Search results
Results from the WOW.Com Content Network
Download as PDF; Printable version; ... Tensor software is a class of mathematical software designed for ... [41] is a lightweight distributed-parallel tensor library ...
Finite element software for structural, geotechnical, heat transfer and seepage analysis: Intuition Software: 5.11: 2016-01: Proprietary software: Free educational version available [17] Mac OS X, Windows: JCMsuite: Finite element software for the analysis of electromagnetic waves, elasticity and heat conduction: JCMwave GmbH: 5.4.3: 2023-03-09 ...
The software runs on Unix and Windows platforms and can be compiled on a large variety of compilers, using the CMake building tool. The solver can also be used in a multi-host parallel mode on platforms that support MPI. Elmer's parallelisation capability is one of the strongest sides of this solver.
General purpose, includes 2D and 3D magnetics solvers, both static and harmonic. 3D solver is based on the Whitney AV formulation of Maxwell's equations. VSimEM: Commercial Yes Yes Yes Yes Yes Automatic, variable mesh FDTD, PIC, finite volume: Simulating electromagnetics, and electrostatics in complex dielectric and metallic environments.
The argument of methods symmetrize() and antisymmetrize() in tensor classes is now directly a sequence of index positions (and no longer a single list/tuple encapsulating such a sequence). Method self_contract() of tensor classes renamed trace(). The code for tensor contractions has been optimized; moreover multiple tensor contractions are now ...
MUMPS (MUltifrontal Massively Parallel sparse direct Solver) is a software application for the solution of large sparse systems of linear algebraic equations on distributed memory parallel computers. It was developed in European project PARASOL (1996–1999) by CERFACS, IRIT-ENSEEIHT and RAL.
Released under the GNU General Public License, Cadabra is free software. Cadabra has extensive functionality for tensor polynomial simplification including multi-term symmetries, fermions and anti-commuting variables, Clifford algebras and Fierz transformations , implicit coordinate dependence, multiple index types and many more.
The library is designed to reduce computing power and memory use and to train large distributed models with better parallelism on existing computer hardware. [2] [3] DeepSpeed is optimized for low latency, high throughput training. It includes the Zero Redundancy Optimizer (ZeRO) for training models with 1 trillion or more parameters. [4]