This post is an entry in The World Isn’t Flat, It’s Parallel series running on nTersect, focused on the GPU’s importance and the future of parallel processing. Today, GPUs can operate faster and more cost-efficiently than CPUs in a range of increasingly important sectors, such as medicine, national security, natural resources and emergency services. For more information on GPUs and their applications, keep your eyes on The World Isn’t Flat, It’s Parallel.
Many people don’t realize that the parallel computing revolution that’s transforming science and industry is also having an effect much closer to home. By allowing state and local governments to make use of sophisticated data analysis, parallel computing is changing the way they do everything from plan for disaster to fight crime.
One of the most powerful tools for local planners and agencies is Geographic Information Systems (GIS) software, which analyzes satellite imagery by applying complex image processing filters. GIS data can help city planners determine emergency plans, and even set recovery and rebuilding efforts at a natural disaster.
But analyzing this complex data is a resource-intensive process. With the parallel processing power of NVIDIA CUDA and GPUs, local governments can unlock the practical value of GIS data much more quickly and cost-effectively than with CPUs alone.
In 2005 a major city in Canada experienced flooding that forced nearly 4,000 residents from their homes and caused $12 million in damages. City officials realized they needed to identify flood-prone areas and plan for ways to limit damage by initiating needed repairs. To do this, they are working with the Associates of Engineering in Ontario and using Manifold GIS software paired with NVIDIA CUDA GPUs to detect flood-prone areas and to better target necessary repairs.
Normally, this process would take several months. But parallel processing on GPUs helped shrink the compute window to a matter of days, enabling preventative maintenance to begin sooner.
Municipalities also use parallel processing to improve public safety. When it comes to crime-prevention, timely analysis of surveillance video or GIS data is critical.
GIS software can be used for tracking 911 emergency calls, overlaying crime statistics with demographics information, creating geo-registered overhead aerial mosaics and making development imaging forecasts. Thanks to parallel processing, calculations that used to take almost a minute are now completed in real time, while more extensive calculations can be compressed from 20 minutes down to 30 seconds. Companies such as DigitalGlobe, PCI Geomatics and Manifold that provide satellite imagery analysis and GIS software are routinely seeing speed-ups of 10x to 60x after recoding their software for CUDA.
Likewise, UK-based Bikal IP CCTV tapped into parallel processing for EyeSoft, its proprietary video analytics software. The resulting efficiencies helped Bikal improve its real-time analysis capabilities and allow its customers to monitor surveillance video with fewer staff.
Bikal’s solutions are used by governments and corporations around the world. Its analytics software is capable of counting objects; motion and object detection; detecting loitering objects or people; and smoke and fire detection, among other capabilities. The GPU computing advantage lets Bikal’s CCTV customers improve their real-time detection capabilities and event-response times, all while freeing up other computing resources.
In applications like video analytics and complex image analysis, old computing approaches simply can’t compete with GPUs on either a cost or performance basis. The traditional method of using a CPU to crunch data in a straight line takes a back seat to the parallel processing power of the GPU, which can better handle simultaneous calculations from multiple data streams. The huge amounts of data involved in these applications – whether from images, videos or audio – are best suited for GPUs, which are optimized from the ground up to process tasks in parallel.
Perhaps even more important for cash-strapped municipalities, these complex tasks can often be run on standard PCs, making the information that can save lives and property more accessible. When it comes to helping local government protect people – whether from crime or from natural disaster – parallel processing has emerged as a real-time, real-world savior for some of our most mission-critical applications.
Tagged: CUDA, GPGPU, Parallel World
