Data-Parallel BLAS as a Basis for LAPACK on Massively Parallel Computers

1993 ◽  
pp. 13-20 ◽  
Author(s):  
P. E. Bjørstad ◽  
T. Sørevik
Keyword(s):  
Parallel Computers ◽  
Massively Parallel ◽  
Data Parallel ◽  
Massively Parallel Computers

POMPC: A C LANGUAGE FOR DATA PARALLELISM

1993 ◽  
Vol 04 (01) ◽  
pp. 85-96
Author(s):  
NICOLAS PARIS
Keyword(s):  
Parallel Computers ◽  
Massively Parallel ◽  
C Language ◽  
Parallel Language ◽  
Data Parallel ◽  
Massively Parallel Computers ◽  
Connection Machine ◽  
Compilation Process ◽  
Points Of View ◽  
Definition Of

POMPC is a parallel language dedicated to the programming of Massively Parallel Computers according to a synchronous Data Parallel model. It is an extension of the ANSI C language and follows its philosophy. Parallelism is explicitly handled by the definition of collections of parallel variables and the definition of communication primitives. A methodology is presented in order to easily port the language on different target architectures. Virtualization is introduced to handle simultaneously several collections of different sizes and shapes. Virtualization management is a key point of the compilation process. Programmer, architecture, compilation and system points of view lead to a special implementation of the virtualization mixing physical and virtual parallel objects. The implementation of the virtualization is adapted for the development of communication libraries and also adapted to enlarge the asynchronous sections of code for SPMD architecture. The portability of the POMPC language is validated by several implementations for mono/multi-process simulation on UNIX machines, for the Connection Machine CM-2, for the MasPar MP-1 and a compiler is in preparation for the iPSC-860.

scholarly journals C++ and Massively Parallel Computers

1993 ◽  
Vol 2 (4) ◽  
pp. 193-202 ◽  
Author(s):  
Daniel J. Lickly ◽  
Philip J. Hatcher
Keyword(s):  
Object Oriented ◽  
Feasibility Studies ◽  
Parallel Computers ◽  
Parallel Architectures ◽  
Parallel Execution ◽  
Object Oriented Programming ◽  
Massively Parallel ◽  
Data Parallel ◽  
Massively Parallel Computers ◽  
Massively Parallel Architectures

Our goal is to apply the software engineering advantages of object-oriented programming to the raw power of massively parallel architectures. To do this we have constructed a hierarchy of C++ classes to support the data-parallel paradigm. Feasibility studies and initial coding can be supported by any serial machine that has a C++ compiler. Parallel execution requires an extended Cfront, which understands the data-parallel classes and generates C*code. (C*is a data-parallel superset of ANSI C developed by Thinking Machines Corporation). This approach provides potential portability across parallel architectures and leverages the existing compiler technology for translating data-parallel programs onto both SIMD and MIMD hardware.

The Fast Multipole Method in Canonical Ensemble Dynamics on Massively Parallel Computers

1992 ◽  
Vol 278 ◽  
Author(s):  
Steven R. Lustig ◽  
J.J. Cristy ◽  
D.A. Pensak
Keyword(s):  
Canonical Ensemble ◽  
Yukawa Potential ◽  
Direct Method ◽  
Fast Multipole Method ◽  
Parallel Computers ◽  
Massively Parallel ◽  
Fast Multipole ◽  
Massively Parallel Computers ◽  
Multipole Method ◽  
Execution Times

AbstractThe fast multipole method (FMM) is implemented in canonical ensemble particle simulations to compute non-bonded interactions efficiently with explicit error control. Multipole and local expansions have been derived to implement the FMM efficiently in Cartesian coordinates for soft-sphere (inverse power law), Lennard- Jones, Morse and Yukawa potential functions. Significant reductions in execution times have been achieved with respect to the direct method. For a given number, N, of particles the execution times of the direct method scale asO(N2). The FMM execution times scale asO(N) on sequential workstations and vector processors and asymptotically0(logN) on massively parallel computers. Connection Machine CM-2 and WAVETRACER-DTC parallel FMM implementations execute faster than the Cray-YMP vectorized FMM for ensemble sizes larger than 28k and 35k, respectively. For 256k particle ensembles the CM-2 parallel FMM is 12 times faster than the Cray-YMP vectorized direct method and 2.2 times faster than the vectorized FMM. For 256k particle ensembles the WAVETRACER-DTC parallel FMM is 33 times faster than the Cray-YMP vectorized direct method.

scholarly journals Car-Parrinello Molecular Dynamics on Massively Parallel Computers

ChemPhysChem ◽  
2005 ◽  
Vol 6 (9) ◽  
pp. 1788-1793 ◽  
Author(s):  
Jürg Hutter ◽  
Alessandro Curioni
Keyword(s):  
Molecular Dynamics ◽  
Parallel Computers ◽  
Massively Parallel ◽  
Massively Parallel Computers

scholarly journals TTN: A High Performance Hierarchical Interconnection Network for Massively Parallel Computers

2009 ◽  
Vol E92-D (5) ◽  
pp. 1062-1078 ◽  
Author(s):  
M.M. Hafizur RAHMAN ◽  
Yasushi INOGUCHI ◽  
Yukinori SATO ◽  
Susumu HORIGUCHI
Keyword(s):  
High Performance ◽  
Interconnection Network ◽  
Parallel Computers ◽  
Massively Parallel ◽  
Massively Parallel Computers
Sign in / Sign up

Export Citation Format

Share Document