 So we're here the ARM booth and who are you? My name is Isaac Yonamoto. So what do you do? So I work for a company called Rstore and what we do is we make container deployment for high-performance computing work. So what are you showing here with this whole range of boards right there? What kind of boards are these? So these are Raspberry Pi 3s. We know that there's going to be a lot of interest in supporting ARM platform for high-performance computing in the future. So it's very important to us to make sure that the containers that we support, singularity containers, function in the ARM environment. So what I did was I built a Pi stack with a backplane of Ethernet. It gives us easy connectivity to test out singularity containers and what I found I was able to do quite easily was to put singularity on these machines and deploy a compute task across the entire stack. So what is a singularity container? What does it do? So what a singularity container does is it creates a consistent environment, consistent and reproducible environment for for you to run your compute in. So one thing that dogs a lot of scientists is the inability to get mix-and-match configurations. So if you specify that you've installed X, Y and Z software and you give that list to somebody else it may not run on their computer because they have other dependencies installed and that conflict causes a lot of consternation. And what are you showing here on the screen? So what I'm showing here is just a shell of the Raspberry Pi and what I'm able to do is I'm able to very quickly just run on the right-hand side here Julia which is my favorite. What does Julia do? It's a programming language. So what you can see is that the programming language is completely encapsulated within the container. And so I can do all the normal things that I would be able to do with the programming language. But the secret is that this isn't actually the Julia binary itself. What it is is what it is is it's a entire environment that is completely firewalled off of the rest of the computer. So inside for example if you're familiar with Unix I can find a root file system that is distinct from the outside root file system. So if we compare that to what I would do if I were outside they're completely different. So this is a big deal for the for the ARM system? I think so because Singularity is currently the most popular containerization software for high-performance computing applications. A lot of supercomputing systems will support it especially over Docker which is the sort of exascale containerization technology for various reasons for example that the setup is more more amenable to the HPC environment. Docker containers have to run a root process that that spawns other other tasks whereas a Singularity container is completely self-contained binary. So that's one of the reasons. There are other reasons as well but that's why Singularity is preferred in the HPC community. And so we see that having support for ARM is a big step for ARM in order to gain it's a good step for ARM to get a bigger adoption in the HPC area. So many people in the HPC community are using Singularity containers? That's right especially among like academic supercomputing facilities. And so it's already fully like a standard kind of people are really using it. That's correct. And it's the best way to utilize the maximum performance on those supercomputers or? It's the best way to create a the performance delta between running on metal and running on Singularity is very low. It's a very low profile container. I mean there is of course a little bit of a cost but the real benefit is that you get a reproducible computing environment.