Mainframes and Supercomputers… Yes there is a difference.

We access and use mainframes on a daily basis. When you purchase an airline ticket , shop at Walmart, or withdraw money from the bank, your information is on a mainframe system. It’s rare that we talk about the other workhorse, in I.T. … the Supercomputer. It (supercomputer) is different than a mainframe, so I wanted to write about both.

Mainframe : aka (BIG IRON)

typical screen when accessing the mainframe

Yes, we know what mainframes are, their processing power and reliability. Mainframes are able to handle large numbers of transactions very quickly. 90% of the banks use Mainframes. 90% of large retailers use mainframes. Most insurance companies and government offices use mainframes.

One of the differences between mainframes and supercomputers is how the workload is handled.

Mainframes focus more on the number of transactions , inputs and output to the system. Supercomputer concentrate on specific instructions and high level computations. Mainframes accepted instructions from different inputs, like punched cards, paper tape and magnetic tape earlier on. They are measured in MIPS(millions of instructions per second).

Mainframes excel in a few areas….

1. Reliability

Through the use of software, Mainframes have self check , and self recovery capabilities. The system is constantly running checks on the hardware and software.

2. Availability

Mainframes are able to recover from failed components without impacting the rest of the running system.

3. Serviceability

The system can determine why a failure occurred. This makes it easier for a system engineer to replace the hardware , or change the software elements , while impacting the system as little aspossible in some cases.

The system also has the ability to “phone home” to IBM to report errors , so the engineer can see the errors first hand and determine the best course of action for repair.

Supercomputers :

Titan Supercomputer from Cray

Supercomputers have been around since the 1960’s with the creation of the CRAY1 supercomputer.

Each one is built for a specific purposes. Earlier on, they used specialized processors, but in recent years they have used off the shelf processors from companies like Intel and AMD. They have also incorporated graphics processors as part of their build also.

A majority of them run the Linux operation system. Each manufacturer will use their own flavor of Linux however. This is due to the hardware used. Manufacturers want to utilize as much of the hardware architecture as possible, for top performance. Supercomputers are measured in FLOPS(floating point operations per second). Floating point operations include adding and checking values of numbers, along with subtraction, and multiplication of numbers.

Supercomputer consume a lot of power, they also create an incredible amount of heat. Liquid cooling is used to handle heat distribution. Later a combination of air conditioning and liquid cooling was introduced. As far as computational ability, supercomputers are more powerful.

In a nutshell, Mainframes are able to handle our day to day tasks with ease. Supercomputers can be built with one or two specific purposes in mind.

This was just a quick look at the two different types of machines. There are a number of other tasks both, mainframes and supercomputers can handle.

Hopefully you learned something you didn’t know before. I plan to publish these short articles every couple of weeks on various topics as they relate to Information Technology.