Posted on 02/16/2004 2:08:15 AM PST by Swordmaker
Conventional wisdom paints the Apple (Nasdaq: AAPL) Macintosh as a "soft" machine. Elementary schools might use it, and those in creative fields might use it, but those who want to do hard-core computation choose workstations running Windows, Unix or Linux, right?
Not necessarily.
"If you pull up a shot of NASA after the [first] Mars landing and look at the desktops, you'll see a couple of PC laptops there, but you'll see more PowerBooks," Jon Rubinstein, senior vice president for hardware engineering at Apple, told the E-Commerce Times.
Although no hard figures are available to chart the Mac's rise in scientific communities, anecdotal evidence suggests various Apple machines, from the Xserve G5 to the PowerBook, have become viable options. For example, Virginia Tech chose last fall to build a supercomputing cluster using Power Mac G5s, then decided to upgrade to Xserve G5s when those machines became available. The university's choice of Apple products stemmed from the computers' attractive price-performance ratio, Virginia Tech spokesperson Lynn Nystrom recently told the E-Commerce Times. As Apple continues to reinvent itself, how are scientists putting Macs to work in research projects and other innovative endeavors?
Macs for Mars
Matt Golombek, a planetary geologist at NASA's Jet Propulsion Laboratory (JPL), has used Apple machines since the Mac SE came out in the late 1980s. Golombek told the E-Commerce Times that 90 percent of his JPL colleagues also use Macs for a host of reasons.
In Golombek's case, he was responsible for choosing the landing sites for both of the Mars rovers, Spirit and Opportunity. To make that determination, he needed to render maps, mosaics and related images into Canvas, a graphics application. This task entailed layering maps together and being able to toggle between different data sets to find parallels between them.
"For our landing site work, we always get the highest-end desktop Mac we can find, so we just got one of the G5s with dual 2-GHz processors and 8 GB of RAM," Golombek told the E-Commerce Times. "We are dealing with Canvas files that are a gigabyte in size. They're huge, huge files, so we need every bit of performance we can get."
According to Golombek, Macs traditionally have been superior in handling images. Although PCs may have caught up in this regard, he added, the Mac's ease of use and computational power ensure he has no plans to switch to another platform. That said, he is eager for Apple to introduce a PowerBook G5 so that he can take advantage of the new chip's ability to handle more RAM.
'A Revelation'
Michael Swenson, life sciences computing analyst at research firm IDC, told the E-Commerce Times that Apple's Unix-based Mac OS X has been the driver of Mac popularity in such areas as bioinformatics and chemistry, mainly because porting open-source applications from Linux and Unix has become a trivial process.
"Having that flexibility to hit a key and go from Microsoft Office to an open-source application and back is appealing" to scientists, Swenson said.
For his part, Theodore Gray, director of user interface technology and co-founder of Mathematica maker Wolfram Research, said that before OS X, some problems were too big for Macs to handle. Now, he noted, Macs can tackle almost any task.
"It's safe to say that, since Mac OS X became widely adopted, there has been an increase of use within [the scientific] community," Gray told the E-Commerce Times. "Scientists tend to prefer Unix (including Linux) over anything else, and Mac OS X is a revelation. You can compile all of your stuff: source code, projects, scientific subroutines. Mac OS X provides a nice environment for all of them."
Indeed, as senior software engineer at the MIT Whitehead Institute for Genome Research, William Van Etten once needed four computers to do his work: a Windows box for productivity , a Unix box for development, a Linux computer for Linux development and a notebook computer. When Mac OS X became viable, Van Etten was able to slim down to a single computer: a Mac.
Hardware BLAST
Gray added that Apple's strategy of designing both the Mac operating system and hardware allows the company to provide a more coherent and polished setup than typical PC configurations.
For example, Stan Gloss, managing director of life sciences IT consultancy The BioTeam, told the E-Commerce Times that certain applications in the bioinformatics industry are optimized for the G4 chip's velocity engine -- a coprocessor that increases application performance from five to 50 times over regular levels.
He added that the new G5 architecture, which supports 64-bit processing and up to 8 GB of RAM per unit (systems that handle 32-bit processing max out at 2 GB of RAM), is an even bigger boon for life sciences.
"In the life sciences, you are dealing with large data sets being input into memory, so you need a fairly large memory space when you are comparing, let's say, the genome of a fruit fly with a human genome," Gloss said.
He provided an example: Basic Local Alignment Search Tool (BLAST), a common bioinformatics application, works as a similarity search engine to match genomes. The program matches sequences of four letters representing a gene that can run almost indefinitely. Gloss said the Mac processes these queries at a speed that significantly improves the overall price-performance ratio of the organization using it.
In one practical application of Mac technology, BioTeam helped Texas A&M researcher David Adelson construct an Xserve cluster so that he and his team could map the cow genome. Adelson is expected to complete this work by the end of 2004.
Personal Preference
Why are these moves toward more Mac use taking place now? Gray contended that Apple occasionally takes a step ahead of other vendors in its price-performance ratio. The company now appears to be in one of those "leapfrog" cycles. Wolfram Research has a G4 cluster installed, and Gray said the company is happy with it because it is easy to maintain and is price competitive.
Indeed, because PCs no longer carry the huge price advantage they once did, choosing a hardware and software configuration now also involves an element of personal preference, Gray said. He noted that with Macs, "you do not have the sort of virus problem as with Windows."
And David J. Stevenson, George Van Osdol Professor of Planetary Science at Caltech, said that he, like Golombek, has used Macs for years because he does not want to be a rocket scientist of system administration. "A lot of scientists are like me -- they may know a lot more about how computers work than the general public, but they don't really care," Stevenson told the E-Commerce Times. "They just want something that works reliably."
As Apple's Mac becomes more and more reliable, its scientific renaissance is likely to gather steam.
You may wish to re-think your remark about these guys using the most expensive computers.
The computers the article talks about are not the common PC used in business. They are maximum performance and maximum memory machines. This type of maximum performance PC has never had much of a cost advantage over a top-of-the-line Mac.
In fact, the machines my company buys to get the maximum advantage out of Mathematica and orbital ephemera software costs $5K - $6K.
I think you are confusing "price" and "cost".
all of those MACs are probably bought with government grant money as opposed to an individuals own
OS X is what made all the difference. Scientists love Unix, and the ability of this OS to run Unix and a high-quality graphic interface at the same time is what pushes them toward Mac.
Now if only hardware price/performance would catch up. A SETI work unit takes 7 hours on my G4 iMac, run from the Unix command ine (no graphics). On my infidel machine, an IBM R50 laptop running Windows XP, it takes 1 hour 55.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.