Gadgetory

hello good people Eber here with hardware connects and welcome to our grand finale or part three of our X 299 bill series brought to you by Intel we wanted to say thanks to all of our partners who helped out in this build it was a complicated one and we really appreciate all their help now if you recall about a month ago we put together an epic skylake X PC featuring the core i9 79-80 Extreme Edition 18 core 36 threaded processor along with some high-end components including two gtx TTI's in sli inside a micro ATX case and while i did experience some challenges billing the system the end result was absolutely stunning I mean I can leave the Bill vide or I'll link the bill video right over here so you guys can check it out but something to note here is that this is our very first Scott ik X PC and we wanted to find out how adaptable the platform was for our workflow which primarily revolves around video production but the one thing that we really wanted to focus was being able to get the best out of a high-end system like this while overclocking could improve performance by a specific margin provided you've considered cooling and power drawn to the equation that wasn't the case with our build as you saw we really didn't gain that much in terms of performance so in this video I'll be walking in through my experience using core affinity within Windows 10 and how you could set parallel workloads at the same time so basically by assigning start number of course and we'll get to that later but at the same time I also want to tackle overclocking just one more time just to see if I could squeeze just a little bit of performance out of this 18 core processor so without any further ado let's get started but first a quick message from our sponsor finally an ultra slim body keyboard with mechanical red or blue switches designed and perfected for work and play that the Sorel Graham access evolved to fit right in check it out below all right so one of the reasons why someone would want to invest in any processor featuring 14 or 16 or 18 cores in this case is primarily due to the sheer amount of horsepower it has when it comes to rendering tasks or multitasking remember this is a $2,000 CPU with way too many cores for any program I want to take advantage of so we're gonna try and optimize all the threads for our own workflow for example if I take this 12-minute 35 second 4k timeline export it to youtube 4k pre send this 18 core processor completes that task in 12 minutes and 10 seconds which is almost a one-to-one ratio and that's amazing now when you compare this result to other processors featuring lower core counts the rendered time is expected to increase which makes sense but there's also the multitasking aspect that's important when it comes to taking full advantage of a cpu like this sometimes professionals don't have the luxury of time to dedicate their system to just one use which in this case is rendering you want to be able to do multiple tasks at the same time while rendering the background without losing a lot on the performance side this not only enhances productivity but it eliminates downtime a system renders and I think here's where skylake X can come in really beneficial from a personal standpoint as a content creator I render videos on a weekly basis while working on thumbnails as well as doing other tasks in fact we do have a possibility of streaming to twitch while gaming so we're really looking for the ultimate mega tasking processor here and I think this is exactly where something like bi9 79-80 XE extreme edition is a perfect fit enter core affinity having spent some time playing around with this feature I absolutely love it and it plays a large role in optimizing for scaling X now typically if you were to render a video but at the same time also work on 3d animation or if you just want to kick back and relax by playing a game or something your Windows 10 task manager distributes those workloads evenly across all 36 threads which could affect performance this is the same case for any lower counts abuse by default the current version of Windows 10 does a pretty good job at thread scheduling in some situations but core affinity takes things to the next level essentially this allows the user to assign a certain number of threads to a particular application so for example I can scan 20 threads for Adobe Media encoder a threats for blender and 8 threads for doom or battlefield 1 this completely eliminates the need for these applications to start fighting for system resources because if you said everything in Auto and if you start running all these programs at the same time they're going to be fighting for those threads and this completely eliminates that you're basically taking control of which one's doing what what what are you assigning or which one of these threads are you assigning to a certain application and it's pretty cool executing this is pretty simple all you have to do is open up the task manager head over to details find your design application right click on that task then head over to set if any and start assigning the threads now one of the things to keep note of is that you need to have your application open up prior to sending core Finity so for example if I sign a certain number of threads to media encoder I need to have that particular application open up before even assigning more affinity because it doesn't show up in the process list within task manager same story goes for blender or just any application in general that you desire or that you want to assign a certain number of threats mind you that this is a temporary solution because as soon as you exit the application everything goes back to default so you need to sort of redo the whole process which could be a little frustrating but there are third-party programs like process lasso that allows the user to permanently assign threads but that's something that we want to look into later so to test out core affinity I'll be running three applications at the same time let's start with Adobe Media encoder blender and doom I've created three use case scenarios just to compare and see how core affinity is efficient so with case one I'm gonna set everything in Auto I'm gonna let windows task manager do its thing and basically just assign the number of threads to those applications running at the same time and with case two I'm gonna be sending twenty threads to Adobe Media encoder a threats to blender and then eight threats for doom our last case is splitting the number of threats evenly across all three applications so I'll have 12 threats for Adobe Media encoder twelve threats for blender and finally twelve threats for doom so how do we do well let's start with case one Adobe Media encoder took around 15 minutes to render the 4k video a blender completed its render in three minutes and 29 seconds while doom was completely unplayable at one per second at 4k set to ultra settings remember we're working with two gtx 10a TTI's in sli as well but it seems like there was not enough cpu resources for the game to run but as soon as blender completed its render I got an average of around 50 7.5 frames per second on doom which is pretty interesting moving up to case 2 where we assigned 20 threats for Adobe 8 for blender and 8 for Doom the 4k render took 17 minutes and 27 seconds to complete while blender did its thing in 11 minutes and 7 seconds and doom surprisingly did not stutter we got around 65 frames per second on average and it was a welcoming a change our last case splitting those threats evenly across all three applications we see media encoder completing that task in 25 minutes blender completing its render in 7 minutes and 18 seconds and we get a slight boost in fps in Doom so that makes sense because we did assign an additional 4 threats so I think Court affinity is an excellent feature to take advantage of especially if you have a multi-core processor like the 79-80 Extreme Edition CPU it's great I mean you can just being able to assign number of threads to a certain applications fantastic it's actually a lot better than having Windows 10 do its thing because as you saw we were barely able to play paddle free one or just doom in general I mean one FPS is that's unplayable you can't do anything so imagine replacing that task with something like streaming or doing something else I don't think you'll be able to do or get away with something like that because we had blender as well running in the background that just taxed out on the CPU so there are different things to take into account so this is where core affinity is an excellent option now that we've taken a look at optimizing your CPUs performance I should also discuss other options that Intel X junetta and chips and brings to the table let's start with storage because it's a crucial element especially when it comes to a workstation VC X 299 offers 44 PCI lanes with 40 being split towards dedicated graphics solutions and the remaining 4 nbme or other types of high bandwidth storage having multiple drives is key if you're working with large format files because it eliminates bottlenecks and with your drivers working in parallel that increases your efficiency in my case I configure the octane 900p SSD as my primary boot drive for the operating system applications like premiere traffic's Photoshop are the important ones you know and I should emphasize that investing in an SSD makes a world of difference when it comes to boot times and application launch times the MDOT to nvme SSD that I've installed on the motherboard is configured as a footage or ingest drive where I would offload my footage from the SD card to the drive and then use that to edit videos in premiere for my scratch disk I've assigned one of those to intel's 545 series SATA base SSD and lastly for game library I've assigned the other SATA SSD so this should help ease my workflow plus it's another way of staying organized another thing that I need to mention is that while you're rendering a video within media encoder the last few seconds just right after it completes that video I think that's where a media encoder just sort of writes that file towards your drive so having a fast drive just for that test so say for example if you have another a dedicated Drive just for exports so if you're just basically just rendering it directly to that drive it would come in beneficial it would obviously reduce many times just a little bit only if you're concerned about many times if you're not you could just use the footage drive or the ingest drive that I talked about earlier and then just keep it that way there are different ways to configuring a workstation PC in my case this works really well another thing to keep in mind is that there are some applications that can fully take advantage of two or more GPUs alongside the CPU resources and a good example of that is 3d rendering the X to 99 chipset allows for high speed GPU interface across multiple programs without sacrificing on bandwidth when compared to their mainstream XIII semi chipset take octane render for example I took a sample spaceship demo render scene and compared the results with a single GPU and combined as you can see the render times are cut by half after enabling the second graphics card now realistically I wouldn't recommend gaming GPUs for these tasks since there are workstation base GP is specifically geared towards 3d GPU acceleration but I just wanted to test out the true potential of these to 1080 T eyes ok so let me explain what I'm doing right now so I've disabled SLI on both TT x 1080 eyes and you can do that through Nvidia control panel regardless of regardless if you have the sli bridge or not so you can go through the control panel and disable SLI the other thing that I've done here is after I've actually a Media encoder to the first GTX 10 a tester so one of them is dedicated directly towards the media encoder and I've said the second GPU for Doom so you know titles like doom they fully take advantage of both SLI and single GP solutions so I've separated those tasks within those two GPUs and here are some interesting results with the same core affinity applied so 20 threads for Adobe Media encoder 8 threats for blender and 8 threats for Doom the framerate difference between both cases so disabled sli versus sli is quite frankly interesting as you can see I got around 84 frames per second compared to 65 frames per second by just disabling sli and assigning that second GPU towards doom and I think I might have an explanation here but I'm not sure if it's the right conclusion so correct me if I'm wrong in the comments down below so here's what I think I think when you enable SLI programs like Adobe Media encoder Doom and other applications they try to spread their workload across two graphics cards and this case premiere really doesn't part Media encoder doesn't really make any sense because we've seen through numerous tests that having a dual GPU solution just for rendering videos is absolutely pointless when Doom is running in the background just as Adobe Media encoder and blender I think there's a little bit of confusion with those two graphics cards because when premiere splits its workload across two GPUs then you know doom might have a little bit of a confusion as to which GPS it should potentially take advantage of so that's kind of what's going on here that's what I think that's causing you know the FPS drop when you disable or when you enable SLI but disabling SLI completely fixed it solved the problem and the results are as you saw are really interesting we got you know 84 frames per second on average at 4k said to Ultra settings while we were rendering a premiere and again rendering results haven't affected whatsoever which is again very interesting but it totally makes sense because you know SLI or non-us allow you're really not going to notice a significant difference with a high end GPU like a gtx 980ti and the last thing that i want to discuss here is overclocking because I did want to squeeze a little bit of performance out of the 18 core processor but unfortunately I wasn't able to I spent countless hours playing around the Ring ratio core voltage ring voltage si voltage and IO voltage and I was unsuccessful getting the CPU above 3.5 gigahertz on all course it seems like MSI decided to use middle-of-the-road MOSFETs on this x2 99m gaming pro motherboard and the lack of proper VR and cooling all added to the equation in fact when I ran the stress test I did notice vrm temperatures going as far as a hundred and ten degrees Celsius which then resulted in CPU throttling so all cores were running at 2.6 gigahertz which is again really unfortunate remember I'm trying to overclock an 18 core processor and a micro ATX motherboard so the limitations are pretty expected and it seems like MSI has purposely limited overclocking on this particular chip but it might not be the case for the lower thread count CPUs that until offers so for example 7900 X or the 79 60 X and that's a wrap for this X 299 series it was definitely worth the while it all started with me exploring the platform and then going through my process just building together our first Scottish X PC although we did face some challenges I think the end result was absolutely beautiful and of course in this video we talked through core affinity and of course taking you through different configurations where you can assign the number of threads which is actually a great feature if you're if you want to take advantage or if you want to optimize a multi-core CPU like the core I 979 ata extreme edition processor the other thing is that you know configuring your storage in a certain way especially if you're a video editor is crucial so you know having one as an interest having a certain drive for ingest and the other one for exports another one for scratches is important it'll definitely help your workflow it help ease your workflow that's the term that I want to emphasize on finally we want to take this time to thank our viewers for watching this series and of course let us know in the comments down below if there are any other ways to evolve this bill and of course your suggestions for robust motherboards and I could use my personal workstation PC will be fantastic I'll gladly take a look at those as well and if you're interested in learning more about the core eye 9 processor and the rest of components check the links in the description down below I mean both hardware connects thank you so much for watching and we'll see you and one