Gadgetory

welcome back to harbor unboxed today we're going to take a look at some benchmarks as well as discuss a few things recently there's been a heap of controversy surrounding rising benchmark results and opinions given by reviewers as well as how the reviewers are actually testing by some miracle I managed to dodge a lot of the rising controversy particularly with my release day video though making this video might change that but I really hope it doesn't because everyone is entitled to their opinion the road to my rise in review was also extremely bumpy and I think I did a pretty good job of showing the current strengths and weaknesses while overall remaining objective after all as I noted in my initial Rison review this is a new platform and a totally new CPU architecture so they're bound to be teething problems that said the entire reason I review products at all is to give you the viewer consumer advice so that advice to carry any merit it has to be backed up by solid testing preferably testing done by myself an impartial party that you guys can trust since the initial rise and reviews went live it almost seems like the Internet has been on fire at least the nerd portion of it anyway so the last week my youtube comment spam filter has been flooded with links to an adored TV video a video I actually watched just moments after it went live so I was already aware of it I'm an adored TV fan I often stop by and let Jim know how much I enjoy his videos and support his work the video a few users were spamming on every tech channel that I was aware of was titled risin the tech press loses the plot starting our video one of the first things Jim talked about was the amazing productivity performance and how the results almost brought a tear to his eye well I have to admit to having similar feelings when I fired that premiere for the first time I was shaking with excitement and almost couldn't believe the result I felt like jumping for joy we finally had a real contender and reviewers like myself could start recommending something else in the same old core i3 Core i5 and core i7 processors however the point of that video was to discuss how the tech media benchmark CPU gaming performance now before I get to that I would just like to make a few things clear things I have been saying since day one I do believed in the future the horizon 7 CPUs will likely beat the sky and cable a core i7 quad-core processors the IPC performance is certainly there and they obviously have more resources to work with so once games become more demanding and take advantage of those resources things should start to head in AMD's favor as I said in my GPU bottlenecking video I just don't know how far into the future that will be and so somebody building a system today that is a consideration if horizon 7 isn't able to hit full stride and 2/3 or maybe even four years in games then for gaming you are better off buying a core i7 processor now and upgrading down the track of course if you do any kind of productivity or content creation with your PC then again as I said in my first rise in review you are much better off going with AMD at this point where I disagree with Jim is how CPU should be tested I strongly believe removing a GPU bottleneck to test actual CPU gaming performance is how it should be done just as it always has been now shortly after that video went live droves of YouTube viewers came to my channel and other tech channels spouting about how the AMD FX series is now faster than Sandy Bridge something like 10 percent faster on average when looking at a large number of modern games that to me just didn't sound right I often benchmarked both these processes when testing new games and it's extremely rare that the FX processes are even able to match the core i5 2500 K and I've never seen them match to Core i7 2600k the claim here was that the FX 83 70 is now faster than the 2500 K and again this had me scratching my head as it wasn't that long ago that I compared the two and it was a one-sided beatdown rewind to when the bulldozer architecture was first announced we noted at the time that gaming performance was down due to a lack of utilization as games of the time back in 2011 we're only using about to cause the FX 83-70 is of course based upon 2011 revised piledriver architecture but the same theory applied so we always expected performance would improve that's that we knew would never get anywhere near the level of the core i7 processes of the time and unless I've missed something it never did the 2500 K on the other hand is just a quad core part and it does lack hyper-threading support so it was possible that the eight core FX processors could eventually nudge ahead last year we started to see this dialect core i7 CPU is pulling ahead of the core i5 models wouldn't pair with the gtx eternity graphics card or something fast like the titan XP in games such as battlefield one civilization 6 deus ex mankind divided f1 2016 Gears of War 4 hitman mafia 3 watchdogs 2 overwatch and the list goes on so we are clearly starting to see games require more than four cores and as a result the quad core Core i7 models are already starting to benefit from their hyper threading support so does this mean after all these years those eight core FX processors are starting to make a comeback we figured they might at some point but after years of getting messed up by modern Core i3 processors I honestly thought the day may never come since I just finished testing the rise in CPUs along with the FX 83 70 and 16 games many of which are modern I thought why not go back and add the core i5 2500 K to the results so this is exactly what I've done and the results are interesting though for the most part not all that surprising for testing I again went with the Pascal Titan X and an effort to remove any GPU bottlenecks so did chest at 1440p as well as 1080p the FX 83-70 was tested with ddr3 2400 memory while the 2500 K was paired with ddr3 2133 memory all the clock speeds are stocks or the 2500 K would boost as high as 3.7 gigahertz and the FX 83-70 as high as 4 point 3 gigahertz ok I think that covers all the important details let's get to it kick starting the benchmarks we have the division and even at 1080p we are primarily GPU bound here the 2500 K did deliver a better result and while an 8% greater minimum frame rate is nothing to scoff at the FX 83-70 did deliver reasonably smooth performance in this title I've leaves out at 1440p the FX 83-70 falls further behind which is unexpected the 2500 case minimum frame rate is now 18 percent greater we have another mostly GPU bound game in for honor though the minimum frame rate is quite telling here the 2500 K despite delivering a more impressive average frame rate was actually 13% slower when looking at the all-important minimum frame rate here we see the 2500 K is on par with the much-loved Pentium G 4560 moving to 1440p the 2500 K is now just 2% sold in the FX 83-71 can minimum frame rate while it delivered the same 109 FPS average next up we have the very CPU intensive Gears of War fall and here the 2500 K was 38% faster than the FX AVC 70 delivering an average of one hundred and 60 FPS compared to just 77 fps the minimum frame rate was much closer though here the 2500 K was 14% faster moving to 1440p we again find the 2500 K well ahead delivering 38% more performance with an average of 105 fps this made the 2500 case slightly faster than the chloro 370 350 K and much faster than the FX 83-70 the 2500 cables are heading our 12 player bot tests which runs for 5 minutes and I take the average number from 3 runs here in the 2500 K was 5% faster when comparing the minimum frame rate and 10% faster for the average moving to 1440p the margins remain much the same so a clear win here for the 2500 K in what is a very CPU demanding game next up we have another CPU demanding title in Civilization 6 and here the FX 83-70 gets left well behind delivering similar performance for the pentium g 4560 meanwhile the 2500 K is well ahead producing 24% more frames the FX processor also gets wiped out completely at 1440p as it turns in another poor result well I actually came back after completing all my tests Andreea round the numbers in Deus Ex mankind divided only to find the same results the minimum frame rate for the 2500 K was 33% greater while the average figure was 46% greater the Sandy Bridge processor leaves the FX 83 74 dead here and the same can also be seen at 1440p as a side note I did compile many of my benchmark numbers before I had any rosin CPUs in hand so at the time I was unaware they had issues with DirectX 12 in this title as well as a few other titles I had actually tested with the assumption that the more modern API would actually favor AMD anyway going forward I will be testing using both DirectX 11 and DirectX 12 the FX 83-70 finds itself playing the role of a back marker in f1 2016 here was 29% sold in the 2500 K when comparing the minimum frame rate and 27% slower for the average bumping the resolution up to 1440p doesn't to help so I cry primal isn't particularly CPU demanding though for some reason AMD has issues with it anyway this time the 2500 K was 19% faster for the minimum and 24% faster for the average much the same as seen at 1440p the FX 83-70 is again swept aside this time when testing Grand Theft Auto 5 as the 2500 K was 33% faster when comparing the average frame rate that said the FX processor was just 8% slow and comparing the minimum again we see the margin slightly widened at 1440p which is most unusual here the FX 83-70 is now 14 percent slower when comparing the minimum frame rate even in hitman the FX 83-70 falls well behind dipping down to 44 FPS from a 3 run average this then made the 2500 K just over 40% faster when comparing the minimum frame rate moving to 1440p we see a rather large drop in the minimum frame rate and now the 2500 K is just 10% faster though admittedly that's a decent margin right there mafia tree is a very interesting title and I think you might start to see more reviewers picking this game up and adding it to their list of games that they test with the game was bashed for it's horrible optimization and rightfully so today it's still pretty average in terms of performance but for some reason AMD CPUs love it we saw the 1800 X beat out the 7700 K in this title and now we see the FX 83-70 outpacing the 2500 K for a rare victory when comparing the minimum frame rate the FX processor was 5% faster that margin is extended out to 10% at 1440p though the average frame rates are quite similar Mirror's Edge catalyst sees a 2500 K jump back in front leading the FX processor by a 26% margin when comparing the average frame rate the minimum was closer though the 2500 case still enjoyed a 16% performance advantage here moving to 1440p the margins close up as they often do and now the 2500 K is just 6% faster when comparing the average frame rate here we have yet another bloodbath total war war Hamas is a 2500 K pull ahead by a 30% margin for both the minimum and average frame rate at 1080p much the same as seen at 1440p and again this was another game where a Verizon had issues when using the dx12 API although the 2500 K was pegged at 100% utilization and watched - the experience was actually less jittery than the FX 83-70 although the minimum frame rates are similar and there is a bigger variance between the minimum and average on the 2500 K I can assure you the experience was better on the Sandy Bridge CPU at 1440p singh's look competitive but again the FX 83-70 suffered more performance issues in this game added purely to satisfy you guys that keep asking for it here is ARMA 3 the 2500 K was 30% faster when comparing the minimum frame rate and 36% faster for the average much the same was seen at 1440p and again it was interesting to see the pentium g 4560 was actually faster though again this is a lightly threaded game ok so now we have the big one battlefield 1 now over a computer base in their risin review which adored t v-- referred to they showed the 2500 K to be 29% slower than the FX 83-70 in the multiplayer portion of the game using DirectX 11 which I felt was pretty shocking the figures you are seeing here from the single-player portion which from what I've seen delivers very similar CPU utilization figures that's what I will be looking at the multiplayer portion of the game in a moment though comparing numbers accurately here is next to impossible but what I found was pretty convincing anyway so with the Titan XP in charge of the rendering work both CPUs looked quite good over 2500 K was 27% faster for the average framerate and 43% faster when comparing the minimum much the same we seen at 1440p a clear win for the 2500 K here and once again I found the experience was less jittery on the Intel CPU so as I just mentioned earlier I have shown in the past the CPU utilization looks exactly the same in the multiplayer portion of the game as it does in our single player test still many of you don't seem to buy that despite the evidence provider so in an effort to remove doubt I have included multiplayer performance as well what you're seeing here is footage from the conquest mode on the map giant shadow and both the AMD and Intel processors were tested in fully populated servers in this little snippet of footage I unfortunately wasn't able to follow the exact same path as guys kept killing me they didn't seem to care that I was just here for benchmarking anyway as you can see here the FX 83-70 is clearly well down on the 20 500k despite the Sandy Bridge processor operating at 100% capacity I should also note that it is still smoother on the Intel processor and there was less input lag here we have a bit more of a direct comparison again although the 2500 K is locked at 100% utilization it didn't suffer any noticeable frame drops in fact I was very impressed by how smooth the experience was although slower the FX 83-70 wasn't bad either having 16 gigabytes of system memory is very beneficial in this title anyway in a fully populated multiplayer game we see that the 2500 K is around 30% faster I have no idea where you would have to test in battlefield 1 to see the FX processor winning by a 30% margin on a final note some reviewers claimed to benchmark the multiplayer portion of the game by loading into an empty map well that's fine for testing GPU performance but using this method to test CPU performance isn't ideal as the CPU load is much lighter in fact even in a multiplayer game full of 64 players if you happen to wander off into a remote section of the map you will notice the CPU load drop away and the frame rates will come up as a result and we did see the FX 83-70 producing considerably better performance when doing this claims that the 8 core FX series had played serious catch up were seriously overrated as far as I can tell little has changed into the initial release back in 2012 Intel's quad-core 2500 K is still manhandling the power-hungry FX 8000 series before wrapping things up in what's going to be a bit of a lengthy wrap-up here is a look at the average performance from the 16 games tested this sums it up quite well of the 16 games tested many of which were modern titles the FX 83-70 was on average 20% slower to the 2500 K not 10% faster even when comparing the minimum frame rate the FX processor was still on average 12 percent slower in fact no minimum frame rate was comparable to the dual core Pentium G 4560 fairly strong showing for the FX 83-70 was seen in mafia 3 though it didn't exactly annihilate the 2500 K here if every game was as long piers mafia 3 then yes the FX 83-70 would indeed be 10% faster than the 2500 K but result is somewhat of an outlier when looking at battlefield one multiplayer performance I did refer to computer base I was able to replicate their week 2500 K performance in this title another test play around that should be fairly easy compared to my own data is f1 2006 and canned benchmark though their actual frame rates were inexplicably lower anyway I found the 2500 K to be 27% faster so at least we agree that in this game standing bridge does provide much stronger performance looking around the internet I also found an interesting article from legit reviews titled AMD FX 95 98 core CPU review last look before Rison this was a cool article worth checking out unfortunately it wasn't heavily focused on gaming so just thief grand theft auto 5 and Deus Ex mankind Nevada were tested and the 2500 K was omitted from the results still the core i3 73 50k was included and in mankind divided it was 42% faster than the FX 83 70 which is in line with my own findings in the end there's a very good reason why the FX series has avoided like the plague on the second-hand market you don't see too many YouTube channels dedicated to creating cheap used build guides recommending their viewers race out and find a cheap AMD FX system and in 2017 this isn't going to change in my opinion reviewers really nailed the FX 8350 s gaming performance way back in 2012 and as far as I can tell nothing has changed going forward I will continue to test CPU gaming performance at 1080p using the fastest crap is cut available at the time of testing if you don't think this is how CPU gaming performance should be tested then the best course of action would be to unsubscribe from the hardware unbox channel and subscribe to a channel that does the kind of testing you deem acceptable the alternative seems to be spamming links on YouTube channels and this isn't exactly an effective strategy the spam filter automatically grabs them so Samuel no need to keep spamming those links mate circling back to Rison and it's gaming performance it makes sense to me at least that we keep testing gaming performance the way we always have this shows the consumer exactly what they're buying into now and gives them solid information to make an informed buying decision once the factoring you can discuss thoughts and even speculate as to how things might play out in the future just I did in my review while I was wrapping this video up over the weekend though guys at PC perspective published a very interesting article that claims that updating the Windows 10 scheduler won't be a silver bullet for Eisen it's a very in-depth article that's well worth the read those guys really know what they're on about they started off by showing that Windows 10 is indeed identifying caused and attempting to balance workloads the next part of their article though was what interests me the most they moved on to ping cores to confirm their initial findings and in doing so uncovered something very interesting using the core i7 59 60 X they found that it took roughly seventy six and nanoseconds for the ping to travel from one physical core to another a ping within the physical core took just 14 nanoseconds here they were pinging the hyper threading thread anyway what's important to note here is that the haswell-e architecture has a latency of roughly 76 nanoseconds between physical cause moving to the risin seven 1,800 ex this is where things got interesting because Rison has split into two cc axes each with 4 physical cores things work a bit differently working within a single CCX they found a 26 nanosecond delay between pinging within physical cores that's slower than Intel but nothing extreme however pinging adjacent calls within the same CCX was very fast taking just 42 nanoseconds or there abouts the issue appears to be when moving from one cc X to another doing so added another hundred nanoseconds to the time taking a total of around 140 nanoseconds to complete the ping they found on average it took 87 percent longer to ping between CC X's on Rison than it does to ping between physical cores on the haswell-e 59 60 X they believe this is what's responsible for the weaker-than-expected gaming performance they explain why this is in their article so I won't go over that and regurgitate the entire thing I will provide a link in the video description so if you are interested please go check it out it's good stuff having said all that though Rison obviously isn't a poor gaming solution I'm yet to play a single game where these new AMD processes are anything less than silky smooth I personally would purchase the 1,700 over the 7700 okay because I'm not gaming on a high refresh rate monitor and I focus more on content creation although I would pick the horizon 717 hundred over the 7700 K please be aware that for the most part we're looking strictly at gaming performance the 7700 K is a superior performer and anyone publishing results or suggests otherwise is trying to mislead consumers the only game I have found so far where the rising CPU has come out on top is mafia 3 whether or not this is an indicator of future gaming performance is yet to be seen anyway I hope this clears things up for my subscribers I really wanted to explain to you guys why it is that I believe testing CPU gaming performance without a GPU bottleneck is very important and I also hope you enjoyed the FX 83-70 versus Core i5 2500 Kay battle it was a pretty interesting revisit if you did enjoy it too consider hitting the like button that'd be cool I'm your host Eve see you next time Oh