Gadgetory

welcome back to harbor unboxed today I had planned to address Verizon 7 memory scaling performance but for now that's been put on the back burner though testing is almost complete and the results are very interesting so yesterday morning when I got out of bed there were several dozen comments emails and Twitter messages from you guys asking me to test Rison using an AMD GPU or perhaps even two with so much demand for these tests I decided to do them the faster single GPU on hand from AMD right now is the fury X and since that delivers similar performance of the geforce gtx 1070 i decided to compare them using the Rison 7 1800 X it seemed that DirectX 12 games we're in demand and oddly at the top of everyone's wish list was rise of the Tomb Raider I say oddly because this is a game I consider quite poor for measuring system performance in particular CPU performance in fact there was a good reason why it wasn't included in my 16 game horizon performance video a few weeks ago the request was they also test a single level of the game called geothermal valley so I have I actually didn't really have many requests to test other DirectX 12 games which seems a bit strange testing one level of one game seems kind of limited and definitely not the hardware unboxed approach so an effort to cover my bases I also throw in hitman Deus Ex mankind of audit ashes of the singularity escalation Tom Clancy's the division and some alternative testing with battlefield 1 and Far Cry primal now don't expect to see too much or any Intel CPU testing here really I wanted to focus on Rison and how it works with the fury acts and gtx 1070 i didn't want to risk turning this into another AMD versus intel clash though i can add those results a bit later if need be at this point pretty much every day next week is booked out already the channel has gotten a bit crazy lately and it's demanding a huge amount of my time for multiple projects anyway with the limited time I have available right now this is what I be able to put together for you guys before we jump to the results please note that both the horizon 7 1800 X and core i7 7700 K processors will locked at 4 gigahertz with any kind of power saving or turbo boost type features disabled so basically the horizon 7 CPU was overclocked and the 7700 K was under clocked though please note I only can the Intel CPUs performance in battlefield 1 and a single rise of the Tomb Raider test both CPUs were running with ddr4 29:33 memory ok so starting with hitman things are very much like what you would expect to see in video recently improved their DirectX 12 performance in this title as you can see the gtx 1070 was 7% faster than the fury acts when comparing the average frame rate similar margins are also seen when testing with the dx11 api so nothing unusual here the division was tested using the ultra quality settings which mostly loads up the GPU to be honest that said we are testing at TEP the fury does enjoy strong performance gains when using the DirectX 12 API though that's not surprising that said the GTX 1070 was also faster using DirectX 12 though the gains weren't as substantial again nothing looks out of the ordinary here I have tested Deus Ex mankind divided using both the canned benchmark as well as in-game performance using present Mon here we can see the canned benchmark results and it's interesting to see if the Fuhrer ex takes quite the pounding on the minimum frame rate that said once again the gtx 1070 is slightly fast using DirectX 12 when compared to DirectX 11 still the margins are much the same between the fury x and gtx 1070 using either API moving to in game testing towards the start of the first level we find a fire less demanding situation the fury X isn't plagued by those low minimums now that said the gtx 1070 is still the superior performer even when using DirectX 12 again the margins are much the same at least when looking at the averages a game of been keen to test with ancho uprisings new found form as ashes of the singularity escalation that testing is still to come and you will likely see it in my rising 5 coverage what you'll see here is the fury X getting completely annihilated when using DirectX 11 moving to DirectX 12 the fury ax performs much better though it is once again the gtx 1070 that is able to extract the most out of the rising 7 1800 X processor okay now for the rise of the Tomb Raider testing which you guys have all been requesting first we'll look at the results from the can benchmark focusing on the geothermal valley section of that benchmark running DirectX 11 the gtx 1070 was 29% fast and if your ex keep in mind this is an nvidia sponsored title that said though AMD how made great strides here at least with the rx4 ad from what I've seen from previous tests anyway moving to DirectX 12 we now see the gtx 1070 is 23% faster so the margin was reduced somewhat that said once again the gtx 1070 is getting the most out of the 1,800 x initially I tested the start of the geothermal valley level but found the frameworks were quite high quite a bit higher than those from the built in benchmark the shocking part here being that the gtx 1070 was now 41 percent faster than the fury acts when using DirectX 11 though it was only 27 percent faster for the DirectX 12 test right so this test takes place right after the first cutscene and runs to 60 seconds it is considerably more demanding than both the previously seen tests the margin between the gtx 1070 and fury ax has been reduced though the GeForce GPU was still 26% faster using DirectX 11 and 21 percent using DirectX 12 so we aren't even seeing the fury X match the gtx 1070 in these titles using an overclocked verizon 7 processor before we move on to look at the gameplay footage compared side-by-side here is a quick look at how the 7700 k and 1800 x compare in the same section of the geothermal valley level using a single GPU the margins look much the same using either the AMD or nvidia hardware so there isn't much more to say here really let's move on now the support the testing done so far here is the geforce gtx 970 and radeon r9 fury x running side-by-side in the very past we just looked at the results for please note the DirectX 12 API is being used there are some interesting things to note here other than the fact the gtx 1070 is consistently ahead in terms of framerate performance you will notice that the CPU usage is considerably higher when using the gtx 1070 though that makes sense giving it is delivering stronger frame rates as for the GPU utilization please note that the first percentage figure on the gtx 1070 side is showing power i should have an ticked this feature and rivatuner to avoid confusion anyway be aware that's what you're seeing here after the temperature we have the actual GP utilization figure you will notice that for the most part the fury X is pegged around a hundred percent while the GTX 1070 often dips down into the 80s in fact it really hits 100% in this past so with a higher soft CPU there is a good chance you will see more performance for those of you interested here is the same pass using the DirectX 11 API again CPU usage in general is high when using the gtx 1070 though at times the fury x does creep ahead it's certainly closer than what it was when using DirectX 12 that's it even with DirectX 11 we still see fairly even loading across all out cause you also notice that the threads aren't as heavily utilized the threads being even numbers and the cause odd GPU utilization jumps all over the place using DirectX 11 and it's rare that we see the gtx 1070 exceeding 90% before wrapping up the testing i have just a few extra things i would like to show so there have been some very interesting reports that the Radeon our X 480 crossfire graphics cards deliver exceptional performance on the horizon 7 platform and allow it to push past what the GeForce graphics cards are capable of seems far-fetched based on what we have just seen but is it true well unfortunately I don't have to rx4 80s right now I did a few days ago but I sent one of them off to Thermaltake with the Rison 7 giveaway PC that you guys will be able to check out next week and of course one of you will be able to win what I do have laying around is this old relic the jackhammer of graphics cards AMD's epic Radeon r9 295x2 essentially what we have here is a pair of 290x GPUs teamed up on a single PCB so in terms of performance this should be similar to the rx 480 duo since a pair of 290x GPU should be faster than the GTX 970 I swap that out for the Titan XP let's go see what we found now you would expect the Titan XP to be faster than the r9 295x2 for the simple fact that it is in over 90% of the games out there particularly when testing at 1080p however here it's not in fact it's a good bit slower at least when using DirectX 12 the 295x2 is at times up to 27% faster and it's allowing the horizon 7 1800 X to hit frame rates previously thought impossible given the Titan XPS performance although Seifer utilization is often higher when using the Titan XP what you will notice is that something is stopping the NVIDIA GPU from hitting full stride on the horizon platform while the dual GPU 295x2 does a great job of keeping all 5,000 132 stream processors occupied the same can't be said for the Titan XP and its 3584 CUDA cores the Titan XP utilization is shockingly low rarely does it reach 70% in fact for the most part utilization can be seen hovering around 55 to 60% so I imagine we could boost the Titan XP framerate by 30 to 40% taking that into account and that would place it consistently ahead of the r9 295x2 so this certainly looks to be some kind of bug here you might be thinking the Titan XP will be fast using DirectX 11 well no in this title that's not the case in the gtx 1070 and furia comparison we saw that when using DirectX 11 frame rates drops for both GPUs that said crossfire support appears to be almost broken using DirectX 11 and as a result the r9 295x2 is all over the place it was basically a stuttery mess we notice the Titan XP isn't performing anywhere near its full potential on the horizon 7 platform and I'm not going to speculate as to why this is it just seems to be some kind of bug for now though let's compare it side-by-side with the gtx 1070 a GPU should be at least 40% faster than at 1080p instead what we see here is that at times performance is exactly the same in fact at the most the Titan XP only manages to pull ahead by 12% I also thought it might be a good idea to check out some DirectX 11 titles one tile has shown really poor performance on rising CPUs as Far Cry primal here we see that when using the gtx 1070 and fury x on both the 1800 x and 7700 k cpus both of which are clocked at the same 4 gigahertz frequency we see that even with the fury x the 1800 ex still falls behind the 7700 k by a slim margin what we are seeing here is a GPU bottleneck so what happens if you emulate Titan XP light performance by lowering the quality settings and resolution does this see the fury ax pull ahead on the rise and platform here we see that no that isn't the case the 1800 X imposes a CPU bottleneck at around 120 fps which as we have said in the past is still very acceptable but not a true reflection of Verizon's true gaming performance either the 7700 K on the other hand clocks just 4 gigahertz pushes well ahead with both the GTX 10 seventy and fury X having seen how the Titan XP and r9 295x2 compared a rise of the Tomb Raider I was very intrigued and wanted to test some more games or more DirectX 12 games I was pressed for time though so they'll probably have to be a follow-up video at this point the biggest issue I had when trying to conduct that testing was the lack of crossfire support in those games for example Deus Ex mankind divided wouldn't work at all in either DirectX 11 or DirectX 12 and I wouldn't work I mean it was a stuttering mess so that was disappointing for me right the Tomb Raider looks to be an outlier or at least that level of the game I actually don't have any other results to go on for this title right now as I never used it for CPU benchmarking and I certainly didn't include it any of my original Rison testing the game is well known for its poor optimization which is a shame giving how amazing it looks I'll admit the r9 295x2 results are very interesting but I don't think this one game tested using one level is enough to say that in video GPU is a handicapping Rison I'm pretty sure that isn't the case what I wanted to test was a game that I knew was well optimized for both Nvidia and AMD hardware and didn't suffer from poor optimization for me that title was battle through one at least when using the DirectX 11 API even at 1080p the Titan XP was often pushed to around 90 percent utilization using the 77 okay and here we see it delivering very impressive frame rates on both the 1800 X and 77 or okay meanwhile the 295x2 utilization was all over the place it's safe to say crossfire isn't working correctly here either I have red mixed report online surrounding battlefield ones crossfire support most seem to be seeing the same kind of performance issues that I am i almost never play around with crossfire or sli is I think both technologies leave a lot to be desired and you guys should definitely avoid them unless you have money to burn so I can confirm there is some funny business going on with the Nvidia base GPUs and rise of the Tomb Raider at least on the level tested when using the Rison platform I will be asking if they are aware of this issue and if a fix is in the works okay so a couple of things I would like to just make note of all the game performance figures that I have gathered using Rison 7 today we are done using the most up-to-date bias revision ddr4 29 memory and a fully up-to-date version of the Windows 10 operating system all this sees performance go unchanged in all the gaming titles are originally tested in my day one risin review so there hasn't been a magical Windows 10 update quietly pushed out that improves risin 7 performance I've heard this so many times over and over again so I thought I might as well address it likewise I did test Windows 7 performance a little while ago and found it to be a bust as well still I don't know why people need to keep making up stories or carry on about Rison 7 gaming performance at all performance is excellent particularly given this is a brand new architecture with no history already we're seeing games such as ashes of the singularity escalation and dota 2 updated to better take advantage of Rison and the performance gains are exceptional so people really need to chill out and enjoy the fact that we now have some real competition on our hands as for my rise in 5 testing I won't be including rise of the Tomb Raider or Far Cry primal instead I will be picking up ashes of the singularity escalation finally once amy has a gtx 980ti challenger then we can do some serious AMD vs. Intel CPU testing we could probably also do some exciting game D vs. NVIDIA GPU testing as well that's all for this one guys I know it's a bit of a quick and dirty benchmark video and I did just slap this together over the weekend in just under 30 hours so forgive me for not going into the usual depth if I didn't get this video out today it might have been a week or two before I have a chance to look at it again so I hope this satisfies those of you who have been asking me to check out AMD verson video GPU performance on the rise and platform I'm your host Steve catch again soon you