- Built in the 1950s, the SAGE system was the largest computer ever constructed at the time. It spanned 20,000 square feet and comprised 56 IBM AN/FSQ-7 computers.
- In 2022, Hewlett Packard unveiled Frontier, the world’s first exascale supercomputer. It can complete at least one quintillion calculations per second, making it the fastest computer in the world.
- IBM built Summit in 2018, a supercomputer for conducting scientific research. Summit can execute more than 200 quadrillion calculations per second.
What are the largest computers ever built? How do they compare to today’s standard models, and what happened to them once they were no longer needed?
In this blog post, we’ll take a look at some of the biggest and most impressive machines to grace the world of computing, from massive mainframes that filled entire rooms to state-of-the-art supercomputers that can fit on a desktop.
So without further ado, let’s get started!
The Largest Computers in the World
There’s no question that we have seen some pretty big computers manufactured in the past. But, which one is the largest computer ever built? If you’re curious about the largest computers in the world, read on.
Semi-Automatic Ground Environment (SAGE) (USA)
In the early days of computing, size mattered. The larger the computer, the more impressive it was. As a result, some of the earliest computers were absolute behemoths.
Though it was ultimately decommissioned in the 1980s, the SAGE system remains an impressive feat of engineering. Its legacy as one of the first large-scale computers holds an important place in history.
The SAGE system was designed to solve a data fusion problem: Radar installations across North America kept watch against Soviet bombers and needed to be networked together and coordinated with air defense missiles and interceptors. But SAGE was more than just a solution to a military problem—it was also the first nationwide computer network, predating the internet by several decades.
Additionally, the system used special leased telephone lines, some of the first modems, and CRT screens, making it the first computer system to use a “touch screen interface” via light pens.
SAGE was the first system to use magnetic core memory, making it an important milestone in the history of computing.
Hewlett Packard Enterprise sure knows how to make a splash. In 2022, they unveiled Frontier, the world’s first exascale supercomputer. This machine can complete at least one quintillion calculations per second, making it the fastest computer in the world. HPE has said that Frontier will be used to solve complex problems in areas like energy, scientific research, and national security.
Moreover, Frontier features 8,730,112 cores, and its EFLOPS scores were 1.1 according to the Linpack benchmark tests. Besides, this system is based on the latest HPE Cray EX235a design, along with a combination of AMD’s 3rd generation 64-core 2GHz CPU and MI250X GPUs.
Despite trillion calculations per second, its power efficiency ratings are also high, around 52.23 gigaflops/watt. To put that into perspective, a standard light bulb has a power efficiency rating of about 15 lumens/watt. In simple words, Frontier can perform over 3,000 times more work than a light bulb using the same energy. Not bad for a machine that weighs almost 4,000 tons and costs $600 million to build.
In 2020, the Japanese IT giant Fujitsu debuted its newest supercomputer, Fugaku. A successor to the company’s older K computer, Fugaku was designed to be the world’s most powerful computer, capable of tackling the world’s biggest problems. Among its many intended applications was slowing down climate change. While Fugaku has yet to achieve that lofty goal, it has made significant progress in other areas.
Furthermore, Fugaku was also the world’s fastest computer before Frontier took its position in May 2022. Fugaku featured 7,630,848 cores, along with an impressive 442 PFLOPS. In simple words, this system can perform 442 quadrillion calculations in one second.
Additionally, the system operates with Fujitsu’s A64FX 48-core 2.2GHz processor and features high power efficiency ratings of over 14.78 gigaflops/watt, making this system quite inefficient compared to today’s standards.
Lastly, the machine’s total cost was over $1 billion, and its cabinets weigh around 1.6 tons, whereas its total weight was 700 tons.
Summit (United States)
When it comes to scientific research, location is everything. That’s why IBM decided to build Summit, its supercomputer for conducting scientific research, at Oak Ridge National Laboratory in the U.S.
The same site as Frontier, Summit was built in 2018 to become the world’s most powerful and scalable computer. And it seems to be living up to its name: Summit can execute more than 200 quadrillion calculations per second.
Furthermore, Summit features 2,414,592 cores along with scores of 148.6 PFLOPS. Specs-wise, Summit performs in collaboration with IBM’s POWER9 22-core 3.07GHz CPUs and utilizes NVIDIA Tesla V100 GPUs. Apart from it, the machine has a high power efficiency rating of around 14 gigaflops/watt.
Summit is designed to be the world’s most powerful supercomputer, and it’s already proving its worth by tackling complex tasks in a variety of fields. For example, Summit is being used to train algorithms that can outperform human experts in various tasks, like image recognition and natural language processing.
With its vast potential for accelerating scientific progress, Summit is poised to impact the world in the years to come significantly.
Sunway TaihuLight (China)
What do weather forecasting, pharmaceutical research, and life sciences have in common? They’re all important fields that can benefit from a top-of-the-line computer. And that’s precisely what Sunway TaihuLight is.
Designed in 2016, it’s currently the fastest supercomputer in the world, giving Chinese scientists a significant leg up in the race to develop new cures and technologies. Of course, building a machine like this doesn’t come cheap—Sunway TaihuLight cost over $100 million to create and develop. But given the potential benefits, it’s likely to be worth every penny.
Furthermore, Sunway TaihuLight features 10,649,600 CPU cores, and its PFLOPS scores are 93.01. It operates on Sunway’s 260-core 1.45GHz processor and features an average power efficiency rating of around 6.05 gigaflops/watts. Thus, this machine is a bit inefficient compared to the others on this list.
Supercomputers are impressive machines, and Sunway TaihuLight is no exception. This massive machine is the fastest supercomputer in the world, and its name is translated as “divine power” or “the light of Taihu Lake.” Sunway TaihuLight was first operational in June 2016 and held the title of the world’s fastest supercomputer for two years until a newer model dethroned it in June 2018.
Even so, Sunway TaihuLight remains an impressive feat of engineering, and its performance is a testament to the power of supercomputers.
Selene (United States)
NVIDIA’s Selene supercomputer made quite a splash when it was first built in 2020. Named after the Greek goddess of the moon, Selene was designed to be a powerful tool for data-intensive research. And it didn’t disappoint: in 2020, Selene was used by the Argonne National Laboratory to study the coronavirus and potential cures.
Moreover, Google also used this supercomputer to train its natural language processing model BERT. Selene was created to support a wide community of users. And it has certainly achieved that! Selene’s platform is used for everything from supercomputer-scale continuous integration for software to research on “big iron AI” work. It’s even being used in the automotive world. All this diversity is a testament to Selene’s versatility and usefulness.
Furthermore, Selene features 555,520 cores, and its scores on various benchmarks are around 63.46 PFLOPS. The machine adopts the structure of NVIDIA DGX SuperPOD and performs with AMD’s 7742 64-core 2.25GHz processors. Besides, it has a high-efficiency rating of up to 24 gigaflops/watt.
Today, Selene remains one of the most powerful computers in the world, and its impact on scientific research is undeniable.
As the largest computer ever built, Tianhe-2A is quite a sight to behold. Developed by China’s National University of Defense Technology (NUDT) and deployed at the National Supercomputer Center in Guangzhou, China, this supercomputer is truly a marvel of engineering.
With a peak performance of over 100 petaflops, Tianhe-2A can perform quadrillions of operations per second. Plus, it operates on Kylin Linux, an updated version of the operating system made by the NUDT.
In addition, the supercomputer is equipped with over 18 million cores and 2,277,376 GB of memory. As impressive as all of these numbers may be, what sets Tianhe-2A apart is its ability to run real-world applications. The machine features Intel Xeon E5-2692v2 12C 2.2GHz processor along with 100,679 TFlop/s that further enhance its performance.
From weather forecasting to financial modeling, this supercomputer can handle just about anything you throw at it. So whether you’re looking for the largest computer ever built or simply the most powerful, Tianhe-2A is sure to impress.
We’re sure you’ll agree that these impressive machines are a far cry from the laptops and desktops of today.
With some of these behemoths taking up an entire room, we can only imagine what the future of computing will look like.
What do you think? Are massive computers here to stay, or is something even bigger on the horizon?
- The Top 10 Games from 2019 That We’re Still Playing 2019 was a great year for gaming, and there’s many of them that are still wildly popular in 2022.
- See All Teslas Ranked, From Fastest To Slowest Find out how Teslas rank in terms of their speedometers.
- How to Mine Bitcoin: A Step-by-Step Guide Jump into the crypto goldrush in this guide to mining Bitcoin.