The annual National Instruments “geek bash” called NI Week was held August 4 through 6 in Austin, Texas. President and co-founder James Truchard (Dr. T) greeted a record crowd in the opening keynote with a remark about how the “financial engineers” who helped create the mess we’re in should go back and do “real engineering” to help get us out. “Math the way it’s been taught in business schools has serious flaws.” Warren Buffet, he said, called it a “false perception” of reality. In a later meeting with editors, he said that business schools are trying to apply the laws of physics to an area with a lot more variation.
Truchard used this a jumping off place to promote the added “real-time” math capabilities in the latest revision of the company’s flagship programming platform–LabView 2009. The new release also moves LabView more heavily into systems design. Truchard then went on to tackle a Business Week headline about not enough innovation by pointing out the many ways LabView users have been innovating.
NI is a company that prides itself on technological innovation. Some companies may be sales-driven, a few may be marketing-driven, but NI is still a nimble, technology-driven company. And nowhere is this seen like the latest release of LabView and its supporting cast of hardware.
NI is a company that prides itself on technological innovation. Some companies may be sales-driven, a few may be marketing-driven, but NI is still a nimble, technology-driven company. And nowhere is this seen like the latest release of LabView and its supporting cast of hardware.
The latest generations of chips are fertile field for NI engineers. Intel’s multicore microprocessors offer the opporunity to show off LabView’s inherent parallelism. With LabView, programmers can target parts of programs to run on separate cores of the processor and then be reassembled to accomplish its task. For example, data acquisition and analysis could run on one core at the same time another core is driving the display. Meanwhile another core could be controlling some I/O.
My introduction to field-programmable gate arrays (FPGAs) many years ago was their application as a prototyping platform–not as a production microprocessor chip. That has changed. LabView has supported FPGAs for a number of years creating products than can change personality through software. Now, the company is demonstrating the power and speed of the devices allowing it to ramp up data acquisition speeds to a level unthought of a few years ago.
NI Co-founder and “father of LabView” Jeff Kodosky wondered during his keynote why computer science departments have not seized upon LabView for courses in parallel programming, since it’s the only graphical platform for doing it. LabView is not inherently sequential as are text-based languages. It operates with parallel data flow rather than sequential control.
NI has dipped into the wireless sensor network arena–but in its own terms and in its own way. It unveiled WSN nodes build with IEEE 802.15.4 radios that connect in a “mesh” networking topology. The protocol stack is ZigBee with modifications for things NI needed, for example NI believes that computing devices communicate using Internet Protocol (IP), so its engineers wanted to incorporate that into the network. Mentioned almost as an aside is one of the most powerful parts of the entire system–LabView embedded on a node. It won’t be a full-blown version of LabView, but it will be enough that a node can be a controller, or at least an analyzer of raw data, before reporting status and information back to the system.
SolidWorks has become a partner to the extent that “soft motion control” in LabView can team up with the computer aided design (CAD) environment at the design stage so that mechanical and electrical/control engineers can collaborate at an early project stage and see what a machine can do before cutting chips on the factory floor. This partnership has the potential for a powerful leap forward in machine design.