Sponsored By

NI Meets Concurrent Processing Challenges with LabView Software

August 8, 2007

1 Min Read
NI Meets Concurrent Processing Challenges with LabView Software

Concurrency is prompting dramatic changes in computing and instrumentation systems, bringing programming challenges for those who need more computing horsepower. Single-processor chips are topping out at around 3 GHz, so performance advances will come from using additional cores.

However, utilizing the improvements offered by adding processor cores requires a significant effort in software. Many programming environments are going to require major alterations before users can gain the benefits of multi-core processors. “A lot of the traditional programming languages may be closed for renovation,” says Dr. James Truchard, National Instruments CEO.

LabView, NI’s programming language, has been redesigned for concurrent processing, leveraging the parallelism that was at the core of the original LabView design. The upgrade introduced at NIWeek in Austin, LabView 8.5, has been optimized for multicore systems.

In demonstrations during the keynote, switching from a single processor to a quad core device running at the same speed yielded between 3.5 and 3.8 times faster speeds, showing little loss for overhead. When coupled with RealTime LabView, performance was consistently 3.8 times faster, with cores running at nearly 100 percent of peak speeds.

This ability to leverage what’s sometimes called symmetric multiprocessing also lets embedded designers automatically load balance tasks across multiple cores without sacrificing determinism. They can also assign portions of code to specific processor cores manually to fine-tune real-time systems or isolate time-critical sections of code.

Sign up for the Design News Daily newsletter.

You May Also Like