The Performance Gap Between Intel and AMD: A Study of CPU and GPU Bottlenecks

In our latest investigation, we put the two flagship platforms from Intel and AMD to the test to see which one emerges victorious. The question on everyone's mind is, will the new 30 series graphics cards from NVIDIA make a significant difference in the performance gap between these two processors? We're here to tell you that the answer is not as simple as it seems.

First, we need to acknowledge that no two titles operate under the same conditions when it comes to CPU and GPU utilization. This means that the numbers we see today may not reflect the real-world experience of gamers. However, this also means that our results should be taken with a grain of salt, and we must consider the unique characteristics of each game and system.

We began by running a series of tests on various titles, including Far Cry 5, which is one of those games that uses multiple cores to achieve high frame rates. In these tests, we saw that the Intel system was able to maintain extremely high CPU utilization, often reaching above 99% in certain scenarios. This suggests that there may be a bottleneck at play, and it's not just about the GPU.

On the other hand, the AMD system showed lower CPU utilization, often hovering around 80-90%. While this is still impressive, it implies that there might be some limitation on the CPU side, which could potentially impact performance. However, as we mentioned earlier, our tests were run with the new 30 series graphics cards from NVIDIA, and these features are what truly brought out the performance difference.

So, why did we run these tests with RTX enabled? The answer is simple: it's a feature that many gamers care about, and we wanted to see how it affects performance. And boy, were we surprised! When running games without DLSS (deep learning super sampling), we saw a significant increase in frame rates, often exceeding 200 FPS or more.

However, when we turned off RTX and ran the same tests, the difference between Intel and AMD was negligible. In fact, the Intel system was only able to achieve a slight edge in certain scenarios, suggesting that the bottleneck may be on the CPU side rather than the GPU.

But what about gaming performance at 1440p and above? Are we going to see a significant difference between these two platforms? Our answer is no. Even with the high-end AMD Ryzen 9 5900X processor paired with the NVIDIA GeForce RTX 3080 graphics card, we couldn't identify any tangible differences between the two systems.

So, who should you choose for your next gaming PC build: Intel or AMD? It ultimately depends on what type of gamer you are. If you're a hardcore esports gamer looking for every last FPS possible, then Intel might be the way to go. But if you're more interested in enjoying beautiful HDR and high-quality graphics, then AMD's PCIE Gen 4 interface and NVMe storage speeds make it a compelling choice.

In conclusion, while there may be some differences between Intel and AMD in terms of raw performance, these differences are largely negligible when it comes to gaming. Our tests showed that both platforms can deliver smooth, high-performance experiences, especially when running games without RTX enabled.

Ultimately, the decision comes down to personal preference and budget. If you're looking for every last FPS possible and don't mind paying a premium for it, then Intel might be the way to go. But if you want to enjoy stunning visuals, fast loading times, and a more affordable price point, AMD's platform is definitely worth considering.

Thanks for watching our latest video, where we dive into the details of this performance gap between Intel and AMD. If you have any questions or comments about our findings, please don't hesitate to leave them in the section below. We'll be back with more videos like this one, so be sure to subscribe to our channel for the latest gaming content.

"WEBVTTKind: captionsLanguage: enit should be no surprise that with 30 series uh 30 series is an announcement and the excitement around the performance per dollar the amount of people that are like hey i'm going to get 30 series i'm gonna get 30 70. i'm gonna get 30 80 but will i bottleneck today we're gonna talk about that because um that that's a very important discussion to have especially right now having trouble figuring out the right parts for your next build then the custom pc builder at microcenter.com is just for you microcenter's free custom pc builder tool guides you through choosing the correct components from start to finish and ensures maximum compatibility for trouble-free installation smart search features allow you to customize your build based on anything you wish including price features and store availability and once you're ready to purchase the items in your cart will be reserved and ready for pickup at the micro center store of your choice not sure which parts are right for you we'll then head to micro center's community forums for help and advice from other pc enthusiasts like yourself to get started planning your next build and to check out the free custom pc builder tool head to the link in the description below all right so the reason why we've been doing this video is if i open my social media or my email and i knew this was coming this happens every time there's a new flagship something or another for graphics cards if i buy a xyz gpu and i pair it with my abc cpu will i bottleneck and i've seen this a million times we see it every time as a launch and it's fine so i'm going to try and do a video here today where we sort of kind of are trying to preempt the strike whether or not you will or won't bottleneck and the other reason why i really think this this video is important is because one thing that we definitely observed during our reacting to subscriber builds which we're going to do more of that in the future so please make sure you subscribe for that and follow me on twitter because that's where we will grab the information from you guys to use in those videos the amount of amd systems out there is super prevalent it i even commented by like the third video i'm like does anyone run intel anymore because i'm not seeing any intel systems i don't think we had like but one or two 10th gen systems to be honest but here's what we know we know of the charts are true which you guys should always take with a grain of salt um until you guys can see independent reviews and not the re not the independent benchmarks that are shared by the brand because of course they're they control those benchmarks and what goes live they get the final say we tested the 30 70s level of performance with this 2080 ti right here on the table we took this exact 2080 ti which we know is going to be basically a 37d but the 370 has three gigahertz gigabytes less of vram um they both use gddr6 not g6x and so it's very equivalent to what you can expect for the performance if the what i started to say if the charts are true and it's not just like in these two tests it was faster so that's what we're going to advertise then we know what 2080 ti and the 30-70 are essentially the same level or lying across the chart on performance so we took our brand new i've never actually had a dedicated amd test bed before i now have one i built it this is running an amd uh 3800 xt 8 core 16 thread cpu and i'll talk about why i chose that cpu in a second um with our msi x570 godlike motherboard which i admittedly like their bios for amd more than and a lot of the control more than what i was getting with the asus board which is kind of ironic because i really do like asus boards but whatever versus our intel test bed which we've had now for a few months running a 10 900k hold your comments i will address the core differences here in a second in terms of the core count which also has uh a well it was a asus maximus 12 motherboard they both have aios that are comparable to each other okay this is 140 and this is a 120. temperatures were not tested in this nothing was thermal throttling but you can also see down here in front of my laptop we were running the exact same memory modules so we took out as many variables as we could between the two systems between ram and graphics card we do have a video showing that two identical graphics cards will perform differently on the same system based on asic quality boost tables and other algorithmic things that happen and you know electricals go in and magicals come out they're not identical you could take two identical cards and put them on the same system get two different behaviors so we use the same card between both systems to eliminate any of that variance we also ran both systems at identical memory speeds which is xmp um fun fact xmp on the amd system instead of being 3200 which the modules show was actually 33.37 so we matched the memory on intel to what the memory xmp showed for amd even though it says 3200 on the dimms they went to 3337 when i enabled xmp so we matched it uh and then what else did we do here that's pretty much it oh overclocking no overclocking whatsoever on the amd no overclocking whatsoever on the intel both of them following their manufacturer specs so that means for asus we enabled all of the intel limits we disabled all of the asus intervention that tries to boost higher boost longer and all that sort of stuff enhancement whatever enhance enhance enhance so i removed any of the the chances of the motherboard skewing our results by allowing the cpu to go higher so it's all intel logic the other thing we made sure of that a lot of people don't do with the ryzen systems and again i know i'm sounding like linus here make sure you subscribe for the video we got coming with that also too both of these systems are on identical versions of windows 2004 they've got all their updates their windows were exactly the same in terms of what programs were running what tasks no things running in the background we checked for all of that parity to make sure that they were the same um last last but not least we did take the extra step that a lot of people don't do with the ryzen systems we're going to be doing a video about this on what you need to do with your ryzen system because you're leaving performance on the table so yeah i know i'm sound like linus make sure you're subscribed for blah blah blah video and we got lots of content the juices are finally flowing again we did make sure that we had the chipset installed enabling our windows based power plans for ryzen if you don't do that you won't get the amd ryzen high performance power plan which really speeds up the way that the task scheduler works if you don't do that believe it or not you're actually leaving performance on the table we'll talk about details on that in a later video the tests that we ran are very rtx heavy because what we're seeing now with comments finally is rtx on is finally viable and that you're not going to take stupid amounts of performance hit and paying 1200 to get 60 fps everyone's super excited about the 499.3070 giving you 2080 ti levels of performance which means rtx and dxr ray tracing and all those functions are finally reachable for many more people than there was before which makes you excited for whatever's potentially coming lower down the stack maybe bringing rtx for all who knows what they're gonna actually call it if they do that regardless so we used the game control which is the only title that we know of right now that truly uses all of the the nvidia design suite it's a beautiful game it's a great title it's not one of those indie titles that uses all of it it's a true title uh metro exodus which doesn't again doesn't use all of it but it does have a built-in benchmark so we can compare the differences between the two uh far cry 5 which is a title that we know uses a lot of cpu shadow of the tomb raider which again is a very demanding title in the way that it uses ray trace shadows and cs go for the i don't care give me the highest fps available type of title and of course we tested all these in 1080p and uh 1440 and 4k because if we're testing for bottlenecks we know the lower the resolution the higher the load on the cpu because of the higher frame rates so with all that out of the way here's the way it panned out so all right a few things to note here and please hold your comments till the end questions at the end put your hand down bailey billy don't make me call your mom all right so a couple things to mention here um but jay you you didn't even talk about the fact that there's a two thread four core advantage on the 10900k versus the 800 xt well here's what we did we did the best we could to try and make sure there were no bottlenecks in the 88 3800 xt the reason why we chose this particular cpu and not the 3900 xt or the 3900 is when we've tested the 3900 xt and the 3900x versus the 3800 xt this is an up to 4.7 turbo clock it's the only ryzen cpu that we have that is actually giving us that the 3900 xt we tested this extensively in another video we could not get the promised boost clocks or turbo clocks period they would not go anywhere near where promised the 4.7 boost clock or turbo clock on this processor 2 is amongst the highest of all of the ryzen family we also noticed that this is the only one that would actually hold those boost clocks for any limit of time some of the other cpus would hit their turbo clock and then immediately clock down like three four five hundred megahertz this one also allowed it to bounce back and forth between the best core and the second best core on the ccx's so this gave us our best possible scenario for ryzen cpus we also on this one turned off two cores which brings it down to an eight core 16 thread processor just like this guy ran our test again and it made no difference the numbers were scary consistent and identical to our 10 core 20 thread fully opened and unlocked cpu so we ran the test both ways the numbers were the same so what you could take away from this was we actually gave an advantage to amd and that we made the best possible scenario in terms of what you're seeing here so the numbers that you saw are well kind of surprising and not i thought that the intel the intel system being boosting up to 5.3 gigahertz and remember that's that's spikes that's not consistent at 5.3 those are those are spiky workloads and games and such that can boost up to that it sits around four eight to five oh most of the time on those particular use uh used cores but the intel system also has something in there that we don't uh we i didn't turn off and i don't think it matters in games and that is the turbo timer that is the amount of time that it's allowed to sit at those particular frequencies but i don't think it really came into play in any of these tests because it's not a sustained workload for 56 seconds that then causes it to wrap way down on the the turbo voltage or the turbo clock speed that way it can keep itself under thermal limits and power limits that didn't really come into play but we did test overclocking as well the overclocked numbers locking all the cores at 5.2 all 10 cores 20 threads didn't yield us any real results that were really worth mentioning to be honest we saw like in control we went from 105 1080p all the way up to 110 so and we chose control because that's actually one of the titles that had the biggest gap between intel and amd so 1080p across the board on this if you've got rtx on 1080p if you're running a high refresh rate monitor still is not going to max out your monitor i mean the only title that got us anywhere near the 144 which was actually 158 in far cry 5 and 128 and amd so we're talking 30 fps reduction on the amd system on that one title and the reason why we chose that title is because it is a very cpu intensive game as well even the benchmark all of the ai stuff happening all the explosions and all that is very cpu bound that's why they even give you a gpu cpu bound type of chart to look at at the end of it so the clock speed on intel the 500 plus megahertz that benefit that it's getting was very evident in that particular title but if you look at cs go that one's kind of weird too that's a we actually got higher fps and 1440p on both systems than we did at 1080p which is just weird we don't know if there's an engine bottleneck in fact we made notes on here that it was very anomalous we don't know why that is um but you're still getting 300 between 295 and 302 fps in 4k there's a bit of an engine limitation kind of happening there but we only chose that title because we needed some sort of esports title that would have stupid high fps to see if there was some bottlenecks happening and both of these systems no surprise were clearly capable of running it at 300 fps and not giving you any sort of of issues there cs go and control though those but those benchmarks they were recorded using uh msi afterburner which means that phil had to do a and we use phil actually did the the test for control because he's played the game he knows where he can go to have a good battle happen the problem is it's very ai driven ai is handling all of the enemies and stuff and it's not consistent they do not spawn in the same spot they do not behave in the same way they do not do the same thing it's not a track so it's important to remember that that particular title was a rolling average over about a two-minute fight that we then take multiple runs and then figure out what the fps really is it's not super consistent now that you've seen the charts let's talk about whether or not you're gonna be bottlenecking yourself as you can see on the amd system nearly across the board it is giving us less less fps not a lot a couple of fps in some instances up to 30 fps in far cry 5. does that mean that anything you buy for this system moving forward would be a 3080 or even a 30 90 if you're running one of these cpus or like i said this is the fastest one we have faster than the ones with more cores faster than the lower end cpus in fact that's what makes me a little concerned about um kyle's video on his 1 000 build running the 3600 it does not boost as high as this cpu and it doesn't sustain it nearly as long i'm concerned that that build that he did for a thousand dollars that's getting a lot of views is bottleneck city i really think you guys should wait before you run out and match his parts and stuff because i do not believe that to be a very optimized system and that's only because i watched that video after we did all the testing and i was like oh kyle oh buddy oh man okay whatever um it is what it is you know youtubers are going to youtube i'm really i don't want people to have to have bottlenecks but what it's going to happen no matter what if you throw a 3080 on this you're going to get more fps the problem is a lot of people misunderstand bottlenecks and think that well we're getting 128 fps that was the highest number we saw in here therefore 128 fps is as far as you're going to get in any game it doesn't work that way there's different engine optimizations there's different ways instructions are utilized with the cpu there's different ways that there's handoff between the the cpu and the gpu and the memory no two titles operate the same that's why you see very different numbers here but that's not to say the intel system isn't also giving us some sort of a potential bottleneck although the overclock by giving us no other no additional uplift in fps really tells us that we're not truly bottlenecking a 3070 level cpu so you can bet your ass we're going to do this video again with 3080s and 3090s when they become available to see what's truly happening is the gap gonna widen between these two platforms i don't know i think there becomes a point where if you truly are hitting a huge bottleneck on your cpu that we will definitely see one pull away from the other when it comes to clock speeds because clock speed is king when it comes to games but in terms of cpu utilization we took notes on what was happening during the test and none of these tests like pegged our cpu now again this is overall cpu usage and that doesn't mean that the other cores that are sitting there doing nothing means that you're not utilizing the cpu you could have the cores being utilized to the game be pegged at 100 um and that was actually the case in some of these tests uh far cry 5 being one of them that's one of those titles that uses maybe four or five cores but those cores are being heavily utilized and leveraged which is why the intel system uplifts so much higher than the amd system and that's my concern if you play titles with rtx off that's what's going to happen what kept the the results really artificially similar was rtx that was the factor there that the performance hit in rtx well jay why didn't you run these tests without rtx then because we believe that rtx now with the new 30 series and more people adopting it is a feature that people are going to care about and so we wanted to see what the performance is going to be with those features turned on you turn it off fps goes up through the roof quite a bit on non-dlss titles that's a whole other discussion what dlss does we didn't see utilization though maxed out on any of these cpus in terms of core count and that's obvious that no games are going to use 20 threads or 16 threads um but our gpu utilization is the is the thing we look at though if gpu utilization is low then that tells us bottlenecks are high and we were seeing 98 99 the only time we saw really 99 or not full utilization is in 1080p yeah we were only seeing 80 utilization in far cry 5 on the intel system so that tells us there was indeed a little bit of a cpu bottleneck happening on far cry 5 only 80 utilized in 1080 but we were seeing uh utilization up to 98 99 on all the other resolutions but basically none of this matters if you're the esports gamer playing those really fast-paced twitch style games you're going to probably go for the intel system because you know that it right now it is giving you the most fps but if you're also the casual gamer like i am or even like phil and you're just out for all the beauty the hdr the the rtx and and all of that then you're probably going to go with the amd system because as we've already shown you're not getting any sort of real tangible bottlenecks i'm almost positive if we if we did a side-by-side taste test with these systems no one could identify which was playing on amd and which was playing on intel um yeah especially at well 1440 and above even um you were probably going to be inclined to go with amd because of the pcie gen 4 that's going to give you ultra fast nvme storage it also means it's going to give you more cores per dollar and you wouldn't actually at the end of the day be able to tell a true difference and the type of gamer that you are really does play heavily into which system you're probably gonna choose anyway anyway guys thanks for watching again we're gonna we're gonna we're gonna stir the pie a little bit which platform are you going to use for your 30 series system are you going to say screw it i don't care about the max fps i just will refuse to buy intel or are you going to buy intel because you want every damn left fps on the table regardless of what it costs and or who makes it sound off in the comments below we do have a lot of videos coming about this subject matter like i said the the creative juices are flowing and i'm excited it's gonna be a fun fall and holiday season thanks for watching guys make sure you subscribe for those videos i mentioned and we'll see you the next oneit should be no surprise that with 30 series uh 30 series is an announcement and the excitement around the performance per dollar the amount of people that are like hey i'm going to get 30 series i'm gonna get 30 70. i'm gonna get 30 80 but will i bottleneck today we're gonna talk about that because um that that's a very important discussion to have especially right now having trouble figuring out the right parts for your next build then the custom pc builder at microcenter.com is just for you microcenter's free custom pc builder tool guides you through choosing the correct components from start to finish and ensures maximum compatibility for trouble-free installation smart search features allow you to customize your build based on anything you wish including price features and store availability and once you're ready to purchase the items in your cart will be reserved and ready for pickup at the micro center store of your choice not sure which parts are right for you we'll then head to micro center's community forums for help and advice from other pc enthusiasts like yourself to get started planning your next build and to check out the free custom pc builder tool head to the link in the description below all right so the reason why we've been doing this video is if i open my social media or my email and i knew this was coming this happens every time there's a new flagship something or another for graphics cards if i buy a xyz gpu and i pair it with my abc cpu will i bottleneck and i've seen this a million times we see it every time as a launch and it's fine so i'm going to try and do a video here today where we sort of kind of are trying to preempt the strike whether or not you will or won't bottleneck and the other reason why i really think this this video is important is because one thing that we definitely observed during our reacting to subscriber builds which we're going to do more of that in the future so please make sure you subscribe for that and follow me on twitter because that's where we will grab the information from you guys to use in those videos the amount of amd systems out there is super prevalent it i even commented by like the third video i'm like does anyone run intel anymore because i'm not seeing any intel systems i don't think we had like but one or two 10th gen systems to be honest but here's what we know we know of the charts are true which you guys should always take with a grain of salt um until you guys can see independent reviews and not the re not the independent benchmarks that are shared by the brand because of course they're they control those benchmarks and what goes live they get the final say we tested the 30 70s level of performance with this 2080 ti right here on the table we took this exact 2080 ti which we know is going to be basically a 37d but the 370 has three gigahertz gigabytes less of vram um they both use gddr6 not g6x and so it's very equivalent to what you can expect for the performance if the what i started to say if the charts are true and it's not just like in these two tests it was faster so that's what we're going to advertise then we know what 2080 ti and the 30-70 are essentially the same level or lying across the chart on performance so we took our brand new i've never actually had a dedicated amd test bed before i now have one i built it this is running an amd uh 3800 xt 8 core 16 thread cpu and i'll talk about why i chose that cpu in a second um with our msi x570 godlike motherboard which i admittedly like their bios for amd more than and a lot of the control more than what i was getting with the asus board which is kind of ironic because i really do like asus boards but whatever versus our intel test bed which we've had now for a few months running a 10 900k hold your comments i will address the core differences here in a second in terms of the core count which also has uh a well it was a asus maximus 12 motherboard they both have aios that are comparable to each other okay this is 140 and this is a 120. temperatures were not tested in this nothing was thermal throttling but you can also see down here in front of my laptop we were running the exact same memory modules so we took out as many variables as we could between the two systems between ram and graphics card we do have a video showing that two identical graphics cards will perform differently on the same system based on asic quality boost tables and other algorithmic things that happen and you know electricals go in and magicals come out they're not identical you could take two identical cards and put them on the same system get two different behaviors so we use the same card between both systems to eliminate any of that variance we also ran both systems at identical memory speeds which is xmp um fun fact xmp on the amd system instead of being 3200 which the modules show was actually 33.37 so we matched the memory on intel to what the memory xmp showed for amd even though it says 3200 on the dimms they went to 3337 when i enabled xmp so we matched it uh and then what else did we do here that's pretty much it oh overclocking no overclocking whatsoever on the amd no overclocking whatsoever on the intel both of them following their manufacturer specs so that means for asus we enabled all of the intel limits we disabled all of the asus intervention that tries to boost higher boost longer and all that sort of stuff enhancement whatever enhance enhance enhance so i removed any of the the chances of the motherboard skewing our results by allowing the cpu to go higher so it's all intel logic the other thing we made sure of that a lot of people don't do with the ryzen systems and again i know i'm sounding like linus here make sure you subscribe for the video we got coming with that also too both of these systems are on identical versions of windows 2004 they've got all their updates their windows were exactly the same in terms of what programs were running what tasks no things running in the background we checked for all of that parity to make sure that they were the same um last last but not least we did take the extra step that a lot of people don't do with the ryzen systems we're going to be doing a video about this on what you need to do with your ryzen system because you're leaving performance on the table so yeah i know i'm sound like linus make sure you're subscribed for blah blah blah video and we got lots of content the juices are finally flowing again we did make sure that we had the chipset installed enabling our windows based power plans for ryzen if you don't do that you won't get the amd ryzen high performance power plan which really speeds up the way that the task scheduler works if you don't do that believe it or not you're actually leaving performance on the table we'll talk about details on that in a later video the tests that we ran are very rtx heavy because what we're seeing now with comments finally is rtx on is finally viable and that you're not going to take stupid amounts of performance hit and paying 1200 to get 60 fps everyone's super excited about the 499.3070 giving you 2080 ti levels of performance which means rtx and dxr ray tracing and all those functions are finally reachable for many more people than there was before which makes you excited for whatever's potentially coming lower down the stack maybe bringing rtx for all who knows what they're gonna actually call it if they do that regardless so we used the game control which is the only title that we know of right now that truly uses all of the the nvidia design suite it's a beautiful game it's a great title it's not one of those indie titles that uses all of it it's a true title uh metro exodus which doesn't again doesn't use all of it but it does have a built-in benchmark so we can compare the differences between the two uh far cry 5 which is a title that we know uses a lot of cpu shadow of the tomb raider which again is a very demanding title in the way that it uses ray trace shadows and cs go for the i don't care give me the highest fps available type of title and of course we tested all these in 1080p and uh 1440 and 4k because if we're testing for bottlenecks we know the lower the resolution the higher the load on the cpu because of the higher frame rates so with all that out of the way here's the way it panned out so all right a few things to note here and please hold your comments till the end questions at the end put your hand down bailey billy don't make me call your mom all right so a couple things to mention here um but jay you you didn't even talk about the fact that there's a two thread four core advantage on the 10900k versus the 800 xt well here's what we did we did the best we could to try and make sure there were no bottlenecks in the 88 3800 xt the reason why we chose this particular cpu and not the 3900 xt or the 3900 is when we've tested the 3900 xt and the 3900x versus the 3800 xt this is an up to 4.7 turbo clock it's the only ryzen cpu that we have that is actually giving us that the 3900 xt we tested this extensively in another video we could not get the promised boost clocks or turbo clocks period they would not go anywhere near where promised the 4.7 boost clock or turbo clock on this processor 2 is amongst the highest of all of the ryzen family we also noticed that this is the only one that would actually hold those boost clocks for any limit of time some of the other cpus would hit their turbo clock and then immediately clock down like three four five hundred megahertz this one also allowed it to bounce back and forth between the best core and the second best core on the ccx's so this gave us our best possible scenario for ryzen cpus we also on this one turned off two cores which brings it down to an eight core 16 thread processor just like this guy ran our test again and it made no difference the numbers were scary consistent and identical to our 10 core 20 thread fully opened and unlocked cpu so we ran the test both ways the numbers were the same so what you could take away from this was we actually gave an advantage to amd and that we made the best possible scenario in terms of what you're seeing here so the numbers that you saw are well kind of surprising and not i thought that the intel the intel system being boosting up to 5.3 gigahertz and remember that's that's spikes that's not consistent at 5.3 those are those are spiky workloads and games and such that can boost up to that it sits around four eight to five oh most of the time on those particular use uh used cores but the intel system also has something in there that we don't uh we i didn't turn off and i don't think it matters in games and that is the turbo timer that is the amount of time that it's allowed to sit at those particular frequencies but i don't think it really came into play in any of these tests because it's not a sustained workload for 56 seconds that then causes it to wrap way down on the the turbo voltage or the turbo clock speed that way it can keep itself under thermal limits and power limits that didn't really come into play but we did test overclocking as well the overclocked numbers locking all the cores at 5.2 all 10 cores 20 threads didn't yield us any real results that were really worth mentioning to be honest we saw like in control we went from 105 1080p all the way up to 110 so and we chose control because that's actually one of the titles that had the biggest gap between intel and amd so 1080p across the board on this if you've got rtx on 1080p if you're running a high refresh rate monitor still is not going to max out your monitor i mean the only title that got us anywhere near the 144 which was actually 158 in far cry 5 and 128 and amd so we're talking 30 fps reduction on the amd system on that one title and the reason why we chose that title is because it is a very cpu intensive game as well even the benchmark all of the ai stuff happening all the explosions and all that is very cpu bound that's why they even give you a gpu cpu bound type of chart to look at at the end of it so the clock speed on intel the 500 plus megahertz that benefit that it's getting was very evident in that particular title but if you look at cs go that one's kind of weird too that's a we actually got higher fps and 1440p on both systems than we did at 1080p which is just weird we don't know if there's an engine bottleneck in fact we made notes on here that it was very anomalous we don't know why that is um but you're still getting 300 between 295 and 302 fps in 4k there's a bit of an engine limitation kind of happening there but we only chose that title because we needed some sort of esports title that would have stupid high fps to see if there was some bottlenecks happening and both of these systems no surprise were clearly capable of running it at 300 fps and not giving you any sort of of issues there cs go and control though those but those benchmarks they were recorded using uh msi afterburner which means that phil had to do a and we use phil actually did the the test for control because he's played the game he knows where he can go to have a good battle happen the problem is it's very ai driven ai is handling all of the enemies and stuff and it's not consistent they do not spawn in the same spot they do not behave in the same way they do not do the same thing it's not a track so it's important to remember that that particular title was a rolling average over about a two-minute fight that we then take multiple runs and then figure out what the fps really is it's not super consistent now that you've seen the charts let's talk about whether or not you're gonna be bottlenecking yourself as you can see on the amd system nearly across the board it is giving us less less fps not a lot a couple of fps in some instances up to 30 fps in far cry 5. does that mean that anything you buy for this system moving forward would be a 3080 or even a 30 90 if you're running one of these cpus or like i said this is the fastest one we have faster than the ones with more cores faster than the lower end cpus in fact that's what makes me a little concerned about um kyle's video on his 1 000 build running the 3600 it does not boost as high as this cpu and it doesn't sustain it nearly as long i'm concerned that that build that he did for a thousand dollars that's getting a lot of views is bottleneck city i really think you guys should wait before you run out and match his parts and stuff because i do not believe that to be a very optimized system and that's only because i watched that video after we did all the testing and i was like oh kyle oh buddy oh man okay whatever um it is what it is you know youtubers are going to youtube i'm really i don't want people to have to have bottlenecks but what it's going to happen no matter what if you throw a 3080 on this you're going to get more fps the problem is a lot of people misunderstand bottlenecks and think that well we're getting 128 fps that was the highest number we saw in here therefore 128 fps is as far as you're going to get in any game it doesn't work that way there's different engine optimizations there's different ways instructions are utilized with the cpu there's different ways that there's handoff between the the cpu and the gpu and the memory no two titles operate the same that's why you see very different numbers here but that's not to say the intel system isn't also giving us some sort of a potential bottleneck although the overclock by giving us no other no additional uplift in fps really tells us that we're not truly bottlenecking a 3070 level cpu so you can bet your ass we're going to do this video again with 3080s and 3090s when they become available to see what's truly happening is the gap gonna widen between these two platforms i don't know i think there becomes a point where if you truly are hitting a huge bottleneck on your cpu that we will definitely see one pull away from the other when it comes to clock speeds because clock speed is king when it comes to games but in terms of cpu utilization we took notes on what was happening during the test and none of these tests like pegged our cpu now again this is overall cpu usage and that doesn't mean that the other cores that are sitting there doing nothing means that you're not utilizing the cpu you could have the cores being utilized to the game be pegged at 100 um and that was actually the case in some of these tests uh far cry 5 being one of them that's one of those titles that uses maybe four or five cores but those cores are being heavily utilized and leveraged which is why the intel system uplifts so much higher than the amd system and that's my concern if you play titles with rtx off that's what's going to happen what kept the the results really artificially similar was rtx that was the factor there that the performance hit in rtx well jay why didn't you run these tests without rtx then because we believe that rtx now with the new 30 series and more people adopting it is a feature that people are going to care about and so we wanted to see what the performance is going to be with those features turned on you turn it off fps goes up through the roof quite a bit on non-dlss titles that's a whole other discussion what dlss does we didn't see utilization though maxed out on any of these cpus in terms of core count and that's obvious that no games are going to use 20 threads or 16 threads um but our gpu utilization is the is the thing we look at though if gpu utilization is low then that tells us bottlenecks are high and we were seeing 98 99 the only time we saw really 99 or not full utilization is in 1080p yeah we were only seeing 80 utilization in far cry 5 on the intel system so that tells us there was indeed a little bit of a cpu bottleneck happening on far cry 5 only 80 utilized in 1080 but we were seeing uh utilization up to 98 99 on all the other resolutions but basically none of this matters if you're the esports gamer playing those really fast-paced twitch style games you're going to probably go for the intel system because you know that it right now it is giving you the most fps but if you're also the casual gamer like i am or even like phil and you're just out for all the beauty the hdr the the rtx and and all of that then you're probably going to go with the amd system because as we've already shown you're not getting any sort of real tangible bottlenecks i'm almost positive if we if we did a side-by-side taste test with these systems no one could identify which was playing on amd and which was playing on intel um yeah especially at well 1440 and above even um you were probably going to be inclined to go with amd because of the pcie gen 4 that's going to give you ultra fast nvme storage it also means it's going to give you more cores per dollar and you wouldn't actually at the end of the day be able to tell a true difference and the type of gamer that you are really does play heavily into which system you're probably gonna choose anyway anyway guys thanks for watching again we're gonna we're gonna we're gonna stir the pie a little bit which platform are you going to use for your 30 series system are you going to say screw it i don't care about the max fps i just will refuse to buy intel or are you going to buy intel because you want every damn left fps on the table regardless of what it costs and or who makes it sound off in the comments below we do have a lot of videos coming about this subject matter like i said the the creative juices are flowing and i'm excited it's gonna be a fun fall and holiday season thanks for watching guys make sure you subscribe for those videos i mentioned and we'll see you the next one\n"