Newsgroups: comp.sys.transputer,comp.parallel.pvm,comp.parallel.mpi
From: D.J.Beckett@ukc.ac.uk (Dave Beckett)
Subject: Parallel Computing Archive at HENSA Unix: NEW FILES
Summary: New files since 8th November 1994. See ADMIN article for other info.
Keywords: transputer, occam, parallel, archive, anonymous ftp, www, gopher
Organization: University of Kent at Canterbury, UK.
Date: Fri, 09 Dec 94 17:00:47 GMT
Message-ID: <133@nutmeg.ukc.ac.uk>

This is the new files list for the Parallel Computing Archive at
HENSA Unix.  Please consult the accompanying article for
administrative information and the various ways to access the
files.

For experts:
     WWW/Mosaic URL: http://www.hensa.ac.uk/parallel/
	       OR
     anonymous ftp to unix.hensa.ac.uk and look in /parallel
	       OR
     gopher to unix.hensa.ac.uk port 70 and go to "Parallel Archive"

Dave


MIRROR SITES
~~~~~~~~~~~~

There are full mirrors of the archive in Japan and France.  See
http://www.hensa.ac.uk/parallel/www/mirror-sites.html for details.


NEW FILES after 8th November 1994 (newest first)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

8th December 1994

/parallel/documents/mpi/anl/sut-1.0.5.tar.Z
	Scalable Unix Tools V1.0.5 by Gropp and Lusk


5th December 1994

/parallel/documents/hippi/hippi-atm_1.4.ps.gz
	Version 1.4 of the "HIPPI Mapping to ATM" working draft of 11th
	August 1994.

/parallel/documents/hippi/hippi-atm_1.4_changes.ps.gz
	Changes for version 1.4 of the above document.

/parallel/documents/hippi/hippi-ph_8.3.ps.gz
	Contains the electrical, signalling, and connector specifications
	for HIPPI.  HIPPI-PH is an approved American National Standard;
	This is a maintenance copy consistent with the standard, and 
	includes a list of proposed changes for future editions. 

29th November 1994

/parallel/software/simulators/chaos/docs/ebn.ps.Z
	"The Express Broadcast Network: A Network for Low-Latency Broadcast
	of Control Messages" by Kevin Bolding and William Yost, Dept. of
	Computer Science and Engineering, University of Washington,
	Seattle, Washington, USA. November 28, 1994.
	ABSTRACT:
	  We present the Express Broadcast Network (EBN), a network used for
	quick and reliable broadcast of control messages in multicomputer
	networks. The EBN can be implemented with a single extra wire per
	network link and with minimal extra hardware at each routing node.
	However, it provides very fast broadcast mechanisms that take
	advantage of all redundancy in the network to deliver messages
	regardless of faulty network components. We present extensions of
	the basic network to include multiple-wire, multiple-bit, and
	bidirectional wire support, as well as describing basic methods of
	using the EBN for various applications.

/parallel/software/os/minix/aachen/lcc.README
	Transputer compiler for Minix overview and news.

/parallel/software/os/minix/aachen/lcc.tz
	Transputer compiler for Minix package (binaries)


28th November 1994

/parallel/books/ieee/programming-langs-for-parallel-programming.html
	 Programming Languages for Parallel Processing, edited by David B.
	 Skillicorn and Domenico Talia, and published by the IEEE Computer
	 Society Press.

/parallel/books/cup/foundations-of-parallel-programming.html
	Foundations of Parallel Programming by David Skillicorn published
	by Cambridge University Press: This book summarises the application
	of initiality, using the categorical data type approach, to
	general-purpose parallel computation.  There is also an extensive
	survey of models of parallel computation.

/parallel/software/paragon/nxlib/NXLibV1_1_3_Solaris.tar.Z
	Solaris version of NXLib: Paragon Message-Passing Environment for
	Workstations.


25th November 1994

/parallel/events/scalable-computers-languages-compilers-rts
	Call for papers and registration details for the Third Workshop on
	Languages, Compilers, and Run-Time Systems for Scalable Computers
	being held from 22nd-24th May 1995 at Rensselaer Polytechnic
	Institute in Troy, New York, USA.  Topics: machine-independent
	parallel programming; support for parallel computations over
	irregular data as well as for coarse grain parallelism; compiler
	and run-time communication optimizations; compiler optimizations to
	expose parallelism; and experience with applying parallel
	programming languages and software systems to applications.
	Deadlines: Abstracts: 10th February 1995; Notification: 24th March
	1995; Final papers: 22nd May 1995.

/parallel/events/ipps95-iopads95
	Call for papers for the 3rd annual Workshop on I/O in Parallel and
	Distributed Systems (IOPADS) being held at the9th International
	Parallel Processing Symposium (IPPS '95) being held from 25th-28th
	April 1995 at Fess Parker's Red Lion Resort, Santa Barbara,
	California.  Topics: I/O characteristics (demand); subsysems;
	performance modelling; design of I/O intensive applications;
	practice and experience; OS, language and compiler support;
	real-time and multimedia I/o; scheduling and resource allocation;
	concurrent and parallel file system.  Deadlines: Full paper: 21st
	February 1995; Notification: 3rd April 1995; Camera-ready paper:
	14th April 1995.
	Entry-Range:1995-04-25-1995-04-28

/parallel/events/pact95
	Call for papers for International Conference on Parallel
	Architectures and Compilation Techniques (PACT'95) being held from
	27th-29th June 1995 at Limassol, Cyprus.  Sponsored by the IFIP WG
	10.3 (Concurrent Systems), ACM SIGARCH and IEEE TC on Computer
	Architecture.  Topics: Novel computation models, architectures
	(fine/medium grain parallelism), compiler and hardware techniques,
	data-driven and multithreaded machines, exploitation of
	application-specific architectures, etc.  Deadlines: Full papers:
	10th January 1995; Notification: 10th March 1995.
	Entry-Range:1995-06-27-1995-06-29

/parallel/events/ispan94-programme
	International Symposium on Parallel Architectures, Algorithms and
	Networks (ISPAN' 94), Kanazawa, Japan programme and registration
	details.

/parallel/events/hpcs95
	Call for papers for High Performance Computing Symposium'95
	(HPCS'95) being held from 10th-12th July 1995 at Montreal, Canada.
	Topics: HPC systems and tools, clusters, execution models,
	compilers, tools, OSs, comms, ATM, comms software, network
	simulation and optimization, numerical modeling and methods, CFD,
	computational chemistry and physics, weather modeling, industrial
	numerical applications, parallel imaging and visualization, GIS,
	medical imaging, VR, solid modeling, parallel databases, parallel
	AI, ...   Deadlines: Papers: 15th February 1995; Notification: 15th
	April 1995; Camera-ready papers: 15th May 1995.	 See also
	ftp://ftp.crim.ca/apar/public/HPCS95/

/parallel/events/ppecc-dist-par-workshop
	Call for papers for the Parallel Processing in Engineering
	Community Club (PPECCC) Workshop on "Distributed vs Parallel:
	convergence or divergence?" being held from 14th-15th March 1995 at
	The Cosener's House, Abingdon, UK.  The Chairman is Professor Peter
	Dew of University of Leeds.  Deadlines: Position papers: 30th
	January 1995; Acceptance: 10th February 1995.


24th November 1994

/parallel/documents/mpi/anl/mpich-1.0.5.tar.Z
	MPI Chameleon implementation version 1.0.5 (23rd November 1994).

/parallel/environments/lam/distribution/xled10.tar.gz
	XLED 1.0 distribution for LAM.  XLED is an X/Motif based LED server
	that emulates good old hardware LEDs.  It is implemented on top of
	the LAM cluster computing environment.  It provides a low cost
	alternative for sexy blinking LEDs demos (popular with supervisors
	and managers!) and quick-n-dirty debugging.


23rd November 1994

/parallel/documents/mpi/anl/misc/pio.tar.Z
	MPI Parallel I/O sources

/parallel/environments/pvm3/emory-vss/epvmugtalk.ps.Z
	Viewgraphs of a talk on threads and parallel I/O in PVM
	presented at the European PVM Users Group meeting, Rome 10/94.


22nd November 1994

/parallel/documents/mpi/anl/mpi-html.tar.Z
	MPI final report in HTML (May 5th including errata)


21st November 1994

/parallel/documents/pario/papers/Kotz/kotz:diskdir.ps.Z
	"Disk-directed I/O for MIMD Multiprocessors" by David Kotz.
	ABSTRACT:
	  Many scientific applications that run on today's multiprocessors
	are bottlenecked by their file I/O needs. Even if the
	multiprocessor is configured with sufficient I/O hardware, the
	file-system software often fails to provide the available bandwidth
	to the application. Although libraries and improved file-system
	interfaces can make a significant improvement, we believe that
	fundamental changes are needed in the file-server software. We
	propose a new technique, disk-directed I/O, that flips the usual
	relationship between server and client to allow the disks
	(actually, disk servers) to determine the flow of data for maximum
	performance. Our simulations show that tremendous performance gains
	are possible. Indeed, disk-directed I/O provided consistent high
	performance that was largely independent of data distribution, and
	close to the maximum disk bandwidth.

/parallel/documents/pario/papers/Kotz/kotz:diskdir-tr.ps.Z
	"Disk-directed I/O for MIMD Multiprocessors" by David Kotz.
	Technical Report.
	ABSTRACT:
	  Many scientific applications that run on today's multiprocessors
	are bottlenecked by their file I/O needs. Even if the
	multiprocessor is configured with sufficient I/O hardware, the
	file-system software often fails to provide the available bandwidth
	to the application. Although libraries and improved file-system
	interfaces can make a significant improvement, we believe that
	fundamental changes are needed in the file-server software. We
	propose a new technique, disk-directed I/O, that flips the
	usual relationship between server and client to allow the disks
	(actually, disk servers) to determine the flow of data for maximum
	performance. Our simulations show that tremendous performance gains
	are possible. Indeed, disk-directed I/O provided consistent high
	performance that was largely independent of data distribution, and
	close to the maximum disk bandwidth.


17th November 1994

/parallel/documents/hpf/hpf2-requirements.ps.Z
/parallel/documents/hpf/hpf2-requirements.tar.Z
	"HPF-2 Scope of Activities and Motivating Applications" by High
	Performance Fortran Forum.  13th November 1994. 88 pages.
	ABSTRACT:
	  This document presents issues recommended by the HPF Forum for
	consideration under the HPF-2 effort. It comprises three parts: a
	set of proposed new capabilities, a proposal for a Kernel HPF, an
	official subset designed for particularly high performance and a
	set of applications motivating the new capabilities.


16th November 1994

/parallel/user-groups/ppc/PPC-1994-November
	 Announcement of the PPC Meeting to be held on 14th November 1994.

/parallel/faqs/simulated-annealing
	Simulated and Adaptive Simulated Annealing (ASA) information from
	several news articles.

/parallel/events/wads95
	Call for papers for 4th Workshop on Algorithms and Data Structures
	being held from 16th-18th August 1995 at Kingston, Ontario, Canada.
	Deadlines: Papers: 17th February 1995; Nofication: 14th April
	1995; Final papers: 12th May 1995.  See also
	http://www.scs.carleton.ca/scs/wads/wads95.html

/parallel/software/os/minix/aachen/asld.ps.Z
	"A new assembler for transputers" by Michale Haardt
	<michael@cantor.informatik.rwth-aachen.de>.
	ABSTRACT:
	  The special architecture of transputers and the requirements of
	fast compilers require to leave traditional implementation
	techniques.  This document describes the design and implementation
	of an assembler and linker for transputers in the UNIX programming
	environment.

/parallel/software/os/minix/aachen/lcc-tx.ps.Z
	"On writing a lcc back end for transputers" by Michale Haardt
	<michael@cantor.informatik.rwth-aachen.de>.
	ABSTRACT:
	  This document explains how special parts of the back end for INMOS
	transputers work and gives reasons for certain design decisions.
	It further explains some differences between transputers and more
	common processors.  It is assumed that the reader is familar with
	the documentation included in the lcc distribution.


14th November 1994

/parallel/environments/pvm3/distribution/
	PVM V3.3.5 distribution

/parallel/environments/pvm3/distribution/pvm3.3.5.tar.Z.uu.Z
	PVM V3.3.5 (compressed uuencoded compressed tar file format)

/parallel/environments/pvm3/distribution/pvm3.3.5.shar.Z
	PVM V3.3.5 (compressed shar format)

/parallel/environments/pvm3/distribution/pvm334to5.Z
	Patch from PVM V3.3.4 to V3.3.5

/parallel/environments/pvm3/distribution/epvmug94
	European PVM Users Group Meeting 1995 details


10th November 1994

/parallel/documents/hpf/hpf2-applications.ps.Z
/parallel/documents/hpf/hpf2-applications.tar.Z
	"HPF-2 Motivating Applications DRAFT" editted by Paul Havlak. 40
	pages.
	ABSTRACT:
	  The efficient solution of many scientific problems on parallel
	processors requires language and compiler support beyond the
	capabilities of High Performance Fortran. Several members of the
	HPF Forum have contributed applications to illustrate the
	limitations of HPF-1. For ease of study, the selected codes that
	are small but include parallel idioms important to full-scale
	applications. This report is available via the World-Wide Web at
	ftp://hpsl.cs.umd.edu/pub/hpf_bench/index.html

/parallel/documents/hpf/hpf-v11.ps.Z
/parallel/documents/hpf/hpf-v11.tar.Z
	High Performance Fortran Language Specification
	Version 1.1 draft. 216 pages.

/parallel/documents/pario/papers/Kotz/kotz:throughput.ps.Z
	"Throughput of Existing Multiprocessor File Systems (An Informal
	Study)" by David Kotz.
	  ABSTRACT:
	Fast file systems are critical for high-performance scientific
	computing, since many scientific applications have tremendous I/O
	requirements [MK91]. Many parallel supercomputers have only
	recently obtained fully parallel I/O architectures and file
	systems, which are necessary for scalable I/O
	performance. Scalability aside, I show here that many systems lack
	sufficient absolute performance.  I examined several the papers in
	the literature that report actual performance measurements on
	existing parallel file systems. For each paper, I chose the highest
	reported throughput. A list of these throughputs, normalized by the
	number and speed of disks, is given in the table on page 2.  First,
	note that the numbers come from different experimental arrangements
	and thus cannot be directly compared with each other. Some timings
	include computation, and some do not; some are for reading, some
	for writing; some are sequential, while some are not. Some of the
	parameters were estimated. (A commonly-used benchmark would help,
	and at least one has been proposed [CCFN92, Fin93]).  Second, my
	normalization is extremely crude: dividing by the number of disks
	and then by the raw disk's peak bandwidth is an
	over-simplification. Few systems can approach 100% of the available
	bandwidth, and then only in special cases such as a long sequential
	read. Still, the results are instructive in pointing out the
	difficulty of obtaining good I/O performance in a parallel system.
	Third, note the wide range of throughputs. In only a few papers,
	and only a few cases in those papers, were the authors able to
	extract a significant percentage of the disk bandwidth. The
	performance for the 64-disk Touchstone Delta is particularly
	disappointing (although a new Intel file system is
	forthcoming). The best cases occurred when the I/O was coordinated
	(as in the CM-2, CM-5, or Nitzberg's limited concurrency [KN93]),
	or sequential reading with prefetching and a buffer size to match
	the block size [FPD93].  These results show us that more work is
	needed on improving the raw performance of multiprocessor file
	systems. I suspect that much of the problem is excess software
	overhead. Throughput is lost to extraneous copying and
	message-passing overhead, inadequate I/O node power, cache
	thrashing and prefetch mistakes [KN93, Nit92], architectural
	bottlenecks [KN93, Kry92], and lack of coordination [dBC93].

/parallel/documents/pario/papers/Kotz/kotay:agents.ps.Z
	"Transportable Agents" by Keith D. Kotay and David Kotz.
	ABSTRACT:
	  As network information resources grow in size, it is often most
	efficient to process queries and updates at the site where the data
	is located.  This processing can be accomplished by using a
	traditional client-server network interface, which constrains the
	client to the set of queries supported by the server, or requires
	the server to send all data to the client for processing.  The
	former is inflexible; the latter is inefficient.  Transportable
	agents, which support the movement of the client computation to the
	location of the remote resource, have the potential to be more
	flexible and more efficient.  Transportable agents are capable of
	suspending their execution, transporting themselves to another host
	on a network, and resuming execution from the point at which they
	were suspended.  Transportable agents consume fewer network
	resources and can support systems that do not have permanent
	network connections, such as mobile computers and personal digital
	assistants.  We describe a prototype transportable-agent
	implementation that facilitates research in this area.  Agents are
	written in a script language that supports agent relocation, and
	the language is processed at each host by an agent interpreter.
	Electronic mail and UNIX remote shell (rsh) are the two current
	transport mechanisms and we plan to explore others.  We present a
	technical-report searching agent as a demonstration of the
	capabilities of our prototype implementation.


9th November 1994


/parallel/vendors/
	New hand-crafted HTML index, with icons for some vendors.

