Some scientific problems require so much computation, that they can not be run on any single supercomputer. Instead, these projects rely on people like you and me to download software which runs on your computer. These scientific applications only use your idle CPU cycles, so you typically don’t notice the program running on your computer at all.
These projects include:
And many more
Don’t get me wrong, I love science, discovery, research, and progress. But it’s important for you to understand that these programs will make your computer use more electricity. When processors are idle, they only consume so much electricity. However, when a processor is working at full load, it can consume about 50 watts of additional power. This actually isn’t too disastrous. Let’s assume you turn your computer off when you’re not using it. Now let’s assume that your computer is on for 6 hours a day and you pay the national average for electricity. In this scenario, you will be paying about $9.81 per year just to run the scientific application.
Protein folding simulation
Some of these projects are very important for humanity. For example, Folding@home is using its results you help understand protein folding in order to better understand how to cure diseases and cancer. Pinching every last penny is one thing, but frankly, ten dollars a year is not an unreasonable amount of money to donate for a worthy cause. Everyone who owns a computer and internet connection should be running programs like Folding@home. Unfortunately, some scientific projects are dramatically less important. For example, GIMPS is dedicated to finding very large Mersenne prime numbers. These numbers may be interesting to statisticians, but will never contribute to a worthy cause like curing diseases. You should choose a program which interests you *and* helps humanity.
With that being said, there is one more thing which is worth bringing up. Graphics cards are extremely good at numerical calculations. NVidia has released something called CUDA, which transforms its graphics cards into number crunchers. In fact, using special software, a single nVidia graphics card can be well over 100 times faster than a CPU. So why am I mentioning this? Well, if you have an 8-series nVidia graphics card, you can download a CUDA-enabled client for Folding@home, and all the scientific calculations will be run on your graphics card!
A typical midstream GeForce 8 series card
Like the CPU, it will still cost you more electricity when you run scientific software on your graphics card, but the performance per watt can be over 100 times better than your CPU! This means that if you spend 10 dollars a year for electricity and run the program on your graphics card, the amount of work you contribute to the project can be well over 100 times greater!
Okay, so you don’t want to cough up 10 dollars a year in donations, I can understand that. If you have a recent nVidia graphics card, you can run Folding@home on your graphics card for just three days a year, and still produce as much work as someone would produce by running for an entire year on a CPU. If you run the program for 3 days a year, you can expect to spend a measly 8 pennies a year, which is nothing.
In conclusion, if you want to save money while contributing to humanity, one of the best things you can do is run Folding@home on an 8-series or higher nVidia graphics card, even if it’s just for a couple days a year.
Click here for more information about the CUDA Folding@home client application.
Posted under Saving Electricity
This post was written by admin on August 14, 2008