Hyper-Threading Needed for Games? — Intel i5 vs i7 — Taking a Second Look

hello and welcome to tech deals hyper-threading for games part 2 follow-up video I recently published a video is hyper-threading good for games I will link that in the video description below and in that video we tested a rise in three 1,300 X versus a rise in five 1500 X both four core CPUs but the 1500 X has eight threads not technically hyper threading since that's an Intel trademark term it's technically SMT but who cares we all understand what hyper threading is good game Intel for making a generic term trademarked in any case in this video we now have the best of the best we're fixing two things and addressing two complaints of the previous video and we're testing for popular games okay two really popular games into sort of popular games in live gameplay not built in benchmarks and this time we're switching from middle-of-the-road CPUs to the best that have ever been made at least in this category the i-5 7600 K and the i7 7700 K in this case I have them running at 4.5 gigahertz fixed on all the cores and threads to make them exactly equal all the rest of the configuration is the same including the new video card previously we used a AMD rx 5700 XT which as I showed you in that video had problems I hate to say it but it's not fixed the 5700 XT is still uneven in its performance at 1440p and a 4k it's not as big a deal but at 1080p it still has some issues which five months post-launch come on AMD you guys make awesome CPUs what's going on with the drivers of your video cards that's neither here nor there this time we're using a middle-of-the-road r-tx 2070 super I realized that's more premium the middle of the road but at least it's fast enough that it's not going to create a bottleneck for at least for what we're doing here it's a more typical graphics card choice for a lot of people especially if you're using Intel CPUs and you're at a higher clock speed then this makes more sense it's still not an RT x20 atti but I don't actually think very many people are buying 1,200 are video cards okay if you're building a 3,000 dollar premium machine you are but that is such the top 1% of the market I'm trying to test what I think most of you might actually consider buying and I think a 500 hour video card is much more reasonable especially if you're looking at 4 core 4 thread versus 4 core 8 thread because if you have a thousand plus dollar video card budget the conversation is over you're buying an eye nine ninety nine hundred K or maybe a risin nine 3900 X at least you should be if you put your RT X 20 80 TI on a seventh-generation KB leg chip stop that you're doing it wrong any case we have fortnight overwatch division 2 and everybody's favorite Ghost Recon breakpoint and this time I'm gonna show you something cool I've mentioned it before I talked about it in the previous video but I didn't show it to you and several people said why didn't you show it because I didn't think of it I'm gonna show you the freezing what happens when the system becomes overwhelmed due to a lack of threads and how it does and does not show up in the benchmark chart so that's actually coming later in the video at the end of the benchmarks I could sit here and talk about hyper threading till I'm blue in the face but if you want a explanation of hyper threading go watch the first video linked down below the first five or seven minutes or so of that video is a detailed discussion of hyper threading i've talked long enough let's get directly to the benchmark shall we our first benchmark is fortnight on the left hand side we have the I 570 600 K and on the right hand side we have the i7 7700 K look at the CPU usage between these two the 7600 K with four cores and forth fred's is basically 100% utilized it does come off it a little bit you can see here it's an 87 or 89 but when we're in combat it's effectively being used 100% the 7700 K you can see is in combat on the right hand side and you'll notice that it's going above 50% quite often there's 60% right there so we're using all four cores and we're using some of the hyper threading please keep in mind that this is a completely clean test machine with an absolutely fresh install of Windows 10 nothing is running in the background whatsoever all we have is the epic game launcher we have fortnight and then we've got MSI Afterburner actually running the numbers up on the screen this is being recorded on an external computer using a hardware capture card so this machine doesn't even know it's being recorded you can absolutely play fortnight on a four thread machine it does suffer from some stutters and slowdowns there are some places where the game just freezes momentarily not for very long but long enough to destroy the point 1% lows but it is very playable it's just not perfect 230 frames per second average on the i-5 204 average on the i7 whoa wait a second you say something's broken no this is live gameplay and in live gameplay the actual combat in the situation and the places on the map are all going to be different and so performance is going to be different depending upon where you go I didn't run a replay I actually sat down and played the game it was completely controllable and completely smooth except for a couple of places where the game just hit and froze a couple of times which is why if you look at the point 1 percent low the 1 is accurate it did freeze a couple of places the one percent low is also accurate 93 versus 108 99% of the time basically performance is great on either CPU it's just the particular combat I was in simply favored the i5s battle in this particular case overall it worked very well regardless moving right along to overwatch overwatch has been out for a couple of years now and it is designed like fortnight to run on a wide variety of computers both very basic as well as high-end now here you can see that the i5 7600 K is again being fully utilized the percentage it just did a hundred percent 98-96 back to 98 again we are using all four cores completely we're not doing anything in the background if you were trying to watch a video listen to music if you had a Chrome web browser open a second monitor if you were trying to livestream the i5 7600 K would fall apart and you get a lot of hitching here if you were doing anything at all with your computer I cannot overemphasize the point that this is a brand new fresh install of Windows with everything disabled and turned off to provide a perfect test bench what are the comments that some people have said in the past is well why don't you try to simulate the average user open Chrome have a YouTube video playing or have a twitch stream playing have a bunch of stuff going on load up the task tray with you know eight things I understand the idea there I get the concept the problem is how far do you take it medium high what's your definition of high if I say typical user case well what's a typical user that could be all kinds of different things it's also not necessarily repeatable because I might have programs that are running in the background that might be auto updating one run and not auto updating another I might have a you know Chrome might be checking for updates or doing a refresh or you know various extensions might be doing something in the background in one run but not another and so it's very easy to say try to simulate a typical user environment but even if you load up the same stuff each time they'll be doing different things in different tasks since you end up with an uneven workload now it's true that you could repeat the tests five times or six times or ten times and over the course of several hours of testing and updating and rebooting yes you would average those differences out unfortunately I don't have two full time employees working for me yet you know benchmarking which is what it would take to do what I just described so it's a nice idea but yeah in any case the i7 7700 K you can see is going about 50 percent and frankly it's highly recommended even for games like this but it is very very smooth even on an i-5 224 versus 220 average 153 versus 161 and 118 versus 110 in terms of a perfect clean benchmark test bench with nothing else going on it doesn't matter I've already discussed where it might matter so we won't rehash that but Wow talk about epic performance moving on to the absolute opposite end of the spectrum we have Ghost Recon breakpoint the follow-up sequel to Ghost Recon wildlands and it is a sequel it takes place in the same universe and even references wildlands this game would use eight cores I didn't say eight threads I said a course if you have an eight core 16 thread CPU it'll go over 50% CPU usage on a say Rison 720 700 X or an I 999 hundred K on an i7 9700 k-8 core 8 thread CPU yep it'll peg it out to a hundred percent which is one reason why if you like to play the latest Triple A games buying an i7 9700 K with 8 cores and eight thread strikes me as kind of a foolish investment because it's already matched it's it a hundred percent and it's got nothing left in today's games right now where's your future there's no future in that so that's why I say an 8 core 16 thread is the new minimum for a new purchase for a serious gaming PC that wants to be able to play the new latest greatest games and the games coming out soon now as you watch the gameplay on the screen you'll notice that the i5 7600 K appears to be doing just fine it runs fine it controls pretty well the extra clock speed definitely helps I showed this game when I did the previous hyper-threading comparison where I was doing a rise in three 1,300 X versus a rise in five 1500 X both for core chips but at 1,300 X has no hyper threading okay SMT because AMD whatever you guys know what I mean and then the 1500 X does but those CPUs run at 3.6 gigahertz these are running at 4.5 and their Intel chips their recent Intel chips it does make a difference we probably have somewhere in the 40 to 50 percent additional performance range in terms of gaming performance maybe 35 to 40 percent in terms of Cinebench performance maybe per se but in terms of gaming performance and I 770 700 K is much much better than a risin 5 1500 X more expensive of course but it is a much better CPU we are being helped by the additional clock speed we are being helped by that Intel optimization that Intel gaming performance which is a real thing Zen 2 is pretty close but a 1500 XS knows then 2 it's Zen 1 first gen and so on and so forth and as you can see here we're not suffering through horrendous hitching on the i-5 7600 K but it's not absent it doesn't show up in the chart because the length of the benchmark but if you look at the real-time numbers on the screen I'm showing you this section of the benchmark for a reason the numbers on the top of the screen the the white numbers the left number is the real-time frame rate as what you're watching on the screen right now the next one over is the frame time is how much milliseconds are between each frame and then the graph beef actually have set to be the frame time graph as well then the middle white number is the average number the second from the right is the ongoing real-time 1% low calculated as the benchmark runs and then the far right number is the real-time calculated 0.1% low notice on the i-5 it's currently 2 frames per second it will not be in the chart there about the C and that's because I played for about 20 ish minutes on each of these configurations and eventually the laws of averaging of adding up all the numbers the hitches that occurred became beneath that point 1 percent low and uplifted it which means that it really wasn't that big of a deal it happened but it happened like twice just briefly and so it really wasn't a big deal not nearly as big a deal as it was on the horizon 3 1300 x because when I did the original video I talked about how horrific ly bad it was and how badly you needed the hyper threading well 4.5 gigahertz plus intel cores makes up for that to some extent and of course we have a much better video card this time around as well which never hurts because the AMD card was kind of yeah 79 frames per second average on the i5 versus 98 on the i7 but the more interesting number is the one percent and point one percent low 45 to 73 is a huge difference and that's legit over 20 minutes of testing there really is that big of a difference the drops the stutters the slowdown on the i-5 is real it's playable because this is basically the best four core for thread Intel chip ever made and it's running at 4.5 gigahertz which is a nice speed and that helps tremendously compared to the performance on the rise in three or if you had something old or like an i5 2400 would be god-awful in this game it would just be horrible old it just does not cut the mustard anymore same thing with the 0.1% low 32 259 is a dramatic difference it's not as bad as rising because we've gone from a middle-of-the-road chip to the best that ever existed and the best that frankly ever will exist to be completely honest because the world is moving on to six and eight core CPUs but it just does demonstrate that given more clock speed and given a better chip it does at least alleviate some of the pain of being thread count limited the division to a frankly better optimized game than Ghost Recon breakpoint I've been very impressed at how well this runs on a variety of machines here on the screen I'm doing two side missions there in different places but it's not major campaigns or just side missions so they're fairly comparable in terms of number of enemies and the size of the environment that I'm fighting with and not if you look at the real-time numbers up on the top of the screen of course the i7 is faster because this game does use more than four threats but it's not terrible on the i-5 at least at this point in the benchmark I'm gonna show you I'm gonna go fullscreen and show you a different spot here in a second but most of the time the i-5 is fine and I suspect this is what happens a lot of times when people say well I've got a fork or four thread i-5 what are you talking about Willis everything's fine I get great benchmark numbers what are you smoking well you do mostly and you've learned to live with hitches and slight delays and longer loading times and poor multitasking performance you've learned to live with that look on the right-hand side on the i7 73% 78% CPU usage that is a lot more than 4 cores 4 threads right there we're using almost all of the threads on the i7 but not all the time most of the time you can get away with not having it most of the time it's ok and not necessary by the way I've we missed it but if you look back at the left hand side on the i-5 noticed that our 1% real-time low is down to 5 yeah hitching whoo-hoo for the wind it really does happen it is a thing it's not game killing when you've got 4 point 5 gigahertz of cable 8 cores most games will kind of push through it not so badly on the lower end chips and older i5 from second third or fourth generation with ddr3 a first-generation Rison they're all gonna suffer much much more but this is sort of the best-case scenario and again this is a completely clean test machine with nothing running on and I'll get point your attention to the CPU usage on that i7 and once again I'll remind you I cannot make this machine any prettier when it comes to the fact that nothing's running in the background no tasks nothing in the tray nothing is installed it's as clean as a whistle so take it for what it is but if you want to play current games and you've got a fork or four thread chip you can get away with it but not for much longer not if you want to keep up with the new stuff we're gonna jump the head here we've gone full screen with the i-5 7600 K because I want you to watch something here this is not paused by me this is the game and this happens several times during the benchmarking this is a freeze the machine just freezes and then you can get control of it again it's not for long and it's not often enough to make it unplayable let me show you that again here we are running towards the police car our point 1 percent low is currently 43 frames per second everything's great right and freezing and freezing and freezing and notice that our 0.1 percent load dropped to 1 notice the 1% low only dropped to 58 and that's because this is a 20 minute benchmark and that level of freezing does not destroy a 20 minute benchmarks 1% low when you're looking at benchmarks how long it was tested how it was tested you cannot just look at a chart and gain everything you need to know and this is why you so much live gameplay testing I do very little replays occasionally I do it really does make a difference now what the pause you just saw if that occurs once every 5 minutes of gameplay does it really matter well in the middle of a firefight it would by the way this is pretty epic an entire group destroyed all in one shot it doesn't really matter but it would kind of suck if it happened and an at a basically a bad time or if you're doing PvP or something else so it's tolerable but I wanted you to see that so you had an idea of what I'm talking about when I talk about like system freezing which happens just not all that often ly you got 4 point 5 gigahertz of cable 8 cores 110 frames per second on the i-5 versus 122 on the I 767 and 93 at the 1% low and yeah 1 276 on the point 1 percent low this right here is why I continued to see people and I think a lot of people who have these older machines go but but my benchmarks fine man what are you talking about well it is but you've learned to live with that and it it's not often it's not all the time it's not so much that it destroys your gaming experience but it does exist how much it exists depends upon the game how much you have running in the background how fresh your windows install is what video card you have what generation CPU you have what games you're playing this might not be an issue lots of you listening to me are probably fine on a fork or for thread chip and then others of you listening long ago left it because it was a completely terrible experience depends upon what you're doing in this case it's playable at least what you've got the best but it's time to upgrade thank you all so much for watching all that hopefully it was informative and interesting if you like these kind of follow-up videos well you know where the subscribe the bell notification icon like button and the comment section so those YouTube BOTS have something to say oh engagement yum-yum links in the video description below well to be completely blunt the only thing on this desk that any of you should buy is an r/t ex 2070 super and frankly the 2060 super is a better deal if you already own them I mean if you already have a 7700 K you can still use it you don't have to replace the 7700 K a 1500 X well because it's so easy to upgrade is probably time to upgrade and you don't have to upgrade a fork or four thread shop I mean as I said during the benchmarks it still does the job for a lot of people if you're playing overwatch who cares unless you're super competitive or you're live streaming or recording or doing three other things in your machine and have multiple monitors and that's a separate conversation completely clean test bench with that absolute fresh install of Windows 10 and nothing running on it so yeah but if you're buying new if you're ready to purchase maybe you have a 5 year old computer a 10 year old computer maybe you don't have a computer and you're just using an Xbox and you're ready to buy don't buy any of this I mean unless it's like you know 20 dollars what you should be buying is like a risin 5 2600 at the bare minimum like a hundred and ten ish dollars at the moment and there's no excuse to buy less I do think the extra money for a risin 720 700 is worth it another 30 $35 or so gets you 25% more cores gets you eight cores instead of six court well technically that's 33% more cores it gets you more cores more threads it gets you a better cooler the 2700 comes with the larger race spire versus the race stealth on the 2600 and while yes it's only a second generation rides in it is better than the first and it kind of bridges the gap between the first and Intel level performance and at the prices they're currently at now I wouldn't buy any four or anything if you're making a purchase today so there's that like this video if you like it share it with your friends if you love it remember to subscribe bail notification comment and everything else down below I appreciate you guys watching I'm looking forward to reading your comments in the comment section below as to what you guys think of this follow-up video I don't do a lot of follow-up videos on this channel so this is gonna kind of be an interesting experiment I do appreciate those of you who have taken the time to watch two gold stars to anybody who got to this point thanks so much and I will see all of you next time

Channel: Tech Deals

Views: 49,772

Rating: 4.867712 out of 5

Keywords: Hyperthreading, hyperthreading vs no hyperthreading, hyperthreading gaming, hyperthreading cpu, hyperthreading vs cores, hyperthreading on vs off, What is hyperthreading?, 1300x vs 1500x, Ryzen 3 vs Ryzen 5, i5 vs i7, i7 vs i9, i5-7600k vs i7-7700k, i5-8600k vs i7-8700k, i9-9700k vs i9-9900k, Cores vs Threads, SMT, What is SMT?, How Hyperthreading Works, How SMT Works

Id: llw4N-8XqFg

Length: 23min 18sec (1398 seconds)

Published: Mon Dec 16 2019