The World's Largest Computer Chip And What It Does

• What Is The Wafer Scale Engine?

  
  via Giphy

  On August 19 2019, California based start up Cerebras Systems released the exciting news: they had built the world's largest computer chip. Named the Wafer Scale Engine, it's beyond anything that has been seen in technology so far. The computer chip is designed to drive complex artificial intelligence (AI) systems, from surveillance software to driverless cars. The future is closer than ever, and it's all thanks to this big computer chip. 

• Advertisement

• Why Is It Important?

  
  via Fortune

  Until now, the focus has been on making smaller and faster computer chips, so they can fit inside smaller devices. Remember how thicc the original Nokia brick was? It had a much larger computer chip than your slim smartphone does now. But size isn't the issue now. 

  The people over at Cerebras Systems wanted to make a computer chip powerful enough to operate complex AI systems. While most high-powered computer chips are a collection of chips stacked on top of a silicon wafer, the Wafer Scale Engine is a single chip on a single wafer, allowing for much faster processing capacity and transfer speeds.

  The most powerful GPUs have 5,000 cores. The Wafer Scale Engine has 400,000 cores. This is a game changer. 

• What Will The Wafer Scale Engine Be Able To Do?

  
  via Giphy

  Now that we can comprehend how much more powerful the world's largest computer chip is, what exactly will it be used for? 

  Firstly, the Wafer Scale Engine will be at an advantage because it can handle more complex processes with less lag and lower power requirements. While this may excite gamers at the notion of life-like graphics, this chip has been built with for one purpose: AI, and specifically, deep learning.

• What Is Deep Learning?

  
  via Giphy

  You may have heard the term being thrown around, but what is deep learning? Deep learning is a branch of machine learning. The idea of machine learning is to design programs that will be able to create a result based on what it has learned from it's history. Machine learning uses structured data.  

  In deep learning, programs learn through an artificial neural network that acts like the human brain, allowing the computer to analyze data without a human programmer telling it how to process it. Deep learning can use data that is structured and unstructured. Unstructured data includes pictures, videos and language.

• Advertisement

• What Can We Expect From It?

  
  via Giphy

  The future technologies that we have been envisioning for centuries all depend on deep learning. And these technologies are on the cusp of becoming actualized. So what kind of technologies will deep learning, with the assistance of the Wafer Scale Engine, bring to the world? 

  Anything you can imagine. Driverless cars, virtual assistants (we already have Siri, Alexa and Cortana), smart houses, facial recognition in place of pins and passwords, medical innovations, and robots that can do our jobs for us. Anything that sci-fi films have predicted in the past 50 years are probably going to be a reality in the next few years. 

• Where Can I Buy One?

  
  via BBC

  The price of the Wafer Scale Engine hasn't been announced yet, but we're guessing that it's going to be more than what you have in your piggy bank. Some critics say that the chip could be impractical to install in many data centers, and it may be harder to keep it cool. This isn't the first AI chip to be created, and it certainly won't be the last. But for now it is undeniably the biggest, and our dreams are getting bigger too. 

Tags