Ice cream has been around a long time and we don't need any help knowing we like it. But, we do need help understanding why we like it and that's why scientists at the University of Edinburgh are firing up the Cray supercomputer to better understand ice cream.
Ok, you science fans know it's more complicated than that. Supercomputers help us understand complex systems and interactions, right? Well, according to Dr. Alan Gray one of the lead scientists on the program, ice cream is more complex than many of us would believe. Its base substances — including different flavoring elements — interact in a variety of ways. Plus, add time into the equation and that affects the interactions as well.
It's enough to give you brain freeze.
That's where the Cray supercomputers at the Edinburgh Parallel Computing Center (EPCC) come in. They started with a massive 200 cabinet Cray XT5 supercomputer using 200,000 CPUs, but switched to a 10 cabinet Cray XK6 that uses a combination of GPUs and CPUs that offered better performance and scalability.
The 10 cabinet hybrid system integrated 936 NVIDIA Tesla GPUs with standard x86 CPUs turbocharged the process — with performance two and a half times faster than just using the CPUs alone.
So, I'm no computer genius and I am not intimate with the workings of Cray supercomputers but my take away is that not only did the team find out more about ice cream, they found out more about their computing systems as well.
Multitasking is a good thing.
Getting back to the ice cream, the team was using the simulations they ran as part of a larger study on soft matter in general and how it interacts in order to improve a vast number of things we all use every day.
So before we get all freaked out about supercomputers being used to make better mint chocolate chip (though personally I'm fine with that), we can rest easier knowing the findings can be used to help improve the quality and shelf life other items. These could be things like paints, cleansers, lubricants, and sauces among other things.
Sauces?
Is it just me, or does the Edinburgh team seem to have a thing about food? Not that I mind in the slightest…perhaps mashed potatoes can be added to the next test.
I may joke about it, but we all know food is a multi-billion dollar a year business so if the scientists at Edinburgh crack some kind of food code, they could probably pay for those supercomputers pretty quickly!
NVIDIA Blog, via Geek.com
It's enough to give you brain freeze.
That's where the Cray supercomputers at the Edinburgh Parallel Computing Center (EPCC) come in. They started with a massive 200 cabinet Cray XT5 supercomputer using 200,000 CPUs, but switched to a 10 cabinet Cray XK6 that uses a combination of GPUs and CPUs that offered better performance and scalability.
The 10 cabinet hybrid system integrated 936 NVIDIA Tesla GPUs with standard x86 CPUs turbocharged the process — with performance two and a half times faster than just using the CPUs alone.
So, I'm no computer genius and I am not intimate with the workings of Cray supercomputers but my take away is that not only did the team find out more about ice cream, they found out more about their computing systems as well.
Multitasking is a good thing.
Getting back to the ice cream, the team was using the simulations they ran as part of a larger study on soft matter in general and how it interacts in order to improve a vast number of things we all use every day.
So before we get all freaked out about supercomputers being used to make better mint chocolate chip (though personally I'm fine with that), we can rest easier knowing the findings can be used to help improve the quality and shelf life other items. These could be things like paints, cleansers, lubricants, and sauces among other things.
Sauces?
Is it just me, or does the Edinburgh team seem to have a thing about food? Not that I mind in the slightest…perhaps mashed potatoes can be added to the next test.
I may joke about it, but we all know food is a multi-billion dollar a year business so if the scientists at Edinburgh crack some kind of food code, they could probably pay for those supercomputers pretty quickly!
NVIDIA Blog, via Geek.com
Posts
