Gadgetory

welcome back to our unboxed it's time for yet another big benchmark video so strap yourself in last week and video released the geforce gtx 1070 Ti and I suffered through the launch with some kind of flu virus but the good news is I've just about fully recovered now because I was sick though that did limit my time and energy I had to throw out the release day coverage so I promised you that I would put together a 34 game benchmark comparing the gtx 1070 TI and nine other GPUs and well this is that video I've actually added a 35th game and that is Call of Duty World War 2 all 35 games were tested at IUP 1440p and 4k resolutions and this means that we have a grand total of 105 grafts comprising of eleven hundred and fifty five individual benchmark runs the results as always are based on an average of three runs and so that means there's three thousand four hundred and sixty five runs in total or well at least to that amount anyway this took weeks of work together and the issue on our face is how to present all this data without taking up a week of your time so what we're gonna do here is focus on a doesn't have the most recently released and most interesting games and then look at a number of head to head comparisons based on all 35 games you know those massive graphs you guys seem to really love they usually show at the end of the big benchmark comparisons anyway because I have exported and manually highlighted the GPUs that we're discussing at all 105 graphs I've decided to include them all at the end of this video with a bit of background music so consider it kind of like showing the workings out if you will for you guys that are interested to see how the frame rates looked in individual games at certain resolutions you can go and check those at the end of the video anyway first that we have 12 games to closely look at before jumping into the full breakdown graphs before that there are a few quick notes on testing the latest AMD and NVIDIA drivers were used and all tests were conducted on my new core i7 ATS owner okay test system previously many of these results were gathered on my 77 okay test system but they've been since retested on the 87 are ok many of them went unchanged but I did see a couple of big changes in games such as ashes of the singularity for example so that pretty much covers all the boring stuff it's now time to jump into the results kick-starting things we have the call of duty at World War 2 results starting with the 1080p resolution now right away you've no doubt noticed something a bit odd here all the Radeon based graphics cards are stuck at an average of 91 fps this is obviously a driver issue as Vegas 64 liquid-cooled should be quite a lot faster than an RX 580 for example in fact all the AMD GPUs were slower than the GTX 1060 at this resolution moving to 1440p we again find that the Radeon GPUs are limited to around 90 FPS on average though this time the RX 580 does drop off a little bit still because of this issue the gtx 1070 TI is a little over 20% faster than the vega 56 graphics card before we even bothered or overclocked it finally at the extreme 4k resolution that apparent 90 FPS cap of the AMD GPUs is no longer an issue as there is no single GPU here able to achieve that level of performance even so Nvidia still has the edge in this tile at least when comparing the gtx 1070 TI and vega 56 GPUs here the gtx 1070 TI was 8 percent faster while it was possible to extract a further 11% via for clocking moving on we have Assassin's Creed origins and here we have another new tire what looks to be favoring the green team the GTX 1070 Ti provided 9 percent more frames in Vega 56 and this meant at 1080p it was pushing well over 60 fps at all times the margin is diminished at 1440p and here the 1070 - I was just 7% faster when comparing the average frame rate that said though the minimum result weighed heavily in the green team's favor then finally at 4k the 1070 - I was just 3% faster with an average of 36 fps though again it was a good bit faster for the minimum frame rate that's it with less than 40 fps on average neither GPU was particularly impressive at the 4 pear resolution when using the very high quality preset next that we have the new Wolfenstein game often signed to the new Colossus like doom this title supports the Vulcan API so you might expect the AMD GPUs to do very well here Vega 56 is able to outpace the gtx 1070 TI even once the NVIDIA GPU is overclocked that said those things do change slightly at 1440p and here Vega 56 is just 7% faster than these stock gtx 1070 TI this time overclocking the 1072 i places at on par with the Vega GPU impressively both GPUs are averaging over 100 FPS using the maximum in game quality preset then finally at 4k we are still able to achieve very playable performance again value 56 is still slightly faster when comparing both GPUs at their default operating frequencies both Vega 56 and the gtx 1070 I performed very well in f1 2017 now this time the green team has a clear advantage in this DirectX 11 title the margins aren't huge but out of the box the 1070 Ti is faster than Vega 56 though the Vega Ravi's card does beat the plain old 1070 the margins close up at 1440p and though the 1072 AI is still faster there's now just a few frames in it so not really much to talk about there again this is a title where previously or I suppose still currently AMD's Vega 56 does enjoy a win over the 1070 but 1070 CI though does tip things back in invidious favour then finally at the extreme 4k resolution the gtx 1070 TI remains in the lead though again there really isn't that much in it Middle Earth shadow of war is another recently released title and this one plays very nicely with AMD Hardware as a result Vega 56 is faster than even the gtx 1070 TI at least a stock gtx 1070 TI here the red team enjoyed a 12% performance advantage over the vanilla 1070 at 1080p and 3% over the new TI model similar margins were seen at 1440p though this time i should know that overclocked in the gtx 1070 TI did boost the frame rates by 11% and consequently this meant that it was 6% faster than the stock vega 56 Jeep you then for those gaming at the 4k resolution you will need to reduce the quality preset as the GTX 1072 Ione Vega 56 GPUs were only good for like 40 fps jumping back in the cockpit this time for project cars too we find another racing simulator that prefers its cars with a shade of green at 1080p the stock out-of-the-box GTX 1072 I matched the Vega sixty-four liquid cooled graphics card while overclocking plates to the $450 GeForce graphics card within the 1080 TI's slipstream I should note though thanks to what is a fairly obvious CPU bottleneck that slow slow 8700 K at 5 gigahertz is unfortunately the culprit here jumping to 1440p the CPU bottleneck is alleviated as the GPU has much more to deal with now the margin between Vega 56 and the gtx 1070 TI grew ever so slightly and now the GeForce graphics card is 13% faster out of the box then finally at the 4k resolution the margin again grows and now the gtx 1070 is 15% faster than vega 56 and as a consequence is now hot on the heels of vega 64 liquid-cooled once again moving on now we have destiny to to check out and things look very similar to what we saw in testing the beater out last month by that I mean this game pretty much favors hardware from the green team that said though previously the GTX 1070 did beat Vega 56 by a pretty handy margin an au value is slightly ahead so AMD has made some improvements unfortunately though for AMD since then nvidia has of course released the gtx 1070 Ti and that kind of hides some of the gains they've made with Vega 56 having said that there at 1440p things really do turnaround for the Radeon GPUs and here we can see Vega 56 beating the GTX 1070 Ti even once it's overclocked so that's a very interesting and certainly very impressive result for AMD what's more interesting there are the 4k results as here AMD continues to gain momentum and now vega 56 is 15 percent faster than the overclocked gtx 1070 TI in fact vega 64 liquid-cooled is able to the GTX 10 atti and well you don't see that too often total war war hammer 2 was tested using the DirectX 12 API and this should give AMD a slight advantage that said in videos drivers are already appear to be pretty much up to speed and working well in this title and as a result the GTX 1070 Ti is just three percent slower than Vega 56 at 1080p so I think we should probably increase the resolution as we've already done numerous times and see if that pushes things further in AMD's favor like what we saw with destiny 2 well this time things actually closed up at the 1440p resolution and now Vega 56 and the stock GTX 1072 I deliver virtually the same performance this is again also seen at the 4k resolution as both GPUs deliver an average of 24 fps he'll blades built upon the Unreal 4 engine typically this engine favors in video hardware and we can certainly see that here the gtx 1070 TI was a whopping 19% fast in the vega 56 in what was one of the most convincing wins for nvidia yet the margin though is reduced ever so slightly at 1440 p.m. now the gtx 1070 TI is 16% faster even before the 1072 I arrived though Nvidia is plain old 1070 had AMD more than covered in this title again we see similar margins even at the extreme 4k resolution though as usual the hoh's gaming here will want to reduce the quality settings if they plan on using either a Vega 56 or gtx 1070 TI type graphics card another game released this year that also uses the Unreal 4 game engine is player on owns battlegrounds and this is yet another title it prefers in video hardware it's not exactly a well-built title though it's certainly not optimized at least to the degree that you would expect from a title of its popularity let's say and well I really only include it because of how popular it is the GTX 1072 I led Vega 56 by 32% margin that's not exactly the norm but it's still a disappointing result for AMD in this hugely popular title as we often see the margin is reduced at 1440p that said it's still quite a massive margin in favour of the G Falls graphics Carter's here it was 28% faster then at the 4k resolution the 1072 I was 24% faster but it hardly matters at this point with almost every graphics card tested struggling Oh chief 30 frames per second from a terribly optimized game to an extremely well optimised game we now have the prey results this games provide us with plenty of competitive results in the past and this is again true here though the updated gtx 1070 TI now has the rendering power to pull ahead of vega 56 the results though are quite interesting as with the 1070 TI you do get the minimum frame rate of the gtx 1080 but the average result is much closer to the gtx 1070 so that's quite curious I have for the most part the margins though were much the same at 1440p here the gtx 1070 I led Vega 56 by 10% margin while it enjoyed a 12% boost once overclocked then at 4k the margins remain much the same though now we're seeing frame rates in the 40s the last game that we're gonna have a close look at is dirt 4 and we found that this title does heavily favor the red team when using any anti-aliasing method other than the ultra heavy 8 times MSAA I've always used CMAA as it looks quite good and it's not very taxing on performance here value 56 berries the gtx 1070 TI and it does seem like in videos driver team are unable to make up this massive deficit at 1440p the margin is reduced but even so Vega 56 was 17% faster than the stock gtx 1070 TI finally at 4k Vega 56 seems to choke a little and now the gtx 1070 TI is ample and matches performance with 61 FPS this is quite an impressive result for these graphics cars at this extreme resolution lastly just a quick recap on power consumption we looked at this a few days ago in the day one coverage of the gtx 1070 TI and there was there are a few more cards let's say in that comparison there is yourself you want a bit more detail go back and check that anyway as you can see the power consumption between vega 56 and the gtx 1070 TI when comparing full-system draw is very similar obviously the pascal base 10 70 TI is much more efficient as it was faster overall but managed to draw less power right so we've now looked 36 of the 105 graphs let's skip the rest now and break down the data I've gathered from all 35 games I've mostly focused on the 1440p resolution as I feel this is the most relevant for GPUs of this caliber overclocking really wasn't the focus this video and I realize I didn't overclock Vega 56 so it's not an overclocking comparison that said though you kind of have to overclock the 1070 TI for it to make sense otherwise you might as well to spend a little extra on the 1080 or save a bit and get the 1070 anyway as you can see here once overclocked the gtx 1070 TI produces on average around 11% more frames which isn't bad it's certainly worth the small amount of effort it takes to achieve of course the gains will vary quite a bit depending on the title use for testing but as you can see for the most part you will see double-digit gains when compared to MSI's factory overclocked gtx 1070 gaming x graphics card these stock gtx 1070 TI was an average just 8% faster which might seem a bit surprising given that the 1070 TI packs 27% more cuda cores there are however two reasons why that margin might be smaller than expected firstly the MSI gaming X version of the gtx 1070 features a factory overclock whereas the 1070 TI doesn't so this does help reduce the margin the other reason being that the gtx 1070 TI is memory starved and this means all those extra criticals aren't really being utilized as effectively as they should be overclocking the 1070 TI's memory really does help out as you're about to see now with my custom overclock applied the 1070 I is 20% faster than the msi gtx 1070 gaming ex model in fact we're now saying up to a 28% gain and tiles such as fir honor f1 2017 doom and the division with both the 1070 and 1070 ti overclocked to the max the TR model should be I would say at least 15% faster on average overclock the 1070 TI is also able to deliver factory overclocked GT X 1080 light performance as it matched the MSI gaming X mile they were evenly matched in six other games tested while the overclock 1072 I was faster in 11 games albeit by just a 1 to 2% margin in most of them additionally the 1070 I was never more than 5 percent slower so you can safely say it delivers at least stock GG X 1080 performance once overclocked in my big 32 game Vega benchmark video though did a few months ago now I found that Vega 56 was an average 5 percent faster than the geforce gtx 1070 founders edition at the time i noted that the factory overclocked custom board partner versions of the gtx 1070 would be around 5% faster than the founders edition model so overall the 1070 and Vega 56 graphics cards would deliver similar performance well this time for the 35 game benchmark which includes many new and previously untested games I've used the MSI gaming X version of the gtx 1070 and here we see it was just 1% slower than vega 56 on average so that's about what I was expecting to find newly-released titles that helped AMD out include Wolfenstein 2 shadow of war Warhammer 2 and destiny 2 the big wins for Nvidia came from player unknowns battlegrounds Mass Effect Andromeda and the newly released Call of Duty World War 2 title so prior to the gtx 1070 TI's release AMD were still hanging in there with a pretty close head-to-head battle with the factory overclocked n 70's however if we now swap to the 1070 Ti we find that on average this new GeForce GPU was 6% faster at 1440p in fact the 1070 TI was faster in 25 of the 30 games test and only suffered a double-digit loss in dirt floor and destiny 2 so in today's games it's fair to say that for the most part the geforce gtx 1070 TI is the superior GPU before wrapping things up i also checked out the margins at 1080p and here the 1072 i made out a little better it's certainly the better option for those seeking extreme frame rates at this resolution the only game to really hurt the 1070 I was again dirt 4 but for those that would claim this is an outlier and it's skews the results I can tell you that it doesn't that's the beauty of testing with so many games of course removing dirt 4 does only change the margin by a percent and that set up we memory of Call of Duty World War two unsurprisingly we're back to a 7% margin and then finally we have the 4k resolution and in my opinion neither of these GPUs are really capable of 4k gaming but you know here are the results anyway the gtx 1070 TI was 6% faster on average at this extreme resolution again the 1072 i was faster in the majority of the games tested well is there anything else that needs to be said I guess at this point it pretty much comes down to pricing availability performance wise you're looking at around 6 percent on average and the difference between these 2 cards and again that is just the average difference across the 35 games tested so if you plan on playing a shipload of pub G for example the gtx 1070 TI is the obvious choice it was 28% faster at 1440p in that title so though if you're gonna mostly play something like destiny 2 then perhaps Vega 56 would be the way to go as the 1072 I was 11% slower in that tile that said that I think if you're gonna play a mix of games and that's probably sort of realistic for most to you guys you're not going to stick to all one title I know some people do but yeah anyway if you're gonna play a mix of titles and I think it's safe to say that the nvidia geforce gtx 1070 TI is the best bet right now as always i suggest you look at the performance in the games that interest you to work out which brands gonna serve you best keeping the performance figures in the back of the old noggin let's talk Mahler in a post cryptocurrency mining apocalypse when MSRP z' actually mean something again for the gtx 1070 TI should cost $450 us the gtx 1070 $380 us and the gtx 10 $8,500 us and finally vega 56 should cost around $400 us actually to my surprise at the time of creating this video it was possible to buy a vega 56 graphics card for $400 us and a gtx 1070 TI for 450 dollars us what kind of madness is this hopefully this is still the case when you watch this video i kind of don't believe it but over Amazon and Newegg those are the prices I'm currently seeing this means the gtx 1070 TI cost around 13% more than a reference vega 56 graphics cutting honestly that makes picking between the two a bit of a challenge here in australia the pricing is much the same for both models you can expect to pay around 700 to 750 dollars Ozzy so keeping those prices in mind I'm not really sure why the gtx 1070 CIA has been copying so much hate especially from some of the big-name reviewers out there sure i'll admit it's an exciting and i did say this in my day 1 review but right now it is only $40 u.s. more than a 1070 and that's 10% more and with both overclocked it is around 15% faster that's hardly exciting gains I know but for someone buying a new graphics card today it is the better value option at least for enthusiasts that keen on overclocking that said that I can certainly understand the argument that if you're gonna spend four hundred and fifty dollars why not just spend five hundred or at the moment it's five hundred and ten dollars to secure the gtx 1080 it is just 13% more after all at the five hundred and ten dollar price as I said in my review the gtx 1070 TI it really does fill a gap that we didn't really need filled but it seems Nvidia just doesn't want to give AMD an inch in the end though if I had to pick between a reference Vega 56 graphics card at $400 or a custom board partner gtx 1070 TI at 450 dollars i think i'd lean towards the GeForce graphics card for the simple fact that in most games does provide superior performance consumes less power albeit only slightly less runs cooler and is a heck of a lot quieter operating volume really is the big issue for me with the reference Vega 56 graphics card it's loud I don't care what anyone says it's bloody loud and I certainly couldn't live with it I am hoping to get my hands on a custom Vega 56 graphics cards soon but boy oh boy it's not an easy task and good luck if you're a consumer in the end I think the gtx 1070 TI wins juda better availability significantly more options for consumers to choose from and better quality options at that there is still hope for Vega but with each passing month that hope is dwindling and I'd say AMD's GPU armors now pretty much on life support everything would have to be hinged on the next generation which could really make or break the Radeon team as it is in videos basically ganged up on Vega 56 the gtx 1070 was already really giving it a hard time I what was a bit of a back-and-forth fight but now the gtx 1070 TI is just shamelessly jumped into the ring to kick Vega while the 1070 holds it down Vega 64 would no doubt like to step in and help but it's already been dropped punted across the ring by the GTX 1080 and the liquid cool model wet itself at the site of the GT X 1080 TOI I could go along with this story but I'll spare you and stop here and instead I'm gonna hit you with 69 more graphs and though that wasn't intentional hey I'm your host Steve again sorry about my voice it is getting better I should be fully recovered in a few days time but until then here is a truckload of blue bar graphs enjoy you