NVIDIA's latest GPU is here and it offers a big performance bump, but what exactly does that power deliver the VR gaming enthusiast? We pit the new Nvidia GTX 1080 Ti against the GTX 1080 to see just how far each card can enhance VR image quality through supersampling. It's frightening the pace at which the GPU industry moves. Here we are, less than one year after Nvidia launched its brand new line of 10-series 'Pascal' architecture graphics cards with the GTX 1080, back with a new card which promises to not only outgun its predecessor by a significant margin, but on paper matches the performance of Nvidia's flagship GPU, the ludicrously pricey and powerful Titan X. [irp posts="60611" name="Analysis: Latest VR Ready GPUs Benchmarked and Compared with NVIDIA 'FCAT VR'"] Table of Contents Introduction GTX 1080 Ti Specs Testing Methodology & FCAT VR Primer Overclocking Benchmarks Standard Gaming Performance Rise of the Tomb Raider (DX11 and DX 12) Deus Ex: Mankind Dividied The Division VR Gaming Performance Dirt Rally VR Showdown Conclusion NVIDIA's GTX 1080 Ti – the Titan X Killer? The new GTX 1080 Ti is here and offers a step change in performance when compared with the last generation, Maxwell architecture GTX 980 Ti. This is certainly impressive, and you can see why Nvidia are keen to emphasise the progress that's been made since the 980 Ti's launch in 2015. But the real story here is that this new card's closest performance stable mate is the current generation $1,200+ ultra-enthusiast card, the Titan X. In fact, the GTX 1080 Ti is built around the same GP102 GPU used in Nvidia’s Titan X released last year. With 12 billion transistors, GP102 is "the most powerful GPU Nvidia has ever made for gaming." [caption id="attachment_60599" align="alignright" width="325"] 1080 Ti block diagram shows the card's underlying architecture[/caption] The GeForce GTX 1080 Ti ships with 3,584 CUDA Cores, 28 Streaming Multiprocessors (SMs), and runs at a base clock frequency of 1,480 MHz, while the GPU Boost clock speed is 1,582 MHz. And as we'll discover, there's quite a bit of headroom in both memory and core base clocks. The 1080Ti sports 11GB of GDDR5X VRAM, just 1GB shy of the Titan X, and that's a spec shaving that you're very unlikely to notice, even when gaming at 4k or supersampling at extreme levels. In other words, the 1080Ti just made the Titan X effectively obsolete. Bear all of that in mind, and consider that the new GTX 1080 Ti shipped last week for $699, the same price as its GTX 1080 predecessor went on sale for just 10 months ago. It's also launching at this price a mere 8 months after the 10-series Titan X, owners of which may justifiably feel their wallet wincing at their short lived performance supremacy. Testing Methodology & 'FCAT VR' The world of cutting edge GPUs may move quickly, but one of the reasons why virtual reality remains fascinating is that it's moving even faster. Last year's GTX 1080 review opened with an apology of sorts, stating that as VR itself was in its infancy, we had no tools to record metrics at the level of empirical detail which standard PC gaming enthusiasts take for granted. As of this week, we're allowed to publish benchmarks based on the newly released FCAT VR tool from Nvidia, a new frame analysis tool which records VR runtime data in detail and lets us peek under the hood at if and when VR rendering safety nets like Asynchronous Spacewarp and Asynchronous Timewarp/Reprojection are kicking in under load. [irp posts="56455" name="NVIDIA Announces 'FCAT VR' Frame Analysis Tool to Help Demystify VR Performance"] As the 1080Ti is considered a high-end GPU for dedicated enthusiasts, we wanted to really get to grips with the benefits such extreme performance could provide VR gamers. Whilst current generation headset displays are limited in terms of overall pixel density (meaning a visible panel structure), one of the biggest immersion breakers are jaggies (aliasing) caused by a low target render resolution. We've therefore concentrated our VR benchmarking efforts to test the limits of the GTX 1080 and 1080Ti and their ability to supersample the image to extreme levels. Supersampling is a compute intensive way to reduce aliasing (the appearance of obvious pixels or stepping on a digital image) by first rendering at a much higher resolution and using that extra detail to down-sample to a lower resolution, but one of a much higher resultant quality. Supersampling is the easiest way outside of game-specific rendering options to improve image quality and immersion. As man cannot live on VR gaming alone, we've also assembled a selection of visually sumptuous and computationally taxing games. each benchmarked with tests designed to highlight the raw grunt each card possesses. Overclocking Although we've only had limited time with the 1080Ti thus far, we did manage to ascertain what we think is a stable (and fairly generous) overclock on our supplied founders edition unit. Pushing the core clock to +170Mhz above base with an additional +400Mhz bump for memory, we cautiously kept fan speed fixed at 80% with temperatures maxing out around the 80-85 degree mark. These numbers are provisional, but provide a healthy boost to performance and that's with no additional cooling or voltage applied - and they proved stable. We've included overclocked results in some of the benchmark breakdowns. Interestingly - for those of you squeamish about damaging such a pricey piece of hardware - you actually only need to lift the cap on the card's power and thermal throttling limits to realise some significant gains. Testing Rig We partnered with AVA Direct to create the Exemplar 2 Ultimate, our high-end VR hardware reference point against which we perform our tests and reviews. Exemplar 2 is designed to push virtual reality experiences above and beyond what’s possible with systems built to lesser recommended VR specifications. Test PC Specifications: SuperNOVA 850 G2 Modular Cables, 80 PLUS® Gold MAXIMUS VIII GENE LGA 1151 Intel Z170 HDMI SATA 6Gb/s USB 3.1 USB 3.0 mATX Intel Motherboard Core™ i7-6700K Quad core (4 Core) 4.0 – 4.20GHz TB, HD Graphics 530, LGA 1151, 8MB L3 Cache, DDR4-2133 ACX mITX CPU Cooler 16GB (2 x 8GB) HyperX Fury PC4-17000 DDR4 2133MHz CL14 500GB 850 EVO SSD, 3D V-NAND, 540/520 MB/s 1TB Barracuda®, SATA 6 Gb/s, 7200 RPM, 64MB cache 2 x 120mm Quiet Case Fan, 1500 RPM, 81.5 CFM, 23 dBA, White LED Custom 20-Color LED Lighting w/ Remote Windows 10 Pro 64-bit Edition Continue to 'Standard Gaming Benchmarks' >> Table of Contents Introduction GTX 1080 Ti Specs Testing Methodology & FCAT VR Primer Overclocking Benchmarks Standard Gaming Performance Rise of the Tomb Raider (DX11 and DX 12) Deus Ex: Mankind Dividied The Division VR Gaming Performance Dirt Rally VR Showdown Conclusion Benchmarks Standard Gaming Performance Even the most ardent of VR fanatics occasionally takes time out to enjoy the odd standard game or two, so rendering performance in standard monitor-bound titles has to be a big factor in whether the new GTX 1080 Ti represents a worthwhile purchase. Also, our chosen titles represent a real workout for even these top-end GPUs, especially at the 3440×1440 native resolution of Asus’ X34 Predator monitor. How well the GTX 1080 Ti does here gives a good indication of how much raw power the GPU brings to the VR table too. Rise of the Tomb Raider (DX11 and DX12) DirectX 12 is on the rise as far as Windows gaming APIs are concerned (no emails please OpenGL fans!) and games with DX12 support are now common. We chose Rise of the Tomb Raider (2015) as a particularly gorgeous (and demanding) example to include. Every setting is set to its highest available, AA and DX level is noted per-graph. Stock clocks, power and heat limits Power limits @120% | Heat limits raised to 90 deg. Fan speed manually set to 80% Overclocked +170Mhz Core / +400Mhz. Fan speed manually set to 80%. These benchmarks not only indicate the significant uptick in performance headroom versus last year's GTX 1080, it also provides existing 980Ti owners a tantalising view at the gains they could enjoy upon upgrading - up to double the performance even before overclocking in fact. Deus Ex: Mankind Divided The second title in the 'next generation' offspring of the first person RPG classic, Deus Ex: Mankind Divided (2016) is a curious technical offering. Its dense, baroque futurescapes are rendered with volumetric lighting and and an impressively effective implementation of temporal anti-aliasing and at higher resolutions push even the most powerful PC hardware. Stock clocks, power and heat limits "No Limits" aka Power limits @120% | Heat limits raised to 90 deg. Fan speed manually set to 80% Overclocked +170Mhz Core / +400Mhz. Fan speed manually set to 80% Deus Ex: MD has been a demanding game to run since launch, and even an overclocked 1080ti doesn't manage this resolution and temporal AA at 60FPS. Included here are figures which highlight that, even if you don't overclock the Ti's core or RAM, just unlocking the temp and power limits gives you a significant boost in performance - a nice option for the nervous (or sensible). The Division Ubisoft's massively multiplayer online third-person shooter, The Division (2016), presents a haunting yet stunning vision of a near-future New York in the aftermath of a viral outbreak. What the game lacks in narrative flair it makes up for in visual fidelity and atmospheric production design. Despite being over a year old, The Division still presents a challenge for hardware when all eye candy is engaged, especially at high resolutions and supersampling levels. The Division sports an in-game benchmark, but it's unfortunately hobbled at 16:9 only aspect ratios, which means we were only able to test this monitor's maximum vertical resolution of 1440p. That's not a resolution which these cards will sweat over, so we decided to run a set of tests at 1:1 pixel scale (a vanilla experience) and another run using the game's 200% pixel scale setting, an effective x2 supersample (or 4x the pixels per frame). The visual results are stunning as fine detail is resolved at greater distances and the image is rendered with extreme stability and barely a hint of jaggies. All settings were run at the in-game 'Ultra' preset. The results below show results for the following: Stock clocks, power and heat limits Power limits @120% | Heat limits raised to 90 deg. Fan speed manually set to 80% Overclocked +170Mhz Core / +400Mhz. Fan speed manually set to 80% The expected gains can be seen when supersampling is not in force above, and again those performance bumps from unlocking power and temp limits are obvious. Visually speaking, running this game in 2x supersampled mode delivers a stunning image and an overclocked Ti will glean playable (if not ideal) framerates here, one to try if you happen to own a G-Sync monitor perhaps. Continue to 'VR Gaming Performance / Conclusion' >> Table of Contents Introduction GTX 1080 Ti Specs Testing Methodology & FCAT VR Primer Overclocking Benchmarks Standard Gaming Performance Rise of the Tomb Raider (DX11 and DX 12) Deus Ex: Mankind Dividied The Division VR Gaming Performance Dirt Rally VR Showdown Conclusion VR Gaming Performance With FCAT VR we finally have a tool which allows us to peek under the hood of today's complex PC VR render pipeline. With VR's lofty technical demands, both main PC VR hardware vendors offer up mitigation strategies for helping hardware and software to achieve the 90FPS requirement for comfortable experiences in both the Oculus Rift and HTC Vive headsets. These strategies (Asynchronous Timewarp / Spacewarp / Reprojection) involve the insertion of synthetic frames when rendering can't hold a consistent 90 FPS. [irp posts="56455" name="NVIDIA Announces 'FCAT VR' Frame Analysis Tool to Help Demystify VR Performance"] FCAT VR exposes not only frametime latency (how long the PC took to render an image) but also dropped frames (where Asynchronous Time Warp kicked in) and importantly how many synthetic/re-projected frames using Oculus' Asynchronous Space Warp were recorded on a benchmark run, and it displays them in nice, pretty graphs. We'll have performance stats on SteamVR and HTC Vive hardware soon too, but for this article we benchmarked using the consumer Oculus Rift headset and version 1.12.x software and runtime. In the metric images below, you'll see each split into two, the top section detailing frametime latency (remember, we need sub 11ms results per frame to hit a solid 'genuine' 90FPS) and a breakdown of how frames were eventually delivered to the headset in the bottom section. You'll see dropped frames (marked in red and where Asynchronous Time Warp will kick in automatically to stabilise the experience) and where synthetic frames had to be generated due to latency rising too high (marked in yellow - indicating where Oculus' Asynchronous Space Warp kicked in). ASW is the absolute last resort in these scenarios and should be avoided if possible. But bear in mind that some synthetic frame generation may be undetectable so it is not the end of the world for the VR experience. Read more on the topic in our group GPU test with FCAT VR right here. Dirt Rally VR One of the most impressive VR gaming experiences to date, Dirt Rally also happens to be one of the best driving games in quite some time. Codemasters' implementation of VR here is excellent and as an argument for how virtual reality can enhance and transform a standard gaming experience, Dirt Rally works brilliantly. Unfortunately, Dirt Rally's in-game benchmark doesn't engage fully when in VR mode, which meant coming up with a repeatable test drive for each run. In this case we drove for 60 seconds per test, increasing Supersampling each time - between 1.0x, 2.0x, 2.5x and 3.0x via Oculus' debug tool. For each run we took as similar a route and speed as possible, keeping our head position mostly fixed throughout. Below are the combined graphs for each run with the raw results numbers below - click to enlarge. 1.0x PPDPO (No Supersampling) Some anomalies here - you can see each card more than capably stayed below the 11ms threshold, and despite this we see occasional dropped frames - but they clearly have no impact on the VR experience. We'll be speaking to NVIDIA about these to try and explain them. 2.0x PPDPO (2x Supersampling) At 2x SS, both cards still turn in a solid performance, but you can see the divergence in frametime beginning to show. There's a performance blip towards the end of the experience here, but clearly the digression wasn't great enough for the runtime to enforce ASW - and the average delivered FPS seems to back that up. 2.5x PPDPO (2x Supersampling) At 2.5X we're pushing some serious pixels and the 1080 now solidly exceeds 11ms in frametime, so much so that ASW has had to step in from the off to stabilise performance. experientially, the qualitative difference in image is negligible when compared to 2x SS. 3.0x PPDPO (3x Supersampling) Now at 3x SS we're in the realms of snake oil in terms of visual fidelity improvement, but as an experiment it's useful to see that by this point the 1080 simply can't deliver a stomach-able performance level even with ASW locked in place. The experience when playing with the 1080Ti however was manageable if clearly not ideal, it remained playable here even if I wouldn't recommend it. Showdown Epic Games' Unreal Engine powered Showdown is a brilliantly simple concept, throw the VR user in slow motion through a futuristic urban battleground, witnessing the chaos as a rampaging robot lets loose on a street full of armed police. It's a little long in the tooth at this point, but it was the star attraction at Oculus Connect 2015 where it was used as the centrepiece demo for Oculus' final publicly shown Oculus Rift prototype, the Crescent Bay. As the experience is on rails we ran through the experience for 60 seconds a time, for each run we increased the supersampling (aka Pixels Per Display Pixel Override) level between 1.0x, 2.0x, 2.5x and 3.0x via Oculus' debug tool. As you can see, these cards are having to push some serious pixels at 90FPS in order to achieve this. Below are the combined graphs for each run with the raw results numbers right underneath - they're a little tiny, so click to enlarge. 1.0x PPDPO (No Supersampling) 2.0x PPDPO (2x Supersampling) 2.5x PPDPO (2.5x Supersampling) Here at 2.5x SS, the 1080Ti takes it all in its stride and delivers a stable 90FPS throughout the experience. The 1080 however has buckled a little at this point with ASW engaged from the off. 3.0x PPDPO (3.0x Supersampling) At the ludicrous 3x SS setting, the 1080 again cannot take the heat, leaving the Ti to deliver an acceptable if far from ideal experience inside the headset with ASW engaged. Conclusion The GTX 1080 Ti provides a compelling price to performance prospect. Whilst Nvidia are still asking you to part with $700 for the privilege of thumbing your nose at Titan X owners, what you get as a standard gamer and as a VR enthusiast is the power to literally max the image quality capabilities of your chosen VR headset and a damn sight more. Existing owners of a GTX 980ti face the proposition of gaining up to and beyond a 100% uplift in standard gaming performance should they upgrade to the 1080ti. But the 1080ti also offers up tantalising and tangible experiential benefits for virtual reality enthusiasts. Yes our tests here are intentionally over the top, at anything over 1.8x supersampling levels, your experience will be limited by the VR headset panels and not the rendered image. But, the ability to push VR titles at these resolutions, achieving in many cases a fully-rendered (i.e. with no need for reprojection and ASW) 90FPS mean your investment will very likely see you through to the next generation of VR headsets and beyond. One thing we've not had time to cover here are the potential benefits gleaned from vendor specific VR architecture features introduced with the 10-series Pascal chipset. Features like Lens Matched Shading are still not found in many titles yet, with developers having to code specifically for those features in order for the application to benefit. We'll revisit this topic again in the future with any luck. The cynics out there will argue that the GTX 1080 Ti is an aggressive pre-emptive strike against the forthcoming AMD 'Vega' GPU architecture, which is rumoured to even the fight between red and green. The first AMD 'Vega' card may arrive as early as May this year, meaning it may be prudent to hang on for a few weeks until we see verified figures for those. That's not too long to wait. Disclosure: NVIDIA provided Road to VR with a GTX 1080 Ti GPU for testing purposes.