Why should I use a server?
Q. Why should I use a compute server? Why not just run my application on my home or office computer?
A1. You may not need to? Refer to the next answer.
A2. The primary reason is the scale of the job. You need to run processes that are bogging down your computer and you need more cores and more RAM. Our computational servers have far more cores, RAM, and storage than the typical desktop system. Currently, the standard is 12-24 cores and 128GB RAM.
Your job takes hours, days, or even weeks to run, and you can't have your computer tied up just on that task. Servers are built to remain stable and run processes for very long periods of time - even months or years! And they are dedicated solely to computational tasks: we're not running iTunes, Office, or other standard applications on the servers.
You need to be able to compute against very large data sets, larger than you can reasonably fit on your computer.
You don't want to run the risk of your computer going to sleep, battery running out, etc. and the processes getting killed.
You need to take advantage of a specialized architecture you don't have such as an Intel Phi processor or Nvidia GP-GPU.
If your system meets your needs, then stick with that. But, if you're running into any of the issues mentioned above, then explore our research computing systems. Keep in mind it won't necessarily run faster than on your computer, but it will most likely scale larger than your computer.
There are additional reasons for using a server outside of hardware requirements, such as:
You need access to software that's not available on your system.
You have a software license that can only be installed on one machine, but several people need to use it.
You don't have the operating system installed that the application requires.
You want to have experience running on a remote enterprise scale system.