Let’s learn what a GPU is, and how it’s differentfrom a CPU, and why GPUs that were once the focus of gaming, are now the focus of ArtificialIntelligence. your easy insightsinto data, bots, and artificial intelligence. We’ve got something called the Graphics Card(or video card).

It refers to one of these. Inside of a graphics card is what helps youcompute everything. That’s what the GPU is. The “Graphical Processing Unit”, which isa silicone based microprocessor. For those of you who are not gamers, you also have asilicone based microprocessor but yours is called a CPU, a “Central Processing Unit”.Now GPUs and CPUs differ in a few ways. CPUs are like Trucks. CPUs are on most of our computersbecause it can handle multiple tasks.

Things like spreadsheets, writing, drawing, listeningto music, etc. GPUs are like race cars. There only on computersthat need to be specialized for one task. Just like a racecar is fast but it can onlyreally carry a few different items, so they’re less good with general purpose computing needs.Ok I know I just compared a GPU to a race car.

 It doesn’t mean a GPU is better thana CPU. You need think of them as two different designed microprocessors that have two differentgoals. CPUs run different types of computations serially.GPUs run similar types of computations in parallel.

This focused race car specializationof the GPU is why they are so important for running games. Every visualization on a gameneeds to be computed mathematically. Every blade of grass, the lighting, the movementof the wind. That’s a lot of computations,

 for every second you move.But because it is so fast at computations, something started happening. GPUs have becomepopular for other computations that have nothing to do with graphics.One of them is bitcoin mining, and the other one is actually artificial intelligence. Trainingmachines. Remember we talked about how GPUs are better at handling parallel computations.Well that’s what AI like machine learning needs. It needs to be able to process multipledata points to be trained. Deep neural networks, for example, can have millions of parametersto train on.

All of which can be done faster by utilizing GPUs.Ok here are 3 very quick things I want you to remember about GPUs. The first thing, GPUsare the physical foundations of artificial intelligence. Which is crazy. Not absolutelynecessary for machine learning. The only problem is if you don’t have a GPU, training yourmachine learning model, and training AI, can take weeks.Second thing to remember is that if you don’t have a physical GPU in your computer and youwant to do these AI computations, you still can. Companies like Google offer GPUs on theCloud. And three.

There’s CUDA. C. U. D. A. that has helped with programming the actualGPU. Because of this introduction, there’s been an explosion of usage of GPUs for artificialintelligence. So you thought GPUs just for gamers.

They’realso for data scientists and anyone interested in building artificially intelligent machines.Comment down below, and tell me if your computer has a GPU or a CPU. Hint, if you have a regularlaptop computer you probably have a CPU. If you’re huge into gamming, then you probablyknow if you have a GPU or not.


Please enter your comment!
Please enter your name here