What makes a supercomputer so super? Can it leap tall buildings in a single bound or protect the rights of the innocent? The truth is a bit more mundane. Supercomputers can process complex calculations very quickly.

As it turns out, that's the secret behind computing power. It all comes down to how fast a machine can perform an operation. Everything a computer does breaks down into math. Your computer's processor interprets any command you execute as a series of math problems. Faster processors can handle more calculations per second than slower ones, and they're also better at handling really tough calculations.

Within your computer's CPU is an electronic clock. The clock's job is to create a series of electrical pulses at regular intervals. This allows the computer to synchronize all its components and it determines the speed at which the computer can pull data from its memory and perform calculations.

When you talk about how many gigahertz your processor has, you're really talking about **clock speed**. The number refers to how many electrical pulses your CPU sends out each second. A 3.2 gigahertz processor sends out around 3.2 billion pulses each second. While it's possible to push some processors to speeds faster than their advertised limits -- a process called **overclocking** -- eventually a clock will hit its limit and will go no faster.

As of March 2010, the record for processing power goes to a Cray XT5 computer called Jaguar. The Jaguar supercomputer can process up to 2.3 quadrillion calculations per second [source: National Center for Computational Sciences].

Computer performance can also be measured in **floating-point operations per second**, or **flops**. Current desktop computers have processors that can handle billions of floating-point operations per second, or gigaflops. Computers with multiple processors have an advantage over single-processor machines, because each processor core can handle a certain number of calculations per second. Multiple-core processors increase computing power while using less electricity [source: Intel]

Even fast computers can take years to complete certain tasks. Finding two prime factors of a very large number is a difficult task for most computers. First, the computer must determine the factors of the large number. Then, the computer must determine if the factors are prime numbers. For incredibly large numbers, this is a laborious task. The calculations can take a computer many years to complete.

Future computers may find such a task relatively simple. A working quantum computer of sufficient power could calculate factors in parallel and then provide the most likely answer in just a few moments. However, quantum computers have their own challenges and wouldn't be suitable for all computing tasks, but they could reshape the way we think of computing power.

Learn more about computers and processors by following the links below.

### Related HowStuffWorks Articles

### More Great Sites

- National Center for Computational Sciences
- Discovery News: Tech

### Sources

- Clay Mathematics Institute. "Prime numbers and cryptography." 2010. (March 22, 2010) http://www.claymath.org/posters/primes/
- Computer Power User. "Processing Power." 2010. (March 22, 2010) http://www.computerpoweruser.com/Editorial/article.asp?article=articles/archive/u0413/08u13/08u13.asp
- Computer User. "Clock Rate." 2010. (March 22, 2010) http://www.computeruser.com/dictionary/clock-rate/
- Intel. Intel® Multi-core Technology. (June 20, 2010)http://www.intel.com/multi-core/
- National Center for Computational Sciences. "World's Most Powerful Computer for Science!" (March 22, 2010) http://www.nccs.gov/jaguar/
- Red Hat. "Red Hat Enterprise Linux 3: Introduction to System Administration." (March 22, 2010) https://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/admin-guide/s1-bandwidth-processing.html
- Shinder, Deb. "What's the best way to increase processing power?" Aug. 2, 2006. (March 22, 2010) http://www.zdnet.co.uk/news/servers/2006/08/02/whats-the-best-way-to-increase-processing-power-39280352/