Gadgetory

welcome back to our inbox now last week we got our first look at the battlefield 5 open beta or beta as my American viewers like to say yes the British pronunciation is usually better well the American pronunciation is beta and as far as I know all Aussie say beta but Steve I'm an Aussie and I say beta don't lie to me anyway what's this vo about again is the battlefield 5 open beta I mean beta cherish ok so last week I benchmarked almost a dozen AMD and NVIDIA GPUs at 1080p 1440p and 4k to see how they compared in battlefield 5 and at the time I'd completely lost my voice or I just recovered from completely I was on my way to recovering from completely losing my voice as you can probably tell I still haven't fully recovered I've been to the doctors it's possibly going to be about a week with this sort of croak ish kind of voice so sorry about that but I'm not gonna stop making videos because we have plenty of cool content to get through so I'm just gonna have to suck it up and you're gonna have to bear with me anyway all testing in the previous video where we compared the GPUs was done our standard GPU test rig which comprises of a core i7 87 okay clocked at 5 gigahertz and I noticed really really high CPU utilization and I concluded at the time that if we're seeing in 87 ok hitting up to around 80% and usually hovering around 70% but running a quad core was likely going to go quite badly you'd see a lot of stuttering and just be a bad experience of course I would have to investigate that further at the time I foolishly promised a very in-depth CPU comparison once the games released in a few months time but hint at the fact that I might be able to do some CPU testing on the beta version of the game and yeah that prompted a huge amount of you to say please do CPU testing as soon as possible so with so many of you demanding a CPU comparison right away as soon as that GPU video went online I got testing I did note in the previous video that gathering accurate data with the single-player portion of the game is extremely difficult and it really is so much so that when it came to the CPU testing I had to take an average of 6 runs to try and smooth things out and get somewhat accurate information the runs were also extended so they're longer runs and as a result it took me three full days of nothing but testing to get just 10 CPUs done at two resolutions nope sorry one resolution using two GPUs so not the amount of data you would usually expect to see from us but the test conditions are very difficult and it's hard to get a lot of data because of it I'm also testing a different section of the map that's closer to where the action takes place does that seem to place a bit more load on the CPU however most of the time was wasted trying to get into the right game and then be out a benchmark at the right place so yeah a lot of time was wasted just trying to get to the benchmark location and then execute the pass honestly while I was really intrigued to see the data I have to admit I didn't enjoy the process at all a rare benchmark session that I mostly didn't enjoy anyway the data is in and it's certainly interesting so far I've only seen one other source provides CPU benchmark results and they came from game GPU and my numbers don't exactly aligned with theirs there is some crossover but for the most part the performance trends are a bit different and again this can be juda multiplayer being very difficult to benchmark accurately and the performance can vary massively depending on what map you are benchmarking where in the map you're benchmarking what the play is and the game are doing there's a lot of variables at play here so again just testing multiplayer is extremely difficult I don't mean that it's just hard work and I prefer to take the easy way out I mean it's nearly impossible to gather accurate data at least for comparison with other components the fluctuation between runs can be bloody massive whereas our battlefield one single player benchmark often delivers the same average framerate over and over again and the same is also true for the one percent low result you don't see that when testing with multiplayer when testing the single-player portion of battlefield 1 it's rare that we see a deviation of more than five frames per second with the battlefield 5 multiplayer though I was often seeing up to a 20 FPS difference for the average framerate I did neglect the extreme outliers and I tested many more times than I normally would to try and report the most accurate performance I could because of this just watching it side-by-side gameplay footage of two different hardware configurations executing a similar pass in my opinion is next to useless if you've got 30 players relatively close to you the frame rates seem to be much lower especially if they're blowing things up and engaging in extreme combat however if the actions taking place much further away from you it seems to be a lot less demanding on the system also making sure the same sort of stuff is going on around the player character for each pass is a serious challenge and it's basically an impossible challenge which is what I said earlier sorry by now you've probably gathered I'm not a big fan of these multiplayer benchmarks at least not for comparing different hardware configurations that said this is the absolute best we can do for now and while I believe the data is reasonably accurate given the conditions and the effort I've put into making them as accurate as possible I'd still only recommend you take them as a rough guide so having said all that let's have a look at what I have ok so these results were all recorded at 1080p using the ultra quality settings with the GTX 10 atti as you can see the core i7 87 okay provides the best results allowing the 1080i to pump out 158 FPS on average with a 1% lower result of 118 fps the core i5 8600 k was equally impressive and Intel's last quad core Core i7 the 77 RK also crushed it then from the 77 ok we see an 18 percent drop in performance to the 2700 acts which honestly is significantly more than I was expecting to see then we say that these second gen Rison processors were only able to edge out the core i5 7600 K as the Intel seventh gen quad core provided first generation rise in performance also please note I'm not using tuned sub timings for the Rison CPUs the second-gen parts are using ddr4 3000 with XMP loaded the same is true for all the Intel CPUs as well so some pretty disappointing results here for AMD given Rison is quite competitive in battlefield 1 we often find that poor rise and performance can be a result of the Nvidia display driver sorry ran all the tests using vega 64 liquid for the most part we are now GPU limited so it's difficult to determine just how much in a video are to blame for aizen's weaker than expected performance with the GTX 280 ti the Rison processors did offer slightly better performance with vega 64 but with frame rates limited to around 120 FPS on average as I said we are GPU bound and this makes it difficult to determine if horizon is better off with the radeon GPU if the game had been a bit easier to test and didn't involve hours of wasted time trying to get into games i would have started over and tested again using the high or possibly even the medium quality settings at least for stuff like shadows for example what I did do before wrapping things up was retest with the GTX 1080i on the Rotterdam map and I did this before realizing that it would have been useful to perhaps test for Vegas 64 with some less demanding graphical settings anyway this benchmark pass is a complete lap of the train carriages which takes roughly 80 seconds to complete this benchmark was slightly more demanding than that of the narvik test but as you can see the margins are much the same Intel's 8 thread CPUs deliver class-leading performance while Verizon trailed by a much greater than expected margin it was disappointing a/c Rison only matching the quad core Core i5 7600 K in our battlefield 5 CPU tests again I realized that there is at least one other set of results out there that contradicts these findings but I spent a lot of time gathering this data and saw no evidence that the horizon processors could hang with the likes of the 77 ok 8600 K and 8700 K also to make sure nothing had gone wrong with the horizon test system I did jump back into battlefield 1 and confirmed the solid performance in that title that we had seen previously I was shocked to find very little stuttering with the quad core processors the quad cores that don't support Smt that is the 7,600 cane 8350 kay did it time suffer a small amount of stuttering but it was very minimal and certainly didn't take away from the gameplay oddly both the risin 5 and risin 7 processors also suffered a small amount of stuttering about the same as the 4 threaded intel parts the higher-end intel parts did provide perfectly smooth performance the semi similar okay for example was buttery smooth I should note that even with Vega 64 there was still a small amount of stuttering with Rison so battlefield 5 clearly needs to be better optimized for Rison and that's quite surprising given how well the battlefield 1 was optimized for the rise in architecture and both games do use the same frostbite 3 engine that said though they did still manage to go backwards on the whole dx12 thing so why not do the same with processor support anyway the game is still only in the beta phase must be fully assimilated now so hopefully the developer can polish it a little bit more before their release in November personally I'm not expecting any real big improvement but I am hopeful that before too long they will address these performance issues finally I would like to just thank patreon member element 83 for generously lending his origin account to help with the testing for those of you unaware origin does this really awesome thing where it locks you out after five Hardware changes and the lockout last 24 hours this makes benchmarking a number of hardware configurations extremely difficult I am in talks with EA right now to see if there's anything we can do about my account for testing in the future for now the workaround is to use multiple accounts and I had four to work with for this video so again a big thank you to element 83 for helping out there can't be easy sacrificing your account when you're very keen to place in battlefield 5 oh and I should note that he's also using the horizon 720 700 X with 32 gigabytes of ddr4 3000 memory with Titan sub timings and he's also reporting a bit of random stuttering here and there with the horizon processor again only a very small amount of stuttering like what I saw but it's good to know that I wasn't the only one seeing this and that is going to do it for this again sorry about my voice and I realize there is only a very limited amount of testing in this video compared to what we normally do but remember it is just the beta version of the game and when it's officially released hopefully we better provide a lot more GPU and CPU benchmarks also coming up on the channel soon I will a couple of you notice that this build has been progressing along nicely I'll have a video later this week showing you where I'm at with this it's coming along very nicely I also have an episode of unboxing boxes coming up for all of you that have been asking you in the next episode of that is and I have a mountain of gear to go through a lot of it we will be reviewing and testing on the channel over the coming weeks so it'll be good to get awesome and early look at those products anyway if you enjoyed this video please feel free hit the like button subscribe for more content just like this and if you appreciate the web do hara box then consider supporting us on patreon thanks for watching I'm your host Steve and I will see you again next time