How a Supercomputing Innovation Accelerated a Business
- Friday, 01 February 2008
Founder and CTO
Proto Labs, Inc.
Maple Plain, MN
Proto Labs is a quick-response custom manufacturer that provides design engineers with prototypes of plastic parts faster than anyone in the world. Our Protomold division produces prototypes (and short run production parts) by injection molding, and our First Cut Prototype division produces prototypes by CNC machining plastic stock.
Both divisions are based on core Proto Labs proprietary software that analyzes customers’ 3D CAD models, prices the work they are requesting, and, where appropriate, suggests design changes to improve part performance and manufacturability. The delivered quotation provides both real-time interactive pricing and sophisticated design analysis of customers’ 3D CAD models, including rotatable 3D models of parts with graphically represented change recommendations, an animated fill analysis showing how liquid resin will flow into the mold, and highlighting of areas where problems could arise. The software then generates toolpaths from the 3D CAD model for milling of the mold or the final part. As you can probably guess, this is all extremely computationally intensive, and reducing execution times to acceptable levels requires highly optimized code.
A couple of years ago, we recognized an exponential increase in the demands on our computing resources. This was attributable to growth in demand for our services due to our geographic expansion, as well as the enhancements we made to our processes to support bigger and more complicated part geometries. It became apparent to us that before long we would hit the proverbial wall and we needed to take action.
We quickly realized that even the “latest and greatest” individual processors were not going to allow us to run the software we were developing and keep up with growing demand. The only real solution was a “supercomputer.” There was a time when that would have meant a mainframe, but with advancements in hardware, software, and system architecture, we realized that the system we were looking for was going to be a large compute cluster of loosely coupled microprocessors.
These are typically reserved for research applications, oil exploration, meteorological and astronomical computation, and the like, and are relatively rare in business or industrial applications. But, due to the size, complexity, and speed requirements of our application, ours was an almost ideal application for the cluster architecture. The system divides a large task among multiple nodes, allowing massive amounts of computation to be completed far faster than would be possible on smaller or non-clustered systems. The Proto Labs system is currently comprised of 64 Intel Core-2 Quad processors and can operate at 1.9 teraflops (trillion floating point operations per second). Processors are linked with Gigabit Ethernet, and the system has a total of 256 gigabytes of main memory and five terabytes of disk storage. The cluster operating system is Windows-64.
Less than three years ago, this cluster would have been one of the world’s 500 fastest computer systems. Less than 10 years ago, it would have been the fastest in the world. Today, it is merely a very powerful supercomputer managing internal and external operations at levels that would be virtually impossible without the high-performance cluster.
The end result is that we smashed through the wall, and despite the continued ramp-up in quoting and manufacturing demands, we were able to make our quotations and deliveries even faster, with many of our quotes entirely automated and parts delivered as quickly as the next day. It’s actually a pretty cool instance of necessity being the mother of invention, as well as a unique marriage of high technology with mainstream manufacturing.
For more information on Proto Labs’ Quick-turn Plastic Parts Business, click here.