Gadgetory

welcome back to harbor unboxed today's video has loads of benchmarks for you i'm confident a lot of the testing is really going to be of interest let's hope so anyway also be aware this is going to be a bit of a long-winded one so if you're after a short sharp and to a point video i can assure you that this isn't that you've been warned ok so recently i've conducted a heap of gaming benchmarks and work out which cpus from AMD and intel provide gamers with the best experience and offer the most value typically we found that intel mainstream cpu still deliver the best gaming performance thanks to a combination of things really are primarily the low latency ring bus architecture and their ability to clock up to around 5 gigahertz despite that though we often hear from viewers when making our head-to-head cpu comparisons that rise ins a better choice because it's more future proof this came up over and over again when comparing the core i5 8400 and risin 5 2600 the Rison cpu supports SMT technology and therefore has twice as many threads we heard it again when comparing the horizon 720 700 x and core i7 87 ok the 2700 x of course has two extra cause now just to be crystal clear we don't doubt for a second that those extra CPU resources won't come into play in the future and give aimed at a serious advantage because they almost certainly will the problem we have is telling you to buy the 2700 X because we think it will be faster for gaming in the future it isn't that we don't think that's true we just don't know exactly when that future will be speaking of future gaming performance people still seem to get all worked up about 720p benchmarking screaming from the rooftops that it's not an indicator of future performance and that's actually true to a certain extent although resolution CPU limited benchmarking doesn't tell us how the 2,700 X 87 hurricane will compare in 10 years time that said we've never claimed that it will I don't think anyone has I believe it is a useful indicator for how these CPUs compare in non GPU limited scenarios for not just the games we have available today the games that we're currently testing but also the majority of the games that will be released over the next year possibly even the next two years and let's be honest those really are the titles that you should care about when buying a CPU today there's also more to low resolution testing than just trying to predict near future performance not everyone plays using ultra quality type settings for example using a gtx 1070 with competitive settings will have a higher ceiling than a gtx 1080i using ultra quality settings in most titles above all else though CPU limited results are obviously very useful when analyzing CPU performance and while it's also very useful to see where the high end GPU limitations come into play those aren't the only results you want to pay attention to anyway this video isn't another 720p or low resolution benchmarking discussion well and truly over those for now and frankly have become suspicious of anyone requesting certain data sets be removed anyway what I want to do with this video or try and do with this video is work out how long it will be before the 2700 X is a more capable gamer than the 87 okay and we'll do this using a few different methods like I said I really don't expect games to evolve greatly in the next year or even the next two years for that matter and we might get some core heavy games perhaps battlefield 5 will be one of them but for the most part were you likely won't see any major change until probably a year or so after the next-generation consoles arrive since I recently compared the horizon 727 her X to the core i7 87 okay in 35 games and many suggested that the 2700 X was the better buyer for games because of those two extra cause I wondered how true that would eventually become once games eventually max out eight core processors and that's the thing it's not good enough for a game to just utilize eight course it has to max them out for the 2700 to have any chance of beating the 87 or okay at least by a convincing margin the 2700 X might have 33% more cores but with both CPUs overclocked the 8700 K clocks around 19% higher couple that with the lower latency method of connecting the course and ends up being a pretty close battle with both CPUs fully utilized truth be told there will only be a narrow window in time before the 2,700 X expires as a usable gaming CPU where it will have a big advantage over the 87 kay any way to explore all this we've tested a number of different configurations firstly I decided to turn the 2700 X into a 2600 X by disabling to cause one from each cc X so essentially I could have just used a 2600 X but don't worry about that too much it doesn't really make any difference so what we have are 6 second generation risin cause with SMT enabled giving us 12 threads then for comparison we have the core i7 8700 okay also with two cores disabled essentially turning it into a 77 or okay you might think well that's a bit silly why not just use a 77 oh okay and well I could have but I chose to use the 87 okay because it has the full 12 megabyte level three cache and therefore is better for comparing how the 2700 X and 8700 K might compare in the future a typical overclock for the unlocked coffee-like CPUs is about 5 gigahertz while a typical all core 2nd gen Rison overclock is 4.2 gigahertz so that's why we've chosen to test at those frequencies unless you're conducting an IPC test it makes no sense at all to under clock the Intel CPU to match the 4.2 gigahertz overclock of the AMD CPU we've done all those kind of IPC tests but for this one that's not what we're trying to do because we're looking at how these CPUs will shape up in the future and anyone running an 87 ook won't be doing it at 4.2 gigahertz on the other side of that coin AMD typically has more cores which is why I'm making this video and while we're giving the AMD CPUs a 2 core advantage over the Intel parts finally both CPUs were paired with ddr4 3400 memory with tightened timings now for this one we've only tested three games but before you say that's not very hard unboxed of me just know I've got almost 40 graphs to go over and like I said a lot of different configurations were tested the game selection includes battlefield 1 ashes of the singularity and Deus Ex mankind of Ireland all of which had been benchmarked using the DirectX 11 and DirectX 12 modes using both a gtx 1080i and vegas 64 graphics card then wrapping everything up we have some additional core scaling performance to look at and that is particularly interesting but get to that after we do all the first batch of results there really is a heap of data to get through would blast through it because it's not that interesting and others will have a good chat about so let's jump into it I'm going to kick things off with ashes of singularity because I know you guys love this title starting with the DirectX 11 testing using the extreme preset with the GTX 1080i this is what we found and I should note that under these conditions the Intel CPU saw an average utilization of 88% with a peak of 100 percent while the AMD CPU averaged 51 percent with a peak of just 66% ok so here we see that the Intel CPU was faster at all three resolutions under these test conditions we're not able to take full advantage of the horizon CPU even with six cores enabled so that is a bit of a problem swapping to Vegas 64 changed a few things firstly the performance between the CPUs is much more competitive and aim Lieven hits the lead in a few tests however the game appears almost broken now at times we see frame time performance taking a serious nosedive Vegas 64 really isn't enjoying ashes of the singularity using the older dx11 API so we'll switch to DirectX 12 and check those results out in a moment before we do jump to the DirectX 12 results here's how the game performs using the standard quality preset here the average framerate is similar but Intel has a notable advantage for the frame time results the Vegas 64 results are again much the same similar average framerate performance as seen with surprisingly poor frame time results so let's move on to the DirectX 12 results to see what they have for us moving to the DirectX 12 API we do see slightly better CPU utilization for the AMD CPU and as a result we do see more competitive performance now the rise in CPU was able to nudge your head at 720p and 1080p using Vegas 64 doesn't change much here the margins are all quite similar so let's move on to check out these standard quality settings with the standard quality settings enabled on the DirectX 12 API we see slightly ever so slightly better performance using the Intel CPU with the GTX 1080i swapping to Vegas 64 provides mixed results for the most part performance was much the same but we do see the AMD and Intel CPUs trading blows under these test conditions okay so that wraps up the ashes of the singularity testing let's move on to Battlefield 1 up first we have the DirectX 11 results using the ultra quality settings with the GTX 10 ATT I hear the four core coffee like CPU was able to beat the six core second gen rise in part though it has to be said the margins were minimal switching to Vega 64 mix things up a bit and now the second generation CPU can be seen delivering stronger performance at 1080p and 1440p the margins aren't huge but Rison is clearly faster here here we see that reducing the GPU bottleneck with the medium quality settings does play into Intel's hands when using the gtx 980ti here the four core coffee like CP was up to 14% faster using vega 64 we see the Intel CPU still takes the lead at 720p and 1080p but was slower at 1440p falling behind the rise in CPU by a 5% margin now we're switching to the DirectX 12 API and once again using the gtx 1080i we see intel has a clear performance advantage when not GPU limited this time we see when using vegas 64 the frame rates are much more comparable but even so having two extra cause doesn't give rise and a noteworthy advantage reducing the GPU bottleneck further with the medium quality setting sees the quad-core Intel CPU delivering up to 18% more performance with the GTX 1080i then with Vegas 64 this reduces that margin quite significantly but even so the Intel CPU is still up to 6 percent faster ok so now we have the Deus Ex mankind divided results and when using the dx11 api the four core Intel system saw a peak CP utilization of 64% with an average utilization of 57% we saw slightly higher numbers for the 6 core AMD system hitting 63% on average with a maximum utilization of 70% using the GTX 280 eye the Intel CPU was up to 15% faster though as we increase the GPU bottleneck with the higher resolutions that margin does shrink with Vegas 64 the GPU bottleneck is introduced much earlier and this means that a DP the CPU performance is pretty well neutralized the lower 720p resolution does help to avoid this GPU limitation and here the Intel CPU was 11% faster dropping down to the medium quality preset with the gtx 1080i reduces the GPU bottleneck and now the Intel quad-core configuration is up 21% faster Vega 64 also allows the CPUs to stretch their legs a bit more with the medium quality settings here the Intel CPU was up to 12 percent faster switching to the DirectX 12 API with the GTX 1080i doesn't help out the rise and CPU and here we see some very odd behavior although we're not CPU bound the results certainly appear as there we are as the Intel configuration is limited to around 87 to 89 fps and the AMD configuration eighty to eighty one FPS we see the same odd behavior with Vegas 64 when using the DirectX 12 API unfortunately not all or rather not many DirectX 12 titles are as well put together as ashes of the singularity even with the medium quality preset we see the same odd behavior with the GTX 1082 I and the same can be said for Vegas 64 as well not much else to say here really so let's move on from the Deus Ex mankind divided testing okay so whopping 2 calls off the 80 seminar okay and the 2700 give the rise and CPA noteworthy advantage in today's games Davis X mankind divided is a bit of a poor example it has to be said of a tunnel that utilizes many cause but it's also fairly typical example of how most games behave battlefield 1 is extremely cool heavy and you'll see further performance improvements when enabling are those extra cores on both CPUs so that is to say enabling all the cores and the 27 hundreds will make it a little bit faster than say the 2600 X even so in those two tiles we saw a very few examples where the six core second gen rising CP was faster than the four core coffee-like CPU in fact when using the GeForce GTX t2i we saw almost no examples with the exception of a single benchmark and that was ashes of the singularity are using the extreme quality settings with the DirectX 12 API the rise in CPU had a little more success with Vegas 64 but even so there wasn't any tests that really favored the AMD CPU by a notable margin the four core Intel CPU did often hit 100% utilization when testing with battlefield one and ashes at the singularity and we saw the average utilization hovering around 90% which is obviously quite high the problem though is that the same tests these same test conditions saw the rise in CPU are very underutilized hitting around 40% on average and I'll fill one about 50 to 60% in ashes of the singularity for those extra cause of the rise and see for you to really be of an advantage we need to see the AMD CPU hitting around 80 to 100 percent utilization and this would mean that the Intel CPU would be overwhelmed at a constant 100% so for such a test we've taken battlefield 1 again and we've tested scaling performance using a range of Core configurations the coffee-like CPU has been benchmarked with two four and six cores enabled while the second gen rising CPU has been tested with four six and eight cores enabled and remember all configurations still have SMT enabled again I've tested using both the DirectX 11 and DirectX 12 api's but this time I will be dropping Vegas 64 and just testing with the GTX 1080i as that shows us ultimately what we are looking for right so these results probably tell us all we need to know but even so I'm gonna go through all the results we have for the CPU scaling mostly because we have them and I'm sure you guys are still interested in taking a look I say these results tell us all we need to know because you can clearly see the tipping point where AMD's core advantage comes into play the drop down from 6 to 4 course for the coffee like architecture in this title see is just a 6 percent reduction in framerate and we're basically seeing the same reduction for these second gen rise in architecture and dropping down from 8 to 6 cause then when dropping down from 6 to 4 course for the second gen rise in architecture we see a rather large 15 percent drop-off in performance however what you want to pay attention to here is the massive 29 percent performance degradation seen for the coffee-like architecture when dropping from 4 cores down to just 2 cores so do remember SMT is enabled for both architectures basically what this means is right now for AMD's core count advantage to be realized we have to be playing with two core for thread Intel CPUs but before getting into that too much let's just check out the rest of the results really quickly using the medium quality preset to help reduce the GPU bond that we see similar performance margins for the 8 and 6 core battles as well as these 6 and 4 core battles but the margin for the 2 versus 4 core battle grows from 15% in favor of horizon 2 almost 30% again we see similar performance margins at 1080p here the eight core rise in CPU isn't that far behind the six core coffee-like CPU while it crushes it when making the four core verses - cool comparison we see pretty much the same thing when using the medium quality settings though the turnaround here for the lower core count comparison looks more dramatic then at 1440p we see that for the most part the rise in CPU is able to maximize the gtx 1080 eyes performance and we only see a slight slip in performance for the four core configuration then using the medium quality settings shows basically the same performance trends switching to the DirectX 12 API shows just how badly in videos drivers hurt their eyes and CPUs as we see very little shift in performance from the four core to the eight core configuration that said as was the case with DirectX 11 testing the Intel coffee-like architecture does fall behind once overwhelmed with just two cores similar performance trends are seen with the medium quality settings at 720p videos drivers do become less of an issue at this slightly more GPU limited Tony P resolution but in any case we see a similar performance trend the same can also be said when looking at the 1080p medium quality results then finally at 1440p we are quite heavily GPU limited and here while slower the two core coffee-like CPU configuration wasn't that much slower even when looking at the frame time performance of course if we reduce the quality settings the margin does open back up and again we see the four core eyes and CP offers a clear performance advantage over the two core Intel CPU something we don't see when comparing the four core Intel CPU to the six core verison CPU so quite a bit to go over here but I'll do my best to explain the results the most revealing of all the tests were the battlefield 1 cost scaling results here we could quite clearly see that having two extra cause isn't particularly useful today for AMD we're looking at the four and six core Intel CPUs so that would be the AMD CPUs with six and eight cores they're just not getting utilized and we just not putting them to good use so unfortunately for AMD that kind of puts Intel in a good position because most of their range does feature four to six cores these days they don't necessarily have SMT enabled that is something that the core i3 and Core i5 models do miss out on anyway what we've quite clearly found is that in order for AMD's extra course to come into play the Intel CPUs have to be completely maxed out it's hardly surprising that one but it is worth being reminded of as we discussed the results essentially what this means is for the horizon 720 700 X to beat the core i7 87 or okay we need games where the 2700 x ccp utilization hovering around 80 to 90 percent and that is opposed to the 20 to 40 percent we are currently seeing right now so how long till the core i7 87 or okay is overwhelmed and the 2700 X offers a noteworthy performance advantage it's really hard to say but it's likely not going to happen in the next few years in fact it's likely going to be a few years before we even see the six core 12 third rising configuration hands down beating a fork or eight thread coffee-like CPU if such a CPU existent let's say the core i7 7700 K and remember we are accounting for intel's current 19% clock speed advantage when overclocked so when some people say get the rise in cpu because it has more cores and therefore will be better for gaming in the future i really have to question just how far into the future they expect this changing of the gaurd to occur if right now in 2018 using some of the most CPU demanding games we have you need to compare a fork or eight threaded AMD CPU to a two core for thread Intel CPU in order for it to win how long is it going to be before we see this with double the core count honestly I hate to think of course when it comes to non-related gaming tasks using applications that can take advantage of the extra cores and threads that rise and offers there is a real advantage to be had and this is why we often recommend the rise in CPUs we're also well aware that for the most part gaming performance is GPU limited and second generalize it is close enough to coffee like in terms of frame rate muhammad's for any difference to go unnoticed and this is largely because under most conditions there is no actual difference that being the case can you fall back on the argument the rise on has more cores and therefore will be better for gaming in the future well given that in today's games there are indeed very similar and one day what will hopefully be a not-too-distant future those extra cause could mean the difference between smooth playable performance and a stuttering mess so that being the case you probably can however I wouldn't hold your breath it's likely you'll upgrade your platform once or twice before those extra calls come in handy for gaming and that is gonna do it for this one I hope you enjoyed this little benchmark exploration and if you did feel free to eat the like button for us subscribe for more content you if we appreciate the work with your horror books then consider supporting us on patreon thanks for watching Armijo Steve and I'll see you next time