What are Supercomputers?
In its simplest definition, supercomputers are computers with much higher performance compared to a general purpose computer. The performance of a supercomputer is typically monitored by floating point operations per second (FLOPS).
FLOPS is a measure of computer performance useful in the scientific fields of computations, i.e. computations involving very small and/or very large real numbers and often requiring fast processing times.
The supercomputer is phenomenal in terms of performance. Today, there are supercomputers capable of performing more than a hundred quadrillion (+100,000,000,000,000) FLOPS, measured as "petaFLOPS".
The most powerful supercomputers are multiple computers performing parallel processing. In general, there are two parallel processing approaches. These are Symmetric Multiprocessing(SMP) and Massive Parallel Processing(MPP).
Supercomputers can be used for anything needed. What turns an ordinary computer into a supercomputer is clustering multiple high-performance, optimized computers, all configured to perform a specific task.
This optimization usually includes a private network, fine-tuned hardware, large amounts of storage. Workloads that require a supercomputer often have two things in common: they either require computation on large amounts of data or are very computationally intensive.
Today, the world's fastest supercomputers are often run by Linux-based operating systems.
History of Supercomputers
The History of Supercomputers goes back to 1964. Designed by American electrical engineer Seymour Cray, the "CDC 6600" (Control Data Corporation) is often considered the first successful supercomputer, as it performed up to three megaFLOPS. Cray used faster-acting silicon instead of germanium transistors. It has solved the overheating problem by including cooling in the supercomputer design. CDC 6600 was followed by CDC 7600 in 1969.
In 1976, four years after leaving the CDC, Cray created the "80 MHz Cray-1", which became one of the most successful supercomputers ever, with performance clocked at an impressive 160 MFLOPS.
What Makes Supercomputers Performance?
Supercomputers succeed by performing calculations in parallel. Calculating in parallel means doing many calculations at once. It's like having thousands of general purpose computers working for you on the same problem at the same time. This is actually a case in point that describes exactly how modern supercomputers work.
It should also be noted that although supercomputers provide enormous computational capacities, they are very expensive to develop, buy and run. For example, the typical power consumption of a supercomputer is several megawatts. That's why it's important to use them as efficiently as possible.
What Are Supercomputers Used For?
Supercomputers are not used for things like website hosting, application hosting, or general business computing that is running all the time. Since today's computers can also do these, it would be unreasonable to use these computers with such high-performance hardware in these areas.
Training models for artificial intelligence, mapping using seismic data in oil and gas exploration, engineering simulations in various branches, climate modeling, life science research, managing large databases are some of the usage areas.
Supercomputers are also used by governments. Especially the USA and China rely on these computers for strategic issues (such as nuclear stocks) for their countries and use advanced simulation and calculation programs.
What Are the Types of Supercomputers?
Supercomputers fall into two categories. These are general purpose supercomputers and special purpose supercomputers.
In general-purpose supercomputers, it is divided into 2 in itself. These are vector processing supercomputers and cluster supercomputers.
Vector processing supercomputers are those based on vector or array processors. These processors are basically like a CPU that can perform mathematical operations on a large number of data items very quickly. Vector processors, common in the computing industry, formed the basis of most supercomputers in the 1980s and early 90s.
Cluster supercomputers refer to groups of attached computers that work together as a unit. These are supercomputers where multiple processors work simultaneously to solve different parts of a single big problem. The "Grid Engine" by Sun Microsystems and Open SSI is a cluster supercomputer.
Special purpose supercomputers are supercomputers created with the explicit purpose to perform a specific task/goal. They often use "Application Specific Integrated Circuits" (ASIC) that offer better performance. Examples of special-purpose supercomputers include Belle, Hydra, and Deep Blue, all built for playing chess, as well as Gravity Pipe for astrophysics.