The data produced by the particle collisions in our experiments get stored on these computers, which are spread all over the world.
The Worldwide LHC Computing Grid goes through 15PB of LHC data per year. A typical home computer might be 1TB. So the Worldwide LHC Computing Grid goes through 15000 times more data than a typical home computer, and that’s just in a year!
If those 15PB were stored on disks like CDs, the CDs would stack up to a height of about 20km, each year!
Lots of my colleagues and students have desktop PCs with three screens on their desk; I just use my trusty laptop for everything. It’s running Linux, which for me is by *far* the most efficient way to get computational science stuff done: anything can do web and email, but I need to write and run my own codes in a bunch of different programming languages, process all sorts of data into convenient forms, and use far-away *really* large computing facilities almost as if they were part of my own machine. MacBooks come an acceptable second place, though: lots of physicists like those. Despite being a laptop, it does have a hefty quad-core processor and scads of memory so I can do some fairly hardcore number crunching before I have to give in and use a big system like the LHC Computing Grid that @Meirin describes.
Comments
Andy commented on :
Lots of my colleagues and students have desktop PCs with three screens on their desk; I just use my trusty laptop for everything. It’s running Linux, which for me is by *far* the most efficient way to get computational science stuff done: anything can do web and email, but I need to write and run my own codes in a bunch of different programming languages, process all sorts of data into convenient forms, and use far-away *really* large computing facilities almost as if they were part of my own machine. MacBooks come an acceptable second place, though: lots of physicists like those. Despite being a laptop, it does have a hefty quad-core processor and scads of memory so I can do some fairly hardcore number crunching before I have to give in and use a big system like the LHC Computing Grid that @Meirin describes.