a vector the same size as the input.
If you require the old behavior, replace instances of nextpow2 with nextpow2(length(v)).
MATLAB no longer provides the libdflapack.dll and libdfblas.dll math libraries for building Fortran MEX-files with a Compaq® Visual Fortran compiler.
If you distribute MEX-files that call BLAS or LAPACK functions built on a Compaq Visual Fortran compiler, you must also distribute the libdflapack.dll and libdfblas.dll library files.
If MATLAB displays errors when loading a MEX-file and the Dependency Walker indicates missing libdflapack.dll and/or libdfblas.dll libraries, contact the MEX-file vendor for copies of the libraries. The third-party product, Dependency Walker, is available from the Web site http://www.dependencywalker.com/.
![]() | Desktop Tools and Development Environment, MATLAB Version 7.10 (R2010a) | Data Analysis, MATLAB Version 7.10 (R2010a) | ![]() |

Includes the most popular MATLAB recorded presentations with Q&A sessions led by MATLAB experts.
| © 1984-2010- The MathWorks, Inc. - Site Help - Patents - Trademarks - Privacy Policy - Preventing Piracy - RSS |