Featured User Site:

Simpson Strong-Tie

Simpson Strong-Tie (http://www.strongtie.com/), a leading designer and manufacturer of construction products for fifty years, is a relatively new user of High Productivity Computing (HPC) products. Though new, Simpson has managed to leverage its small HPC cluster in ways that have allowed the company to reduce its products' time to market and increase its product diversity.

Overview

Simpson Strong-Tie was started in the 1950s by members of the Simpson family, who were long-time participants in the Bay Area construction community. The company began as a small family owned and operated window screen business and over the past fifty years evolved into an $800 million public corporation focused on producing state of the art structural connectors and other construction components.

Simpson Strong-Tie's goal is to be a leading producer of construction products that can withstand the earthquakes, hurricanes and other forces generated by Mother Nature. To accomplish its goal Simpson rigorously tests its products before sending them off to market. The testing process currently consists of real physical laboratory tests and HPC aided computer simulations.

Simpson Strong-Tie's facility in Pleasanton, CA is currently using a five-node (one head node, four compute nodes) Linux Networx cluster. Frank Ding, Research and Development Engineer and Finite Element Analysis (FEA) Specialist is in charge of the company's HPC operations and is the sole user at the site and the person who helped bring HPC to Simpson.

Simpson's primary focus is on structural connectors that tie various types of construction members together, for example: wood connected to wood via a steel connector and wood to concrete using a steel connector. Prior to the introduction of their Linux Networx cluster, Simpson's research process relied more heavily on labor intensive testing processes. The processes involving wood required specially trained workers to not only assemble the wood to connectors, but often cut wood in complicated ways that very often took up a great deal of time.

Additionally, test processes involving concrete required even more time and labor. For Simpson, testing concrete blocks involved calling in a cement mixer, pouring the mix into an appropriate form and then waiting approximately a month for the concrete to cure. Once the curing process was over the connectors could then begin testing in the lab. The addition of the Linux Networx cluster, while not eliminating the need for real-world laboratory testing, allows engineers to eliminate certain models that would have most certainly failed in a real-world test and suggest which models have the highest probability of success.

Source: Simpson Strong-Tie

Tabor Research Analysis

Simpson Strong-Tie is pioneering the use of HPC in testing building construction materials. Although the applications (such as ABAQUS) are familiar within HPC, the materials (wood, concrete) are less so, and Simpson is forging new ground with its modest installation.

Though the five-node Linux Networx cluster is small by traditional HPC standards, the impact the system has had on Simpson's testing and design processes has been significant. The cluster allows the company to reduce its research and design hase by simulating many processes that had only been done before in the laboratory shop. By reducing the overall duration of the R&D phase, many of Simpson's products now make it to market much sooner than they had prior to the application of HPC.

Simpson Strong-Tie's acquisition of a cluster could be indicative of a trend of increased HPC use by market actors who had never considered using HPC before. In Simpson's case HPC has taken the place of more costly, time consuming, labor intensive activities and substituted them with faster, lower marginal cost HPC procedures. Tabor Research is encouraged by the expansion of HPC into new markets and believes that as HPC products become cheaper there will be more instances of companies, such as Simpson Strong-Tie, who innovate their research methods with HPC products.

Table: HPC System at Simpson Strong-Tie

Would you like to nominate your user organization to be featured on the Tabor Research web site? Please email your suggestion, including a brief description of the technology in use and why it is noteworthy, to Addison Snell, Addison@TaborResearch.com.