addHeader("Matrix Template Library 4"); $t->addFile("Beta-1 [1][2]", "mtl4-beta-1-r6916.tar.gz", "76456a777b52d20da25e7ed7152807fb"); $t->addFile("Beta-1 [1][2]", "mtl4-beta-1-r6916.zip", "4133c76052fe5ecb76dc47fe561faa4e"); $t->addFile("Alpha-2 [3]", "mtl4-alpha-2-r6668.tar.gz", "6b1175ac2427650bce2bdbc4179801be"); $t->addFile("Alpha-2 [3]", "mtl4-alpha-2-r6668.zip", "0bc0ee4406ba73f4474a0fe1981c52e5"); $t->addFile("Alpha-1 [4][5]", "mtl4-alpha-1-r6418.tar.gz", "0995b7fd3d9a48327e50e6eaf3de6af7"); $t->addFile("Alpha-1 [4][5]", "mtl4-alpha-1-r6418.zip", "6ede39100eaf31e1d7aa66965687e016"); $t->addFile("Alpha-1 [6]", "mtl4-alpha-1-r6398.tar.gz", "0cd0d5a57fbc9835feb76529a729757a"); $t->printMe(); ?>

1. Tested on gcc 4.1.2 and MSVC 2008. There are problems with Intel C++ 9.1 and 10.1 (helpful suggestions are welcome).
2. Compiled with paranoid warnings on gcc: -Werror -Wreorder -pedantic -Wpointer-arith -Wcast-align -Wcast-qual -Wwrite-strings
3. Tested on gcc 4.0.1, gcc 4.1.1, gcc 4.1.2, gcc 4.2.0, gcc 4.2.1, gcc 4.2.2, gcc 4.3.2, Intel C++ 9.0, MSVC 2005, and MSVC 2008 (so far).
4. Matrix products of more than two factors enabled, e.g., A+= B * C * D (not yet available on MSVC).
5. Memory alignment behaves less dangerously now.
6. We thank Torsten Höfler for setting up the configuration of BLAS in scons and Bill Greene for his help to port MTL 4 to MS VC 8.0. Hui Li from Princeton implemented the division by scalars and the multiplication from right by scalars.

Alternatively you can check out the repository if you wish a more convenient manner to update the library. The trunk of the MTL repository is publicly available as read-only via subversion. The command line to check out is:

svn co https://svn.osl.iu.edu/tlc/trunk/mtl4/trunk mtl4

Disclaimer: Please be aware that the repository might be sometimes broken. We will do our best to avoid this or at least warn you if we suspect trouble. Thus, if you experience any problem please read the log file and roll back to the last (allegedly) stable version.