Apple Targets High-Performance MarketApple Targets High-Performance Market
New servers and clustering software could aid scientists, engineers
Researchers at NASA and biotechnology company Genentech Inc. are among the first users of new Apple Computer software for clustering Macintosh computers to run scientific applications. Apple's release of the test software, and the arrival next month of a new server designed for clustering, show how Apple is building products to capture a greater share of the market for high-performance computing.
CEO Jobs says 9.3 million Apple users have upgraded to Mac OS X.Photo by Noah Berger/Bloomberg News/Landov |
At the MacWorld Expo conference last week, CEO Steve Jobs said 9.3 million Apple users--about 40%--have upgraded to Mac OS X, an operating system based on Unix, a mainstay of the scientific and engineering computer markets. Unix on the Mac and faster G5 processors are attracting more buyers in these markets, the company says. "We're getting a lot of interest in using our systems in computationally intense environments," says Tom Goguen, a marketing director for server software at Apple. "We wanted to make it really easy for scientists to use a rack of Macs" as a cluster.
Apple last week released a test version of new Xgrid software for distributing batch jobs such as the gene-sequencing program Blast across Mac servers, desktops, and laptops. Platform Computing and Sun Microsystems also make distributed computing software for the Mac, but Goguen says Apple's release is designed for use in small workgroups such as university computer labs.
Apple has about 5% of the technical desktop market because its machines perform well on benchmarks there, says Earl Joseph, an analyst at IDC. But for server-side apps, "Apple needs to do a lot of catching up." Hewlett-Packard, IBM, and SGI supply more tools and middleware, and little third-party software is optimized for Mac clusters.
Next month, Apple plans to ship versions of its Xserve server computer with its latest 2-GHz G5 processor. A two-CPU, cluster-optimized version will sell for $3,000.
About the Author
You May Also Like