Subject:
Re: Request to review a paper C508
From:
CAPA <ralf.gruber@epfl.ch>
Date:
Fri, 18 May 2001 14:22:35 +0200
To:
Geoffrey Fox <fox@mailer.csit.fsu.edu>
CC:
jacques.menu@epfl.ch, ralf.gruber@epfl.ch

Dear Geoffrey,

here are Jacques Menu's and my remarks on the paper C508 by Kavas and
Feitelson:

.. The paper's main interest is in the QNX real-time microkernel system
.. The NT system is not further supported on the Alphas
.. The decision to choose a system is generally taken in a very pragmatic
way. Existance of efficient mathematical and communication libraries are
most important. What can be said about MPI, ScaLAPACK, BLAS, etc on QNX?
.. Modern clusters for industry must guarantee availability and scalability.
This is the reason why an OS like TruCluster running on a multi-processor
frontend becomes more and more popular. The cluster itself is connected to
the frontend and jobs are submitted via a resource management system such
as LSF or GRD/Codine. Can NT, QNX or Linux guarantee availability?
.. How is job submission handled with the different OS in the paper? 
.. What are the application that have been run besides the Web search and
what were the experiences made?

We can propose to consider this paper for publishing if the authors answer
our questions.

With kind regards,

Ralf Gruber and Jacques Menu, EPFL




At 07:12 24.04.01 -0400, you wrote:

>Sorry. I uploaded again. Now it is OK I think
>  > Geoffrey Fox  gcf@cs.fsu.edu or fox@csit.fsu.edu > Phones Cell 315-254-6387 FSU Office 850-644-4587 FAX 850-644-0098