Transcript:
I have downloaded already the exact same image. So if I jump over here to the Odroid xu4 queue. I’m gonna see the exact same thing in my downloads there. They are so a couple of differences in the operating system. Obviously the Raspberry. Pi Raspbian just gives me a thumbnail kind of like, circa Windows 98 whereas this is more like a it’s a full operating system with I can see the actual preview of the image. So what I want to do is I want to right. Click, and I’m going to hover over opening it with the GNU Image Manipulation program. I’m going to do that on the Raspberry Pi as well and then we’re gonna go side by side and just see how they compare as far as speed goes to load this up in the GNU Image manipulation program. So as long as I don’t bump the mouse and click in three-two-one, the Battle of the SPCs begins Odroid XU4. Q came up quicker and started loading pretty much identical, yeah? I would say I did even though the screen seemed to come up quicker on the Odroid. Exactly the image was in one of the things that I do need to address with the Odroid XU4. Q is you’ll see that? I’ve got some strange like the backgrounds. Here are all blue. That’s like the theme of the operating system. I’ve just got to fix that on the Raspberry Pi out-of-the-box. It looks nicer whether or not that’s an issue. Couldn’t tell. You don’t think so, alright? Let’s go, we’re gonna right-click and we’re gonna try something here. I want to do a scaling operation. So usually scaling is a pretty intensive thing to do. That’s like increasing or deep, you know, changing the size of an image. So on my ex U4Q. I’m gonna go scale image and I’m gonna set the height to 8,000 pixels and we’re gonna use 72 DPI and cubic interpolation. It’s important to note these things because we want it to be an exact side-by-side head-to-head comparison. We want to set the settings exactly the same on the Raspberry Pi 3 so over here on the Raspberry Pi 3 image scale image again, 8,000 pixels, height, 72 DPI cubic interpolation. And I’m going to click scale because it’s prompting me. Are you sure you want to create this image? I’m gonna do the same thing over here. We’re gonna be prepared then to begin working. You’ll see that each one says it’s going to be eight hundred and ninety six point three megabytes. So on the raspberry Pi, I’m gonna hover over scale and on the Odroid X u4q I’m already hovering over scale, so if we’re ready when do the countdown with the Drumroll three two one click. OK, the Odroid xu4 cue is moving. I can see that kind of down at the bottom. There you see that progress indicator? It’s a little more significant as far as speed goes to the very PI. Raspberry Pis doing it, though. This is a great opportunity for us to have a sip of our coffee and for you to just stare at that progress indicator. Yeah, we need some suspenseful music here. I think here we go, it’s real time. Folks, keep in mind. This is a heavy operation on any system, so the this is that even happening on single board computers just shows. How far technology has come? The Odroid is ever so slightly ahead. How’s that gonna see real time, lets? See what happens, boom? Oh, droid is done, Raspberry. Pi still chugging along still. Check along that vape. I I can almost smell the heat coming off of that processor and come on. You can do it. You can do it right there we go. Raspberry Pi pulled it off way to go. That’s the Raspberry Pi 3 OK, So now next up. I want to lets lets. Try something really, really tough. This is a big image, and this is obviously gonna be really really heavy on any system. I’m gonna go filters and we’re just gonna add a 5% Gaussian blur, nice and simple. Oh, what does it say? Failed to run plug in blur, gauze failed to fork cannot allocate memory. This is the Odroid XU4 Cue that we’re looking at so it looks like I’ve hit that memory threshold. So can I just hit OK and just do other stuff? How’s the responsiveness here? If I’m clicking around, Looks like I can still do other things. I have created a rather significant image, though, Haven’t I? It is responsive. Can I do anything with it blur? Gaussian blur not gonna not going to do it, so we literally do not have enough memory to be able to work on this image now that it is, a ginormous image we’re talking 12,000 by 8,000 in pixels, so that is an unreasonable size for for any case. Let’s see if I can do anything at all with it blur. No, it’s not gonna let me do it. So we have maxed out the Odroid XU4. It is still very responsive, though as you can see. Can I scale down now, maybe? I can scale down. All right, Let’s jump back over to the Raspberry Pi 3 and we’re gonna try the same thing, right, Click, and in note filters, blur Gaussian blur, what the PI is going to allow it and the Odroid XU4 is not. How can that be? Let’s get a look at our system tools, system monitor and see what this is telling me. CPU is really low memory. We’ve we’re using one point five gigabytes, 77.4 percent of the What does the Raspberry Pi say? Do we have the same tool we do not? We may need to use that each time. Jump into the terminal. H Top tells me that we have used. We’ve exhausted the RAM completely. Perhaps there’s some swapping going on here. Yeah, eighteen point five. Megs has been swapped. Not sure if there’s any swap going on with the Odroid XU4 Q There is not no swapping happening and there is still some RAM there. So why is it not letting me go filters blur Gaussian blur, so that’s a disappointing thing right there, and that could be software, that’s not necessarily the board itself, but we do need to see in a real-world environment now. It was faster! It got me there faster. But surprisingly, it seems that gif doesn’t want to proceed. It doesn’t want to go any further, so that’s one test where, okay, the Raspberry Pi has passed and the Odroid xu4 queue while it was very much faster in the rescale operation it it failed when it came to adding the Gaussian blur to that massive image. So let’s go back to the Raspberry Pi 3 here. We’re gonna get back out of that, and we’re gonna go back to a reasonable image size. I’m gonna control Z and control Z here. So keep in mind that now we’re back at a standard image size of it. Looks like three thousand by two thousand pixels. So if I does it release the RAM, let’s find out. Yeah, okay, so it is going to allow me to now that the image is back at a more reasonable size, So let’s undo everything, get back so that we’re again comparing one to one. Okay, there’s that and there’s that so over at X U4Q, let’s add a gauzy and blur of what say he 20% Let’s make it heavy and the blur method is RLE using the wrong keyboard there, so I’ve probably zoomed in on the Raspberry. Pi, here we go and I’ve closed out of a couple things. No worries, all right, right, Click filters, blur Gaussian blur. I closed my tools there, just so you know, that’s why the Raspberry Pi now looks a little bit different, but side-by-side they are doing exactly the same thing. Raspberry Pi. We’re gonna go hover over, okay, And the Odroid X U4Q we’re gonna do the same and three two one watch those progress indicators again. The X U4Q just blast past the Raspberry Pi, though the Raspberry Pi is doing very very well. It seems at the very last second, though the Raspberry Pi locks up and then waits for a moment, and then the the thing happens whether it be blurring or rescaling. So I think the the Odroid X U4Q. I mean, the the the one test that it failed. Is this huge, massive image that nobody’s going to ever work with a twelve thousand by eight thousand picture on an SBC anyways? You need a supercomputer for that, but it did fail that. Where the Raspberry? Pi kept going so so do keep that in mind that the Raspberry Pi did pretty well. And that is again. Probably an operating system or software difference between them [Music] [Music] [Applause] [Music].