Changing the interface of scientific computing

SANTA CLARA, Calif. — Dr. Guy “Bud” Tribble has one of those great names you don’t easily forget. Back in the early 1980s, he and his partner Andy Hertzfeld led the team at Apple Computer Co. that came up with the Macintosh user interface, one that literally changed the face of computing forever.

After a couple of stops at other places — CTO at Sun-Netscape Alliance and at now-defunct Eazel, Inc. (the latter with Hertzfeld) — Tribble is back home in Cupertino and is in charge of the long-term technical direction of Apple’s software engineering strategy. And, being an actual scientist, he is also charge with seeing that Apple continues to ascend in the scientific computing market.

At the recent O’Reilly Mac OS X Conference, Tribble offered an overview of how Apple has been working with public and private institutions to do high-end computational work — such as computing how fast an object will burn up upon re-etry into Earth’s atmosphere, for example. Or how determining how much of the sun’s energy hits the Earth, and what happens to it when it arrives.

Open source: Originator of many powerful scientific apps

Bud Tribble

“Most of the key scientific applications being used today originated in the open source world,” Tribble said. “Because they are mostly so project- and data-specific in nature, these applications obviously are not for everyone. But there are a number of highly successful ongoing projects that are serving the scientific community very well right now.”

Tribble said one of the trends he sees is that the scientific and academic communities are relying more on the Mac to do their daily work, for a number of reasons. “Apple is dedicated to designing products that best serve the professional image processing, audio, and video markets has also given the scientific community a product that is well suited for its needs,” he said. “The high-end processing needs of these users closely matches the processing needs of scientists.”

Because scientists often work with huge data stores and humongous individual files, the fact that the Apple G5 can address more memory and work on larger data structures is a key factor, Tribble said. For example, Tribble said, the human genome, when held in memory, needs anywhere from 10GB to 12GB of physical memory; the G5 can handle that with ease. In addition, the G5 sports two floating-point and two arithmetic units, giving users a substantial amount of mathematical processing power. That’s not all; built into the G4 and G5 is a velocity engine called Altivec, which scientists use to speed up calculations. So there’s now quite a lot of horsepower to go with the Mac OS’s universally respected graphics capabilities.

“Bioinformatics is one of the burgeoning areas we’re well invested in,” Tribble said. “We as a culture have just begun to see what we can do in that science.”

Scientific apps and some of their uses

Here is a listing of some of the companies and their scientific applications — most of which started as open source projects — that have been optimized for use on the Mac OS X system:

• MathWorks’ MATLAB: MATLAB uses an intuitively designed interface to accelerate research, shorten the time invested in analysis and development, reduce project costs, and produce effective solutions. The MATLAB environment enables users to quickly test and compare multiple alternatives.
• Wolfram Research’s Mathematica: Mathematica integrates a numeric and symbolic computational engine, graphics system, programming language, documentation system, and advanced connectivity to other applications.
• PyMOL: PyMOL is a Python-based, user-sponsored molecular graphics modeling system hosted at SourceForge.com. It is used most often in drug discovery and testing projects. The latest release, PyMOL 0.93 for Linux, Mac G4 and G5, Windows, SGI, Solaris, is now available for download. The source code is also available. If you’re working on discovering a new drug, this is the app for you.
• Pixlock’s Exbem: Exbem is ultra-high end image-processing and analysis software, designed for scientific, technical, and medical applications. It allows for close-up examination of single images, as well as QuickTime movies, or live video sequences. When we say “close-up,” we mean it; Exbem is often used in gene-splicing experiments.
• TetrUSS: TetrUSS is a suite of applications used to make transonic calculations in fluid dynamics and aerodynamics analysis and design. The system was developed at NASA Langley Research Center in the 1990s to bring state-of-the-art in computational fluid dynamics (CFD) to a higher level of utility. The software originally was developed for NASA internal applications but has evolved into an efficient CFD tool used by engineers and scientists on many other projects.
• Genzyme: This is a drug research and discovery company which uses a number of open source components in its toolset in efforts to identify new medications for cancer, immune mediated diseases, and severe burns.

Tribble also described smaller projects using custom applications designed for the Mac OS that do things like simulate human brain electrical activity (“to watch in real-time which cells are firing up and which ones aren’t, Tribble said). “When the patient gets a sedative, you suddenly see much less brain activity,” Tribble said.

Makes perfect sense.