NEW O'REILLY RELEASE MAXIMIZES RISC PERFORMANCE

If you work with computers, you own it to yourself to understand the new
directions that workstations architecture has taken in the last half
decade.

SEBASTOPOL, CA -- O'Reilly & Associates announces the release of the first
book for application programmers and purchasing managers facing the
confusing array of RISC-based workstation architecture, such as the DEC
Alpha/AXP, the IBM RS/6000, Sun's SuperSPARC, and the HP 9000/700 series.
HIGH PERFORMANCE COMPUTING, by Kevin Dowd, a principal with the Atlantic
Computing Technology Corporation of Wethersfield, Conn., not only helps
make sense of the newest generation of workstations, it tells how to
maximize their performance.

HIGH PERFORMANCE COMPUTING covers everything from the basics of modern
workstation architecture, to structuring benchmarks, to squeezing more
performance out of critical applications. It also explains how optimizing
compilers work: it discusses what a good compiler can do and, more
important, what a good compiler can't do -- and what programmers have to
do themselves. The book closes with a look at the high-performance future:
parallel computers, including exotic distributed memory multiprocessors,
and the more "garden-variety" shared memory processors that are already
appearing on desktops.

Tricks are revealed that have been used in supercomputer labs for years,
such as loop unrolling, loop interchange, and other methods that improve
memory access patterns and take advantage of parallelism latent in code.
HIGH PERFORMANCE COMPUTING pays special attention to memory issues,
including perhaps the most important story in high performance computing
(and not likely to be told by vendors), the increasing disparity between
CPU speeds and memory speeds.

Another valuable section of the book discusses the benchmarking process:
how to evaluate a computer's performance. Kevin Dowd discusses several of
the "standard" industry benchmarks, explaining what they measure and what
they don't, so that readers won't be overcome by a fog of statistics when
the next salesperson walks in the door. He also shows how to set up a
benchmark: how to strucutre the code, how to measure the results, and how
to interpret them.

HIGH PERFORMANCE COMPUTING is must reading for anyone who needs to worry
about computer performance, either as a software developer or as a buyer.
But it also provides valuable insights for those who do relatively little
programming and run mostly third-party application software. Even if they
never touch a line of code, the book gives them a feel for what "makes
things tick": it provides valuable insights into how the most recent
generation of computer hardware works.

O'Reilly & Associates is recognized worldwide for its definitive books on
UNIX, The X Window System and the Internet. Working closely with
developers of new technologies, O'Reilly's editors are "computer people"
who use the software and systems they write about. The company's planning
and review cycles link together authors, computer vendors, and technical
experts throughout the industry in a creative collaboration that mirrors
the strengths of the "open systems" philosophy itself.

HIGH PERFORMANCE COMPUTING
by Kevin Dowd
350 pages (estimate)
ISBN 1-56592-032-5
Price: $25.95
Publication Date: June 14, 1993

O'Reilly & Associates Inc
103 Morris St, Suite A, Sebastopol, CA 95472
707-829-0515;   800-998-9938;  fax: 707-829-0104

==========================================================
From the America Online -- New Product Information Service
==========================================================
This information was processed from data provided by the
above mentioned company. For additional details, contact 
the company at the address or telephone number indicated.
==========================================================
All submissions for this service should be addressed to:
BAKER ENTERPRISES, 20 Ferro Dr, Sewell, NJ 08080 U.S.A.
Email:  RBakerPC  (AOL),   rbakerpc@aol.com  (Internet)
==========================================================
