next up previous contents
Next: The Cray Research Inc. T3E. Up: Distributed-memory MIMD systems Previous: The Avalon A12.

The C-DAC PARAM OpenFrame system.

Machine type RISC-based distributed-memory multi-processor
Models PARAM OpenFrame 9000 system
Operating system PARAS micro kernel based Unix (compatible with Sun's Solaris)
Connection structure Multistage variable (see remarks)
Compilers Fortran 77, Fortran 90, HPF, ANSI C, C++
Vendors information Web page www.soft.net/cdac/html/newprod.html#PARAM
Year of introduction 1996.

System parameters:

Model OpenFrame 9000
Clock cycle 6 ns?
Theor. peak performance
Per proc. (64-bit) 300 Mflop/s?
Maximal (64-bit) 300 Gflop/s?
Memory/node ---
Memory (maximal) ---
Communication bandwidth
Point-to-point 3.2 GB/s
Bisectional (maximal) ---
No. of processors 1--1024

Remarks:

The OpenFrame 9000 system is the fourth generation of CDAC machines that is developed by CDAC, the Centre for Development of Advanced Computing, an institute in India that has as its mission to develop an manufacture "state-of-the-art open architecture supercomputers". This system is the second generation that is marketed abroad. In the predecessor, the PARAM 9000/SS SuperSPARC II processors were used. In the present model the UltraSPARC is employed. Also, the maximum possible number of processors has been increased from 200 to 1024. The OpenFrame 9000 uses Sun's UltraSPARC processors but it is not clear from the scarce documentation at what clock cycle(s) they are running. Assuming the same basic processors as Sun employs in most of its Enterprise servers, the estimates as given in the parameter list above seem the most probable. The documentation also does not reveal any details of the type of the interconnection network. Only a point-to-point communication bandwidth is quoted.

The amount of available software shows that the PARAM OpenFrame system is not a first-generation system. Apart from Fortran 77, Fortran 90, HPF, and C++ are available and the CORE, MPI, and PVM message passing interfaces are available. There is a parallel debugger, a proprietary performance evaluation tool called AIDE, while TOTALVIEW can be delivered on request.

In addition, a library of parallel routines, PARUL, is available. This library contains PVM versions of dense linear algebra routines, eigenvalue routines, and FFTs.

Measured Performances: No measured performances of the PARAM OpenFrame 9000 are available.



next up previous contents
Next: The Cray Research Inc. T3E. Up: Distributed-memory MIMD systems Previous: The Avalon A12.



Aad van der Steen
Thu Feb 12 11:58:21 MET 1998