MAGMA Downloads
MAGMA provides implementations for CUDA, HIP, Intel Xeon Phi, and OpenCL. The latest releases are MAGMA 2.7.1 for CUDA and HIP, MAGMA MIC 1.4.0 for Intel Xeon Phi, and clMAGMA 1.3 for OpenCL. The libraries available for download are listed below in the order of their release dates.
Please use any of the following publications to reference MAGMA.
MAGMA Bitbucket repository: https://bitbucket.org/icl/magma.
|
|
MAGMA 2.7.1
|
2023-02-22
|
MAGMA 2.7.1 is now released. This release includes:
- Add support for CUDA 12;
- Add a new interface for batch GEMV that accepts a pointer + stride;
- Add sparse test matrices to the release tarball;
- Performance improvement for batch GEMV targeting square sizes up to 32;
- Update CMakeLists compiler flags for Windows.
|
|
|
|
MagmaDNN 1.4
|
2022-12-01
|
MagmaDNN 1.4 is now available. MagnaDNN provides HP data analytics and machine learning tools using MAGMA as its computational backend. Updates in this release include:
- Bug fixes and performance improvements;
- Added ResnetBlocks for Resnet 18/34 models;
- Added Concat Operations for 3D and 4D layers;
- Added Unet Blocks for Unet models;
- Added LSTM functionality;
- Added OpenCV compatability.
MagmaDNN's repository is on Bitbucket: https://bitbucket.org/icl/magmadnn. |
|
|
|
MAGMA MIC 1.4.0
|
2015-07-12
|
MAGMA MIC 1.4.0 is now available. This release provides implementations for MAGMA's one-sided (LU, QR, and Cholesky) and two-sided (Hessenberg, bi- and tridiagonal reductions) dense matrix factorizations, as well as linear and eigenproblem solver for Intel Xeon Phi Coprocessors. More information on the approach is given in this presentation. |
|
|
|
clMAGMA 1.3
|
2014-11-14
|
clMAGMA is an OpenCL port of MAGMA. It supports AMD GPUs. The clMAGMA library dependancies, in particular optimized GPU OpenCL BLAS and CPU optimized BLAS and LAPACK for AMD hardware, can be found in the AMD clMath Libraries (formerly APPML).
Included in the clMAGMA 1.3 release are routines for the following algorithms:
- LU, QR, and Cholesky factorizations in both real and complex arithmetic (single and double);
- Linear and least squares solvers based on correspondingly the LU/Cholesky and QR factorizations in both real and complex arithmetic (single and double);
- Reductions to Hessenberg, bidiagonal, and tridiagonal forms using orthgonal similarity transformationsin both real and complex arithmetic (single and double);
- Eigen and singular value problem solvers in both real and complex arithmetic (single and double);
- Orthogonal transformation routines.
|
|
|
License
Copyright © 2023 The University of Tennessee. All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
· Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
· Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer listed in this license in the documentation and/or other materials provided with the distribution.
· Neither the name of the copyright holders nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
This software is provided by the copyright holders and contributors "as is" and any express or implied warranties, including, but not limited to, the implied warranties of merchantability and fitness for a particular purpose are disclaimed. in no event shall the copyright owner or contributors be liable for any direct, indirect, incidental, special, exemplary, or consequential damages (including, but not limited to, procurement of substitute goods or services; loss of use, data, or profits; or business interruption) however caused and on any theory of liability, whether in contract, strict liability, or tort (including negligence or otherwise) arising in any way out of the use of this software, even if advised of the possibility of such damage.
