Graphics processing units (GPUs) have come a long way. What used to be a tool that simply took binary data from the central processor to render images … could now help fight cancer. Now that’s progress.
We went to the frontlines to chat with Samuel Alt, technical support specialist at Ingram Micro, who gave us his take on the past, present and future of GPUs. He’s fielded thousands of critical tech support calls and has extensive, hands-on experience with GPUs, CPUs and beyond.
Everyone has their definition of GPU. What’s yours?
To boil it down, it’s a piece of hardware that accelerates specific workloads. To expand on that, a GPU is made up of a programmable logic chip specialized for display functions. It renders images, animations and video for a computer's screen. GPUs are located on plug-in cards, in a chipset on the motherboard or in the same chip as the CPU.
The casual consumer probably thinks about gaming when it comes to GPUs and graphics cards. What else is part of the conversation?
So much more. Modern GPUs are engaged in the most complex calculations. To name just a few, GPUs play a role in deep learning, big data research, AI, data centers, mobile phones, workstations and, yes, gaming.
How far have GPUs come?
Over many decades, the GPU has evolved from a single core, fixed-function hardware used solely for graphics to a set of programmable parallel cores. We’ve come a long way.
What GPUs can do today is mind-blowing. What do you credit for the extreme progression?
First, GPUs and CPUs working together. For example, in 2010, a supercomputer in China achieved the record for top speed using more than 7,000 GPUs in addition to its CPUs. A better example for the average consumer is this: you return from a vacation where you recorded tons of 4k footage. It would take you awhile to put a movie together using just a CPU, but a GPU can speak that language and vastly increase your speed.
What else helped speed up the GPU evolution?
Competition. For years, it was an intense rivalry between two major players. The pressure to grab market share forced both companies to push the limits of what the technology could do, including vertex blending, specular shading, waves, refraction, shadow volumes, bump mapping and elevation mapping. The competition resulted in more efficient products.
What’s an example of GPUs being used to better humanity?
Cancer research teams are using high-performance computing and GPU-accelerated deep learning to better understand the role of “driver” genes in cancer. They use DNA sequencing data to rapidly build a cancer’s “family tree,” which reveals mutations that turn healthy cells into malignant tumors. It’s amazing how GPU technology may enable us to predict how these tumors evolve.
Speaking of tech bettering lives, read 4 IoT tips to make your city safer.