Parallel Implementation of a Symmetric Eigensolver Based on the Yau and Lu Method
نویسندگان
چکیده
In this paper, we present preliminary results on a complete eigensolver based on the Yau and Lu method. We rst give an overview of this invariant subspace decomposition method for dense symmetric matrices followed by numerical results and work in progress of a distributed-memory implementation. We expect that the algorithm's heavy reliance on matrix-matrix multiplication, coupled with FFT should yield a highly parallelizable algorithm. We present performance results for the dominant computation kernel on the Intel Paragon.
منابع مشابه
Parallel Implementation of the Yau and Lu Method for Eigenvalue Computation
In this paper, parallel extensions of a complete symmetric eigensolver, proposed by Yau and Lu in ' l 993' are pre' sented. First, an overview of this invariant subspace decomposition method for dense symmetric matrices is g iven, fo l lowed by numer ical resul ts . Then' works are exposed in progress on distributed-memory implementation. The algorithm's heavy reliance on matrix-matrix multipli...
متن کاملAn efficient implementation of parallel eigenvalue computation for massively parallel processing
This article describes an e cient implementation and evaluation of a parallel eigensolver for computing all eigenvalues of dense symmetric matrices. Our eigensolver uses a Householder tridiagonalization method, which has higher parallelism and performance than conventional methods when problem size is relatively small, e.g. the order of 10,000. This is very important for relevant practical appl...
متن کاملTowards a Parallel Multilevel Preconditioned Maxwell Eigensolver
We report on a parallel implementation of the Jacobi–Davidson (JD) to compute a few eigenpairs of a large real symmetric generalized matrix eigenvalue problem
متن کاملA new CPA resistant software implementation for symmetric ciphers with smoothed power consumption: SIMON case study
In this paper we propose a new method for applying hiding countermeasure against CPA attacks. This method is for software implementation, based on smoothing power consumption of the device. This method is evaluated on the SIMON scheme as a case study; however, it is not relying on any specific SIMON features. Our new method includes only AND equivalent and XOR equivalent operations since every ...
متن کاملA Parallel Implementation of the Trace Minimization Eigensolver
In this paper we describe a parallel implementation of the trace minimization method for symmetric generalized eigenvalue problems proposed by Sameh and Wisniewski. The implementation includes several techniques proposed in a later article of Sameh, such as multishifting, preconditioning and adaptive inner solver termination, which accelerate the method and make it much more robust. A Davidson-...
متن کامل