Arrival of the Desktop Supercomputer17 Mar, 2010 By: Randall S. Newton
Nehalem workstations offer a strategic advantage we haven't seen in years, and now — yes, now — is the time to ramp up.
Editor's note:This article was originally published in the Winter 2010 edition of Cadalyst magazine.
Here we are in the middle of a big fat recession, and the most empowering advance in CPU technology in 20 years has arrived. I'm not talking about a new model that's a little bit better, faster, or cheaper; I'm talking about the advent of the desktop supercomputer. The boss says "do more with less," but your favorite vendor has just introduced new workstations that are radically redefining the value proposition. What do you do?
Be sneaky. Be bold. Be whatever it takes. Start equipping your key technical professionals immediately. The Nehalem generation of CPUs from Intel, combined with new graphics processing unit (GPU) technology from NVIDIA or AMD and the most stable version of Windows ever, will revolutionize the workflow of the people who run the most compute-intensive applications, including:
- the gal who creates those stunning renderings the executives keep requesting,
- the graybeard who runs those complicated multiphysics simulations,
- the surveyors and civil engineers who shoot gigabytes of 3D laser data, and
- the young architects who do their initial design work in 3D.
These people — and others like them — offer their firms experience, skill, and creativity based on years of training and practice. Often they are more than the final arbiter; they also are important sources of innovation in a firm.
New Value Proposition
In years past, these technical and creative professionals — in fields as diverse as oil and gas exploration, architecture, product development, and animation — were valued for their expertise, but often placed in advisory or support roles. They were consulted to analyze, prove, or finish existing projects or designs but not considered part of the mainstream production or design process. In effect, these valuable team members were treated as consultants in their own companies. Why? Because their work was so compute-intensive it would interrupt the project workflow if mainstreamed. Their computer requirements were expensive and thus used sparingly. It requires much more computer-processing time to analyze a part than to model it. Rendering a scene on a computer is much more demanding than drawing it. The increased need for processing power is one reason many of these specialists are outside the production workflow. Their demands on existing IT are so great that their work would slow down other work processes.
In these rough economic times, keeping these experts separate is no longer viable. The challenge today is how to equip these valuable professionals appropriately and make each a full contributor. The new-generation workstations are key to this proposition. Renderings or analysis runs that would take hours with last year's technology now require only minutes on a Nehalem system.
It gets better. The provable return on investment for a Nehalem-class workstation is without recent precedent. Not only are there directly measurable increases in productivity, but these increases open the door to reorganizing entire workflows to take better advantage of specialized knowledge. You can empower your innovators in a way never possible before.
Consider an example from the oil and gas exploration industry. Schlumberger offers a variety of tools for complex geophysical visualization and analysis. Some run on Windows, others on Linux. Until the Nehalem workstations arrived, it wasn't feasible to run dual operating systems on one workstation using virtualization technology; the result was just too slow. But on a Nehalem workstation, a geophysicist can run a Linux session and a Windows session side by side. In one test I observed, a last-generation workstation could process one frame from one of these applications every 50 seconds. By comparison, a single Nehalem-generation workstation running virtualization software can process 30 frames per second in two applications, one in Windows and the other in Linux. That's roughly a 3,600% productivity gain.
This example is just one case of how it is possible to put near-supercomputer capabilities inside a single technical workstation at an affordable price. The resulting price-to-performance ratio improvement in most cases will eliminate the cost of computer processing as an impediment to use. Return on investment in a new computer workstation will be measurable in days instead of months or years.
One way to view the new generation of technology is to avoid putting software, hardware, and peripherals into categories and instead to think of them as an integrated whole. A term starting to come into use for this is digital workbench. The phrase originated in science laboratories, but it is beginning to spread throughout the technical and creative professions.
In the digital workbench concept, hardware and software vendors become partners in making sure that a workstation is a unified system and a proper fit. A digital workbench in product development would do more than computer-aided design, for example; it also would be a workstation for digital prototyping, analysis-driven design, and design-based simulation.
Now the value decision point is about maximizing the value of the creative or technical professional who sits at the digital workbench all day. A multicore, multi-GPU workstation has the power to run a simulation in the background while an engineer does design work in another window, instead of tying up two (or more) computers or waiting for the simulation to finish.
Much innovation is the result of seemingly random explorations, the what-if sense of inquiry. When a computer system is fast enough, it becomes possible to take the time to explore new ideas and design iterations, to allow creativity to benefit from serendipity.
For years now, most technical professionals haven't thought of their computers as offering a strategic advantage. Perhaps — at least to gain first-mover advantage in a slow economy — it's time to rethink that notion.