Tim Sweeney, Johan Andersson & John Carmack @ NV Montreal - Live Blog

03:05PM EDT - Q&A is over now, thank you all for reading!

03:04PM EDT - Sweeney: the next-gen of games will be developer centric, not retail centric

03:04PM EDT - Sweeney: it's a great time to be a game developer, distribution systems like steam (or just the web), not having to deal with the publisher or retailer you can do an enormous amount and build a company now like I did Epic back in 1991 without investors, this is a great time for this, the pace of innovation in the industry with mobile devices and oculus it's going to be a very interesting time to see all of this come together

03:03PM EDT - Carmack: all of us our building the future we want to see

03:03PM EDT - If there's anything in the game industry you'd hold your breath for a decade for, what would it be?

03:03PM EDT - Johan: I'll use both

03:03PM EDT - Carmack: I've got a lightboost monitor on my desktop, it'll probably move to G-Sync

03:02PM EDT - Sweeney: we primarily use NV hardware at Epic, and we buy most of it

03:02PM EDT - NV: we haven't precluded that

03:02PM EDT - Carmack: is G-Sync hardware licensable?

03:02PM EDT - Johan: I would play the game on both

03:01PM EDT - Is it thus fair to say that you'll be playing the games you're developing on NVIDIA hardware from this point on due to the enthusiasm for G-Sync?

03:01PM EDT - Ryan: Me too!

03:00PM EDT - I sort of want all life decisions to be handled by a panel consisting of Tim Sweeney, Johan Andersson and John Carmack

02:58PM EDT - Everyone seems to agree that Mantle alone isn't bad, if everyone builds their own it's bad, but hopefully it influences OpenGL/DX direction

02:57PM EDT - Sweeney is totally taking over the control of the panel now and asking everyone else questions this is awesome

02:57PM EDT - Johan: We see lots of stalls, performance gone missing, we want to learn how to program a machine on a really low level that has a good architecture for it

02:56PM EDT - Johan: They wouldn't have to pay us anything to use it, I've been pitching these ideas to any vendor for years

02:56PM EDT - Can you tell us how much they paid for you to use Mantle?

02:54PM EDT - He will magically know the answer at AMD's developer conference?

02:54PM EDT - Wow it sounds like Johan just played the NDA card on Sweeney??

02:54PM EDT - Johan: it's too early to quantify that, we'll have a lot more info in mid-November

02:54PM EDT - Sweeney: Johan, you've used the API, what's the cost in man-years and ultimate gain of using Mantle?

02:53PM EDT - Johan: if NV and Intel and Qualcomm did their own APIs, that would be a bad future, MS and Khronos play an important part in all of this, I hope Mantle will be quite influential in many of these aspects, what we're doing with AMD is just a start

02:53PM EDT - Johan: it's not a replacement, the idea is that we can solve some of the long term problems we have been having on the PC, those are things we can experiment and work with Mantle, this is another avenue, opening up something we're quite familiar with, we've spent the last few years working on next-gen consoles, even though we're not done with our work with it, I still see it as a success even right now because of these conversations and the amount of development and enthusiasm (and the opposite) - it's been a bit stale in the PC graphics space, MS switched focus for quite some time, but now going forward it's an exciting opportunity

02:51PM EDT - Sweeney: there's some good ideas in Mantle we really liked the idea of having low overhead access to the GPU, if you look back at DX/OGL, there's a lot of overhead in those APIs and the fact that they date back to the old SGI model of rendering which is very different than the current model, potentially unified memory, good ideas there I hope it really helps the OGL community and MS evolve their APIs. Would I like to have a low level API? Yes. Five of them? Absolutely not. To have yet another API on the PC that's still different than the DX API and OpenGL, and OpenGL ES and different than PS4 low level API, I don't think it's a good idea.

02:49PM EDT - Carmack: Mantle only became interesting because of their dual console wins, the landscape does matter that they have both major console wins with similar architectures, not a stupid thing that AMD is doing at this point, could have some implications for Steam, if MS and Sony embraced it that would be very powerful for AMD but it doesn't look like they're going to (at least MS), if I was still doing all of the major tech coding I probably would not be embracing Mantle right now but there would be days where it would be extremely tempting

02:48PM EDT - Carmack: My opinion is, in terms of action items, should NV have a response to Mantle? Unequivocally no, a horrible mistake if NV got panicked by this and made some lower level API, you already have good low level access through NV GPU extensions, give you as much performance as you want if you're fine with proprietary extensions

02:47PM EDT - What do you think of AMD's Mantle?

02:46PM EDT - Sweeney: when you get 20 TFLOPS in a single GPU, that's when you want to go 4K

02:46PM EDT - Sweeney: I see 4K as a workstation application for the next few years, we'll use those monitors for building games, but I don't think it's the right output resolution for rendering until you get significantly faster GPUs for rendering

02:46PM EDT - NV: it's a good thing that it doesn't have to be either or...

02:45PM EDT - Carmack: for a conventional desktop monitor, I want to see this absolutely ubiquitous

02:45PM EDT - Everyone answers G-Sync

02:45PM EDT - 4K vs. G-Sync what is more important to gaming?

02:45PM EDT - Johan is talking about the idea of having a GPU server in the home and use it to handle all rendering for all of your devices at home

02:42PM EDT - Carmack: another platform issue that's relevant to G-Sync, cloud gaming, I still think cloud gaming is going to be very important, it has immense convenience factors, G-Sync will be very useful for that since you already have enough latency and tear horribly

02:41PM EDT - Johan: PC has moved from grey box to something that looks like an Ultrabook, then you have lots of nice black boxes with graphics cards in it, not the same old boring boxes

02:40PM EDT - Carmack: MS or Sony would not approve an Oculus device plugged into consoles, no opportunity to do something off the beaten path like that, large scale macro economic trends in PCs, I really have no idea

02:39PM EDT - Carmack: it's not inconceivable that something like Steambox could come along and be the most cost effective way to deliver good computing for games

02:38PM EDT - Carmack: in many ways I look at the PC as the muscle car/hotrod of gaming, you got the people who just want their honda civic (they have a shiny phone or console), clearly the PC ecosystem has warts on it, the fact that you can put together that system there, you don't need to be a conglomerate with a research lab, you can just buy the cards and plug them all together that's pretty awesome

02:35PM EDT - Sweeney: other platforms come and go, smartphones/tablets are here forever but it's purely a content consumption tool, when we go to really participate deeply in complex games, or have a long experience or build a game it's always going to be on a computer of some sort

02:34PM EDT - Sweeney: the very fact that the whole thing works is critically important is going to be a big driver in the growth of the PC as a gaming platform

02:34PM EDT - Sweeney: when you need to do really serious work we always go back to our PCs, everything we do tends to gravitate around PC, it can be depressing as the PC market ebbs and flows, quality of drivers has really changed dramatically, things generally work - even the lowest end Intel integrated laptop you can go out and buy, it works

02:32PM EDT - Johan is quite positive

02:32PM EDT - Question about the future of the PC as a gaming platform?

02:30PM EDT - Now Carmack gets to use the bathroom

02:30PM EDT - Jen-Hsun is thanking the trio for everything they've done for the game industry

02:29PM EDT - Jen-Hsun has to catch a flight, thanking everyone now

02:29PM EDT - Oh wait, Jen-Hsun stopped him

02:29PM EDT - He just walked off the stage

02:29PM EDT - Carmack asked for a bathroom break

02:29PM EDT - Sweeney: memory bw seems to be in good balance as long as you're using your mobile device as it's built, once you start streaming to a TV you're going to want more, a lot of our techniques that we're using aren't translatable to a mobile platform, it's just too slow due to bandwidth limitations. bw also translates into power right now, die stacking is going to be very interesting

02:27PM EDT - Johan: A lot of vendors on mobile with varying graphics architectures, unclear what the best type of model for overall chipset, we may have to switch the pipeline quite significantly in hardware/software

02:27PM EDT - Carmack: in the old days it was exciting to hear Intel talk about getting things like FPUs, it's been a while since I've cared. one of those things I don't need to pay attention to since it's pretty adequate. I'm dearly looking forward to the day when we get to quantum nano tech that gets us 50GHz CPUs, but I'm not holding my breath.

02:26PM EDT - Is CPU power/memory bandwidth expanding quickly enough to accommodate a GPU like Kepler?

02:26PM EDT - Sweeney: important in these systems to look at how GPU is balanced with all of the other components of the system, improving the whole system level performance

02:25PM EDT - Sweeney: "they're losing it all in software"

02:25PM EDT - Sweeney: NV coming in with Kepler on mobile devices is going to be a really important thing for the whole industry, NVIDIA's unsung hero is the quality of its drivers, this is something that's really lacking in the rest of the industry, quality is there but performance overhead is severe, a lot of platforms we're losing a factor of 3 or 4 just in driver overhead. the only mobile platform we're able to run our next-gen mobile stuff on is NVIDIA's hardware

02:23PM EDT - Johan: we do a lot of experiments, one of the issues we're facing isn't technical, coming up with different game concepts because of differences in input paradigms and business models in mobile. difficult to break out and do these really small types of games that have interesting value there, we'll probably start seeing hybrid games

02:22PM EDT - Carmack: slight slowing in growth of GPU power in mobile, approaching thermal/power limits, still going to have a ton of additional progress, going to be really happy if NVIDIA can capture significant marketshare with Kepler on that. I'm working with a lot of the mobile vendors on GPU stuff, but NVIDIA has dev rel, tools and driver team that's better than pretty much everyone else in the industry put together, it'd be nice to get that in mobile GPUs

02:21PM EDT - Carmack: will we eventually end up where your only computation device is a mobile thing and some cloud assets, as far as seeing the latest Battlefield whatever on some mobile device, it's still a long ways off, you can do a UE4 demo on mobile but it's not like you just compiled for mobile on that

02:20PM EDT - Carmack: they're still not really there (console class mobile performance), but it's a tremendous slope they're on. lots of other challenges, if you look at NV Shield, CPU/GPU are pretty potent but if you play games on there it still drops frames/hitches, due in part to power management, G-Sync would be a wonderful thing to have on those Android platforms

02:19PM EDT - Carmack: I'm spending most of my time right now on an Android platform, I have a lot of real current experience, there's still a huge gap where the best mobile GPU now is still less than 1/10th of a pretty decent PC

02:18PM EDT - Next year we've got mobile Kepler, NV says it's more powerful graphically than the PS3, I'm curious when will it be viable to have games like Gears of War or Rage on a smartphone or tablet?

02:18PM EDT - Tim's take: we've seen this sort of problem throughout history, it's a form of artistic expression

02:17PM EDT - Johan's take: people complaining about violence in video games are getting older...the problem will go away eventually

02:17PM EDT - Carmack's take: I don't think there's a correlation, even if there was I don't think it's right to censor

02:16PM EDT - The trio are talking about violence in video games right now

02:13PM EDT - Tim Sweeney: I thought ZZT was this tradeoffs that couldn't be matched, then Minecraft comes along

02:12PM EDT - Tim: I grew up playing old computer games, there were some really interesting platform games - Commander Keen came along, when Doom came out I gave up on programming for a year - this is some unimaginable witchcraft

02:12PM EDT - Johan: massive inspiration from Unreal in Mirror's Edge, toolset overall was really inspiring and motivating for us to develop our own toolsets to focus on really great workflows for artists

02:11PM EDT - Johan: Both John and Tim have been massive inspiration, very early work on virtual texturing with Rage and idea of having entire worlds being textured that way, really inspiring to us

02:11PM EDT - Carmack: I'm proud to be a peer to the people in the companies that do these products

02:10PM EDT - Carmack: It was an issue at id software, so many years between releases, I very much did look at Gears of War, I would point out stuff to the artists. I'm very in awe of the high end titles that get made now, limited by fidelity of imagination, what artists bring to the games today I couldn't have imagined it

02:09PM EDT - A lot of people outside the industry think game developers are competitors, could each of you mention an influence from one or both of the other people on the panel from games they made?

02:08PM EDT - Carmack: Rage did suffer by not having those moments where we let the frame rate drop, but I felt like it was a slippery slope

02:08PM EDT - Carmack: it is a little unclear how you would make the cuts if you're still trying to be a 60Hz game on next-gen consoles, something like this would let lower end PCs play the same settings without falling apart like that, I'm not sure how much you'd throw in in terms of extra things, it'd be like optimizing for a niche platform at the beginning

02:07PM EDT - NVIDIA: pretty much game that ships already has that mode, they all have settings where you turn everything on and run it on a 4K monitor, it ain't 60Hz. It already exists, then you get into a firefight where it's not a 16ms frame, to some degree it already exists.

02:06PM EDT - Do you think that it's possible for current games that you're developing, because it's not going to be picked up right away, could you have a mode enabled where you enable G-Sync and blow us away with better environments?

02:05PM EDT - Johan: when you go into these retail stores selling tons and tons of TV monitors, I would like to remove showroom mode and instead have this variable refresh rate, false colors you don't want that

02:05PM EDT - Sweeney: it's got to be a major initiative here

02:05PM EDT - Sweeney: every display device, every platform (including mobile platforms), will need to adopt this technology over the next few years, it's really striking

02:04PM EDT - Sweeney: the rest of the industry needs to hardware up

02:04PM EDT - Johan: when people actually see/play it, there will be a lot of movement to transition to something like this (G-Sync) really quickly, then we can target games specifically for this

02:04PM EDT - Carmack: in the space of 5 years I'd hope that something like this is ubiquitous

02:03PM EDT - Carmack: I think this is going to be broadly picked up, it's not that expensive of a technology, I don't know NV's licensing costs, this is something that should be broadly adopted, this is just the right thing

02:03PM EDT - Question: when do you get to forget about a target frame rate in a game?

02:03PM EDT - 3-way Titan SLI rig attached to your face, it sounds nuts but it's inevitable

02:02PM EDT - Tony Tamasi (NV): 4K or 8K per eye, and you're going to want to walk around so it has to come in a mobile device, it's going to keep NVIDIA busy

02:01PM EDT - Carmack: we know where the better pixels go, it goes to movie rendering, we can spend another 6x graphics power and we get Avengers, there are other things we can do with that horsepower that might be a little more valuable

02:01PM EDT - Carmack: from my point of view, we're already past the knee of the cost-benefit curve of doing more advanced graphics, want to see more efforts in reducing latency

02:00PM EDT - Carmack: I think the pictures we're making today are very good, we can always make things better, personal belief we need to start pushing for higher frame rate/lower latency sort of things

02:00PM EDT - Johan: Getting to that performance, very few games are running at 100 fps, we want to get there with variable refresh rate, taking that leap forward in our games takes a lot more performance, we need to get that from the hardware vendors, OSes, how we design our games, etc...

01:59PM EDT - Tim Sweeney: This solves the monitor problem, future problem is in this stuff that John/Oculus that others are working on, that's what you need to go beyond what's possible on a monitor

01:59PM EDT - What is the next visual fluidity challenge now that tearing/lag are solved so to speak?

01:58PM EDT - Johan: initially skeptical about SteamOS, very positive to it, will take a long time to get it there

01:58PM EDT - Tim Sweeney: Valve is a great factor in keeping all of them honest, can't judge this over the next 6 months but instead over the next decade

01:57PM EDT - Tim (Cont): Will go a long way in steering the console manufacturers in pursuing an enlightened path

01:57PM EDT - Tim Sweeney: first reaction, building a mainstream around Linux is nuts, but month to month Linux is now a popular mainstream OS. With Valve it's not like MS/Sony, I think they are going to look at this as a very long term effort, build it up over time...there's a lot of fear from the major pubs/developers of being tied down to platforms that are ultimately controlled by MS/Sony or others, it's scary, we'd like to have direct relationships with our customers, possibility of SteamOS being a truly open platform...is very interesting

01:56PM EDT - Carmack: Valve approached id at the very beginning of Steam, asking about do we wanna put Doom 3 on Steam for launch - we said are you crazy? Clearly Valve has played a good strong long game and I'm afraid I may be at that same point right now, making your own console/OS - are you crazy? Maybe 10 years from now they'll look like brilliant prophets again. It still seems a little dicey to me (moving things to Linux), given their track record I'm a little hesitant...if it was another random company I'd be pseudo-scornful but since it's Valve I'm not

01:54PM EDT - Question for all 3: What do you think about Steam machines and Steam OS?

01:54PM EDT - Johan: We've done experiments, have pretty interesting game worlds (can imagine something like Mirror's Edge), rendering side we have a good trajectory, how does it effect our animation? Huge amounts of work to just do that for current gen consoles, but what about when you do that with VR? Takes a lot of time, not necessarily as part of the game but analyzing these things and getting people familiar, surprising amount of things that change within a game

01:53PM EDT - Question for Johan: How much sense does it make for a huge studio that makes mainstream games like Dice to get into VR now

01:52PM EDT - Carmack: something that's not immediately obvious, because all of this is driven by mobile displays we're tied to the mobile dev cycle, causes a little bit of distress internally since it'd be great to ship a new headset every year but that's a big commitment

01:52PM EDT - Tim Sweeney: If you look at what happened with iPhone over the past few generations, I think we'll see rapid evolution of VR in a similar way

01:52PM EDT - Tim Sweeney: Epic is looking at this entirely on the software side, we've had the opportunity to work with a lot of the early stuff, it's mind boggling, hardware needs another generation or two, software needs a lot of work, not ready for prime time yet, Infiltrator demo on Oculus Rift is amazing, we're really exciting about where that's going, will play a major role in the entire industry. Still in the research phases now and how it fits with our engines and games

01:51PM EDT - Question for Tim: Speak to VR at Epic

01:50PM EDT - Now Q&A time

01:50PM EDT - Sweeney: Carmack said something clever earlier, when you put all of these technologies together you don't have a product you have a superpower

01:50PM EDT - Sweeney: Ultimate trajectory is you have dual 4K or 8K displays in the form factor of Oakley sunglasses

01:49PM EDT - Sweeney: content creation challenges are severe, next-gen game budgets are going to $100M per game

01:49PM EDT - Johan: Procedural amplification of content, procedurally creating things like guard rails in a racing game are easy/obvious, but we want that across an entire world

01:48PM EDT - Carmack: Game content creation, does it become like fab companies where only a few companies in the world do it?

01:47PM EDT - I can see how we render these beautiful environments, but how do we build these worlds? It's not clear whose problem this is and how to solve it

01:47PM EDT - Johan: I have a good idea of what we can do visually over the next 4 years, one of the hardest challenge is not necessarily the visuals but we would really like to solve how we create our content and our game worlds, that's a massive issue

01:45PM EDT - We'll have 4K resolution tablets and head mounted displays, still will be able to buy xbone/PS4

01:44PM EDT - Carmack: we had this in an email thread at Oculus, 30 years ago state of the art was the movie Tron, what used to take half hour to render a frame we can render that in real time, better looking on a wristwatch today, A to A comparisons aren't the most persuasive there

01:43PM EDT - What does the world look like 5 - 10 years away?

01:43PM EDT - Similar nature of problem to what NVIDIA is tackling with G-Sync, reducing latency is very important

01:43PM EDT - "this is some of the coolest research happening in the world right now"

01:42PM EDT - Tim Sweeney: VR applications

01:42PM EDT - Asking the panelists about what's surprised them over the past few years

01:42PM EDT - Carmack: something valve did when they were working on one of their research head mounts, if we get to 120Hz that should be good enough, they were very positive that the persistence of the display was a very important thing, I was somewhat skeptical, but they had it setup there where you could turn it on/off, I didn't expect it to be that significant

01:40PM EDT - Carmack is very interested in what can be done around 90 - 120 fps

01:40PM EDT - 120Hz low persistence mode is another big advantage

01:40PM EDT - "A year ago I had no concept of this, but persistence is something huge"

01:40PM EDT - Tear lines in VR are horrible

01:38PM EDT - "It can be valuable to the user to sacrifice frame rate for some things"

01:38PM EDT - That was Carmack, huge endorsement

01:38PM EDT - G-Sync will "unequivocally make for better games"

01:37PM EDT - Carmack on G-Sync: I fought the crusade for 60 fps locked frame rate last gen, there are a lot of significant sacrifices that get made along the way (we can't have more creatures here, can't do this effect here), the game would've been better if we could say 90% were 60 fps and 10% would be awesome but at a lower frame rate, we couldn't do that before because going from 60 to 30 was a cliff you fall off of, if we had G-Sync we would be free to do that

01:36PM EDT - GPU makers and monitor makers need to work together to get all of these visual artifacts out of the entire pipeline

01:36PM EDT - You see tearing, jittery frame rates on consoles/PCs today, we've reached a point where technology allows us to move beyond that (visual glitches)

01:36PM EDT - Now Sweeney: people have come to expect more from these experiences, talking about how the iPhone changed everyone's expectations, you no longer expect graphical glitches

01:36PM EDT - G-Sync really enables game designers to have adaptive scenes/workloads

01:35PM EDT - Johan is first: games nowadays are giant, so many things we want to do with them, don't have control over everything, targeting a fixed frame rate is impossible

01:35PM EDT - The three folks just introduced themselves, now we're getting their opinions on G-Sync