Gadgetory

you're watching Jays to sentence as a PC builder of nearly 25 years is one thing I can promise you about the term bottlenecking it is overused and under understood Misun misunderstood in order to understand why I say the term bottlenecking is overused and misunderstood your first have to understand exactly what bottlenecking is but first let's go ahead and talk about some of the questions I get asked every single day regarding bottlenecking hey Jay I want to get a gtx 660ti but I only have an AMD 4100 am I going to bottleneck my GPU the term bottlenecking realistically is just like it sounds you take something big and you're shoving it through something small generally bottlenecking is used in terms of the graphics card being slowed down by the processing speed or power of your CPU there are tons and tons of hardware configurations and combos out in the market there's literally live little overly endless combination of different parts that you can use but we're going to do today is we are going to take my GTX 680 and my Core i7 3770k and we are going to recreate some bottlenecking scenarios and you are going to learn exactly what it is that happens when you bottleneck okay enough with the stupid announcer guy whatever talking dude I was doing I really want to make a point here people people are terrified of building a computer because I'm learning that not only are they afraid to break it but they're afraid to get the wrong parts because they're they're so terrified of this term bottlenecking and that's a term that most of the time is definitely missed most of the time 80% of the time it works every time bottlenecking if you see that term being used generally it's by somebody who has absolutely no clue what they're doing when it comes to computers and you might as well just run for the hills or get advice from somebody else people who truly know how to build computers tend to not use the term bottlenecking instead of bottlenecking why don't we go ahead and use the analogy or the term of non optimized hardware because basically what bottlenecking is is a situation where you take your graphics card and it has to slow down because your CPU says whoa whoa whoa whoa whoa - simmer down a little bit I can't keep up with what it is you're trying to do and a bottleneck is when your GPU no longer runs at a hundred percent utilization simply because the CPU is sending a signal to the GPU saying I can't handle all this mess I got bronchitis and oh my god that everybody's computer is going to perform differently every game is going to run different on every single machine because of different things that are running on your machine memory allocation your operating system the amount of programs you have running just because you're not getting the same results of somebody else does not mean that you have a bottleneck we're going to take my system back here we're going to set up a couple of cameras and I'm going to do the ghetto method of filming my screen rather than using a capture device because we don't want to have the capture device slowed down any of the speeds that are going to be represented in this bottlenecking recreation scenario we're going to take this PC we're going to disable cores we're going to disable hyper threading we're going to slow it down and we're going to find out exactly what point do we truly create a bottleneck okay everybody so the first thing we're going to do is we're going to go ahead and load up Valley benchmark it's just going to test the GPU only no physics no you know CPU processing any of that we're not adding any additional stuff we are just going to see exactly what the raw CPU rendering of the GPUs informational inputs it can do so this is going to take out you know it's this is all GPU based right here so if there's anything slowing down the GPU they're only Vanka possibly possibly be is the CPU you can't blame physics or any of that so we're going to go ahead and start up this benchmark and we're going to do this here in one clean take no cuts that way you can see we didn't change anything over there obviously I guess it wouldn't really matter because we are doing our base test so let's go ahead and load up the benchmark and as you can see this is all still one one take here we'll go ahead and fast forward this thing just magic special effects stuff and fast forward it and let us see what happens okay so there it is guys right there we benchmarked of average frames per second a forty four point nine base mark score or baseline score of 1877 minimum score of twenty six point nine maximum score of minimum fps of twenty six point nine maximum fps of sixty three point nine so there is our baseline and that's what we are going to use now for the rest of these tests to see exactly what changes alright we are back and we just reloaded from the BIOS and basically we took my i7 3770k and we turned it into an i5 35 70 K by slowing the cores down to 3.8 gigahertz to match a stock I 535 seventy K's turbo speed and we disabled hyper-threading all together because that's essentially all in i7 and an i-5 is compared to each other as hyper-threading and no hyper-threading so let's go ahead and see now if we're going to bottleneck this gtx 680 at all on the GPU score or the valley benchmark here for the GPU and let's see if anything change by turning off that hyper-threading okay guys so there you have it by turning my I overclocked i7 into by turning my overclock i7 into a standard clocked I 5 we've got virtually the same results we lost it looks like a couple of frames per second on the minimum I have to go back and check the maximum but the score is almost identical and as you can see right there the GPU the GPU managed to stay at you know a hundred percent most of the time the reason why it dips down is when it's changing between scenes but there you go at the i-5 speeds exactly the same now you must be asking yourself ok well then what does it take to actually slow down this test well we're going to go ahead and take things to a little bit of an extreme ok guys and for this next test we basically turned my i7 3770k that normally runs at 4.8 gigahertz into an i5 3217 u which is a 1.8 gigahertz or with hyper-threading as you can see right here in CPU Z we are running at 1.8 gigahertz 2 cores 4 threads and there are the 4 threads right over there as you can see core 0 and core 1 active only so let's go ahead and start up our benchmark and see just what happens with the performance something tells me we're not going to see anything too drastically different here and if that's the case then we're just going to go ahead and have to move on to the next test okay guys so there we go there definitely was a dip in performance we went from an average down to thirty-eight point to a score of 1600 a minimum frames per second dropped from I think 23 ish 22 ish frames per second down to 14 but look at our max frames per second didn't really change much so what does that mean that means there's something in this test there's something graphically that happened in this test at some point that made the minimum FPS drop down to fourteen point seven because the CPU couldn't handle that much data that came at it now who knows what point during the test that was it could have been the light bloom it could be these clouds it could be just about anything but as I'm sitting here watching this it's running at 40 frames per second here on an ultra HD benchmark I mean the the benchmark on this is on Ultra HD it's everything turned on 1080p light blue motion all that crap is going and we just did it right now tricking the computer into thinking that we're running on basically an i3 a dual-core with hyper-threading now I can hear the haters already I can hear you guys I can hear you through the videos that technically not even done yet saying flawed tests flawed tests that's a flawed test and I'll tell you why because you ran a synthetic benchmark and because of that the computer already knew it was coming and therefore it didn't have to actually process any raw data that's being thrown at it from staying you know battlefield 3 64 player server because the data is constantly changing it's left oh it has a lot lots of processing it has to do on top of what the graphics card doing see the graphics card you knew was coming because it was a synthetic benchmark and so it really wasn't much of a test being thrown out to CPU so it's fraud yeah that's pretty much what I hear you get all you nerds out there saying yeah I know what's happening here and it was happening here because Billy Arnold does gaming and stuff like that but ok so that's what you guys want to test next let's go ahead and do it and just like we did last time we've got to go ahead and do our baseline test for I'm going to go ahead and throw my overclocking setup back on this so that we can see exactly where battlefield 3 on 64-player Maps starts to bottleneck so now we're in the game we're gonna go ahead and turn on the draw yes command so that we can be frames per second we're getting here as we can see in this map I'm gonna eighty in a map or in the menu system I should say so it's gonna get here in the game in state we got if I can't kill some noobs here so we just spiked over 100 frames per second there for a second I just showed that new person's response am dead and 150 hundred sixty two frames per second right there good lord ah didn't see that guy sitting there hundred sixty-six frames per second Wow okay well my you guys are wondering why I don't have a GTX 780 yet okay my lion hundred seven hundred ninety two hundred forty point of the wall do you want it frames per se yeah I can cut you pull it of a different conversation here but you tell me why on your GTX 780 so she guys were talking I'm having a 680 JZ disrepair daddy's GPU yeah go get me started on a rant about those you guys you know nothing about hardware whatsoever trying to give me a hard time about what I use don't even go there sister don't even go there I'm getting 60 frames per second so you're dead what's gonna get mark around him with rockets let's kill these guys let me tell you what's on there I'll put it aside you can see well over a hundred miles per second I mean let me take 150 the thumbs and this is a 64 man in Memphis you can see like every five and reviving a looking into a puddle cutting he did ah I'm good okay so there you go I 7 right there 130 160 s 64 playin that explosion okay so that was with all settings on ultra we were sitting up in the 100's in the menu I'm getting 162 frames per second epic menu frames so yeah anyway let's go enter this thing into an i-5 and see what happens guys we are back and now we have turned once again my i7 into an i-5 you can see it right there right in 4 cores 4 threads turbo speed right now it slowed down but it is maxed out at there it is 3.8 gigahertz it's not showing anything because it's not running any load but anyway we're at stock speeds take my word for 3.8 gigahertz is where we are and there are the four core showing over there with four threads only so let's go ahead and load up some more battlefield here and let's see what happens when you turn an i7 into an i-5 and let's see if you get them that their bottlenecking as everybody likes to say bottle necking sounds like a sound like a dirty makeout term hey baby do you want a bottleneck yeah okay the game has loaded so once again let's go ahead and turn on our why did it go one earth all right get out of windowed mode please there we go let's go ahead and turn on our draw FPS draw FPS there we go FPS has turned on as you can see in the menu we're sitting at 70 frames per second this is a different map but it shouldn't really make too much of a difference here important thing to note is let's take a look at CPU usage as you can see right there it is running at 3.8 gigahertz as I mentioned 4 cores 4 threads and we are sitting at about 75% CPU usage so we will kind of we'll come back and revisit that as we're playing the game and see what happens to that number so let's go ahead and get the game going you can see the phrase you can see the shape the frames per second right there in the corner so keep your eye on that as we do our thing here once again as I mentioned 64 players in this bad boy and let's go see what happens here let's go kill some noobs whoa I died I was boobie no way I killed him I soaked killed him hmm Ian price per second they're looking pretty good were sitting in the 80s sometimes pop up into the 90s I got there in the hundreds silly guy looking at a wall I'm sitting at 100 110 there yeah let's get camping we're gonna eat up their necks pretty pretty crow yeah frames per second as you can see drop down a little bit not a lot a couple frames per second but I mean that could be a little bit of variance there in the map but for the most part you don't want your frames per second to be greater than the refresh rate of your monitor anyway that's going to create screen tearing which in my opinion is worse than this whole bottlenecking concept but let's go ahead and start to see let's let's look now at slowing this down and see where we start to truly create a bottleneck because as I bring up the second camera here and show you the history on the GPU here while we were playing look at that it was still pegged at 100% the entire time so those little those little dips there where you see it go down that's basically where I died and when it goes into the menu it's not really rendering much so bottlenecking is going to be when this GPU usage is no longer being fully utilized where the GPU says I have to slow down because the cpu can't keep up so let's find where that mark starts to happen shall we okay tech addicts we're back and we've got this AI 7 turned into an i3 3220 CPU I went ahead and one with the 32 20 on this game test instead of the 32 17 you because that's a mobile processor this here is now more commonly which you find people asking 30 to 20 or you know low-end I 5 and as you can see we've got 2 cores 4 threads running at it says 1.6 but this 3 point three turbo boost and there you go there's your four threads and core one or core zero and core one proving it is a dual core now okay so the game is loaded up let's go ahead and do our whoops our draw FPS on here as you can see the menu we're now getting 46 to 60 frames per second dish let's go ahead and deploy as you can see our face for second half dropped down a little bit here into the 30s and 40s it looks like especially all these explosions it's not entirely choppy but it does start to feel a little bit jittery and clearly it does feel like there's something being held back here but it's sitting in the sixties anymore right there 100 frames per second again whoa I'm getting good story my life engine died everybody died oh my god double kill though with my 93 are a pistol it's a pistol now take a look at the frames per second here you can see they clearly are playable when you were sitting between the 70s 50s we do get some below spice into the 30s which is definitely a little bit less than desirable but it's still very playable okay guys so I've loaded up another 64 player map right here as you can see as I'm getting shot and I'm dying but this is what bottlenecking looks like in terms of specs guys do you see how the CPU usage right there in those green bars it's completely pegged out it's sitting at ninety four point four frames or 94 percent utilized and you see how the GPU is just sitting there bouncing up and down going okay whenever you're ready it's going to hop into the game here and you're going to see once again this GPU usage is going to drop down slightly especially as I start moving around which is a lot difficult this is hard to do with one hand but look at that GPU usage how did I not die there was a ok well anyway GPU usage is sitting there in the 50 CPU usage is sitting there in the 90s guys that's a bottleneck now let's go ahead and talk about what we've learned in this video there you go guys bottlenecking as best as J's two cents can bring it to you in real-world scenario where you can see exactly what bottlenecking is and what bottlenecking isn't so guys here you go I hope you take this information I hope you use it you've enjoyed this video hit the like button if you know somebody who is just horribly misusing the term bottlenecking please show them this video and share this video with anyone that you think it may help when it comes to building their PC and together we can educate people on the proper terminology of building your own PC and when it comes to bottleneck remember guys it's hardware miss optimization or whatever it was I said I'm getting poor Hardware optimization there you go guys I'll see you next time it's been Jase descents