link to homepage

Institute for Advanced Simulation (IAS)

Navigation and service


Coupled SMP-Systems for Scientific High-performance Computing - Status and Needs for Development


Shared memory parallel computers (SMPs, symmetric multiprocessors, shared memory multiprocessors) are offered by many vendors. In the near future, they will represent the standard systems for simulations with medium performance requirements. They are also broadly used in commercial applications. In order to achieve very high performance, SMPs are tightly coupled over a dedicated network to form a parallel computer with SMP nodes or they are temporarily connected (clustered SMPs). Both types of architecture are called coupled SMP systems. The new hierarchical level in the computer architecture means more complexity for application development.

In this analysis, the following work areas are investigated, in which software has to be enhanced or new software has to be developed in order to guarantee the efficient use of coupled SMP systems in science and industry: programming models and languages, tools for program development and optimization, building blocks for applications, porting of applications, resource management and usage models, parallel I/O and data management.

As the activities of the computer vendors in developing the required software functionality and tools will not be sufficient to guarantee the efficient application of such parallel computers, a substantial financial support of projects in this area by the BMBF is considered absolutely necessary.

Partners / Grants

The project was funded by the German Federal Ministry of Education and Research (BMBF).

The grant period was June 1999 until November 1999.


Gekoppelte SMP-Systeme im wissenschaftlich-technischen Hochleistungsrechnen - Status und Entwicklungsbedarf -
Berrendorf, R.; Bischof, C.; Brunst, H.; Bücker, M.; Detert, U.; Esser, R.; Gerndt, M.; Grotendorst, J.; Gutheil, I.; Hoppe, H.-C.; Hoßfeld, F.; Körfgen, B.; Lang, B.; an Mey, D.; Mohr, B.; Nagel, W.E.; Solchenbach, K.; Sutmann, G.; Tikko, V.; Wollschläger, L.
BMBF Report, 01 IR 903, (2000)