 So what do we have here? So what's the cluster? Well, we talked about this yesterday, but let's discuss with our schematic picture. So Seema, what are the different parts we see here? Yeah, so here we have like a technical overview or this kind of like hardware overview of the cluster. So cluster is more than some of the parts. So basically cluster is a collection of different things that you can then use. So it has a lot of compute nodes. So it has a lot of these computers that you can use to run your code. It might have some GPU nodes that have these special GPU accelerators. It has usually some home drive, which is mainly for like your SSH keys and that sort of things. It might be that it doesn't have it, but in our cluster we have this home drive. And then it usually has like a bigger faster file system there. And all of these things are accessed via this login node. And all of the resources or all of the computations are managed usually by this Q system that we'll be talking about later. Oh, I guess we shouldn't forget there's a hidden element in this picture, which is the network. So we see all these lines connecting things, but that network is actually a pretty big deal also. And has is very fast to move all this data and stuff around both fast and low latency. So what's the difference between a cluster and a bunch of computers you can connect to? Yeah, so so the difference is maybe we'll talk about this next like this is the next lesson. Yeah, let's. Yeah, in the next lesson we'll talk about it a bit more, but it's even more than just this queue that we'll be talking about later. It's also about like organizing the whole thing. So with a cluster you usually have people already also involved in maintaining the cluster. So the cluster is like if you just put like a classroom full of computers and let everybody go to those computers. And if you have just like you just collect a random bunch of computers somewhere that doesn't have any organization for it. And it doesn't have any like a cool install software who manages that the system is up. What sort of things like all kinds of things that you might need to manage in a cluster. So cluster is also the people behind it. So us in this case. So people who try to make it so it's usable for every user and try to make it sort of like like managing this kind of like a big system. It requires a lot of resources. So that's why it needs to be certain in a certain sense homogenized. So it needs to have certain common things in order to be able to manage it. And that needs us to do choices like choose things that we want in the cluster and make some choices like sort of the average user can have the best use case. But of course we are dealing with the averages, but every one of you is an individual. So you might have your own individual thing that you need to run in the cluster. So it's this kind of like a combination of all of this. Yeah, let's talk about this in the next lesson. So if we scroll down, let's talk about our cluster. So the one at Alter University. So it's pretty mid-sized like it's large for universities in Finland, but not that large compared to some other university clusters. But a big reason of that is because Finland has CSC, which is the national super computing center for academia. And the way it's organized here is in Finland, they have the biggest computers so that way universities don't need to run their own so much. So our cluster is a bit unique because it's very heterogeneous, which basically means there's different types of computers in it. And in fact, I'll claim that Triton is a very environmentally friendly cluster because we, let's see, because many clusters are made, they're used and then they're all throw away at the same time, and then they get a new one. But here we have the same cluster. We use every part as long as it's worth it for the electricity cost. And then we throw away just that part and get new nodes, processors, GPUs to replace that. So we extract all value out of every component before recycling it. And that makes it literally a ship of the Zeus. So once I was asking, it's the same cluster as it was 10 years ago, but every part has been changed, except maybe a few cables here and there, which I think is pretty cool. And also, like, similarly to, let's say, your research group might be like people change for the research group, but similarly, the cluster, a lot of the cluster is maintained in the minds of the people who manage it, or like the people who manage it. So I'd say the oldest part in the cluster are actually us. The oldest part are the people who actually manage it. Yeah. Okay, so there's a good next part down here, which is called getting help. So this is something that I think we could talk a bit more about. So I guess as we've seen from our bugs we've talked about in the icebreaker part. I mean, using technology is hard, and that's what we're here for. We actually really like it when people come and ask us questions. So don't feel ashamed to ask us. So what are different ways of getting help? Or how would you say, how would you recommend someone to approach us? Well, it usually like the easiest way is to join our daily garage. So we, you know, to have like a daily garage where people can ask us questions, but there are other ways of connecting to us. So we have an issue tracker, like an interviewer in another university, your site, most likely has an issue tracker or some sort of like a health email address or something where you can send a question and somebody will answer it. So the easiest way is just to get in contact and usually check the health pages. So what do we recommend for getting help? That is usually the best way. So can you tell us more about this garage? Like what is it exactly? Yes, so the garage that we organize a daily is basically like a Zoom meeting room where you can come with your problems. You can ask about your problems and there will always be someone there. And then based on the problem we usually like go into breakout rooms and try to solve the problem. So you can come with whatever kind of a problem you have regarding cluster computations or maybe computations in general. So optimization problems, you can't get your script to run, you miss some software or something. If the problem is something that we can fix during the garage time, we usually fix it. If it is something that requires like longer like testing procedure, for example, installing a new software, something we usually create an issue out of it so that we can follow up on it. And if it's something that requires like a lot of extra work, then we usually involve RSCs or research software engineers so that they can help the user, let's say, optimize their code or do some modifications to the code that they need. So what's the difference between you and the research software engineer? We talked about what, and you can summarize what the research software engineer is again for those who weren't here yesterday. So yeah, I'd say that the difference is that research software engineers, they can use their time to actually code stuff. So I can give usually consultation on what they should do, what sort of things would fix the problem, but I don't usually have the time to actually fix the problem or actually optimize the code or something. I can give an intro, tell what should be done. But the research software engineers, they are paid and they are hired for actually making like projects, consultation projects for the researchers so that the researchers can then, yeah, they code done by somebody who knows how to code. Yeah. And then there's, if we scroll down, there is a quick reference here, which you should probably open, and that's basically that for this lesson. Yes. And we also have printable cheat sheets for this too, so that might be helpful. Okay. Should we go on? Yeah. I guess, what's the point of the story here? Ask someone. Yeah, basically, like the cluster is more than just the machines, it's also like the whole community, and you can be part of the community as well, because it's a shared system. Everybody shares the problems in the cluster, like somebody has encountered the problems before, ask for help. Yeah. Okay. We're skipping this page. It's not in the schedule. It's.