NowGamer: Epic Games and Nvidia show off Samaritan running on a single gfx card amid heightened Unreal Engine 4 speculation.
The saga of the legal battle that sees Epic Games fight Apple in the attempt to bring Fortnite back to iOS has just gained another chapter.
In an interview with Lex Fridman, Epic Games' Tim Sweeney shared the first details about the next version of Unreal Engine, Unreal Engine 6.
It’s going to come packed with a bunch of flashy, buzzword-filled features that no one will actually be able to use without tanking performance. And just like every iteration of that engine before it, the excuse won’t be that it’s poorly optimized, no, it’s "forward-thinking" and the hardware just isn’t ready to keep up.
But since it saves studios from having to invest in developing their own internal engines, it’ll still end up being widely adopted across the industry.
What cracks me up, is a lot of games utilize Unreal 5 and yet gaming has become more expensive. So all that BS that they shoveled out the last big reveal hasn’t translated into savings and if it has, then the industry is just plain ol’ lying.
You mean like “going digital will bring down costs for customer dramatically. Because there will be no packaging/distribution.” Or maybe the “games going forward, will be cross-buy so you buy it once and will be able to access it across all platforms you own.” Or even the “if we increase the rrp it will mean we can get rid of micro transactions altogether.”
… I could be here all day quoting the lies from this industry.
Scope of games are way bigger than even just 10 years ago. Also keep in mind that Epic charges 5% for using their engine, Steam charges 30% just like Sony, Nintendo and Microsoft's stores. That's 35% of your revenue gone. Steam infuriates me because they don't have nearly the overhead console manufacturers have but they know people are unlikely to migrate to Epic Games Store (which charges 15% instead but has a shit storefront compared to steam). I love UE5 (for the most part) and it has pushed the envelope in ways that would be too long to list here. I think UE6 will push things further and make it possible for devs who don't have Rockstar resources to make amazing games even further. Time will tell.
@1nsomniac Going digital did make things cheaper though. Games are dropping in prices at much faster rate than before and you’ll find plenty of sub $10 games on sale all the time, whereas before we had to wait for “greatest hits” label. Not to mention the indies basically have a levelled playing field now.
In the same way that you make all these assumptions and judgments on the future of UE, I see you making these assumptions and disregard any opinion you hold.
I see no value.
There is nothing constructive, just ire on what was and the willingness to believe nothing will get better.
You have given up on the possibility of joy and will not find it.
I find this odd. How am I expected to be excited with future promises when mired by the current legacy of UE5 and its myriad of technical shortcomings that have yet to be solved, even years after release.
Of course they should be working towards the future, but talking about it while UE5 still has many unsolved issues years after it has been the de facto standard? An engine used by so many, after so many years, with the backing of a company as grossly cash-rich as Epic shouldn't have so many problems still.
And the optics - even if not the truth of the matter - is you're putting time & resources into UE6 at the expense of UE5; your current product still needs quite a lot of attention. Unless the message is "we're abandoning UE5 because it's issues are systemic, and we hope UE6 can address that mess by moving on as quickly as possible".
I was attempting to reframe my comment as I watched more of the video, but the edit timed out. So here is a nearly completely different comment lol:
The number forks/fragmentations of UE5 feels like - from a laymen's perspective - a plausible explanation for why the engine, 3 years post release, has continued to have the same problems today as it did from day 1. Sounding as if they can't really find a way to cleanly coalesce each of the seven disparate variants, it seems hopes lie with being able to do so in the years leading up to the launch of UE6.
That said, if they have so many specific versions, then it does still kind of boggle the mind why issues, like compilation stutter, are still so pervasive. Seems in this specific scenario, the fragmentation could potentially be useful for at least helping to narrow down platform specific issues/solutions.
Clearly not the case, so hopefully they can make UE6 more unified to allow for more focused, streamline engine development.
Compilation stutter hasn't really been much of an issue for a couple years now if the devs know what they're doing. The problem is not all the devs know what they're doing in that regard. The real problem is traversal stutter. That is nearly universal.
I sort of ended up mentally putting both of those under the category of compilation stutter, which is surely too reductive. I should have just said "stuttering/fametime issues in all their incarnations". Because while there are improvements to comp stutter, even games that force you through long, even 30 minutes shader compilation stages before playing haven't managed to fully solve that issue. Heck, even consoles, with fixed hardware that can ship with pre-compiled shaders can't even seem to fully escape it.
Traversal stutter is definitely its own issue, though, and has only been exacerbated thanks to older cards being held onto longer, and companies - primarily Nvidia - opting to put 8GB VRAM buffers into cards for way the eff too long. If you don't have the top of the line CPU and high-end, overclocked RAM kits - most of the PC playing population - to help shuffle that info between system memory and the GPU, you're more screwed than most. And Nvidia could help the issue as well, if they could improve their years-long issue with high driver overhead. Freeing up any extra CPU usage, especially for those with weaker CPUs, would really benefit.
I really hope these things can have some kind of solution found for them sooner than later. As it is, it just feels like games are taking two steps forwards and two steps back a lot of the time. Improved pixel quality (world detail, lighting, etc), at the expense of degraded image clarity (softer image, heavy reliance on upscaling, increased artificing) and smoothness/performance (stuttering/poor frametimes).
And the fact this stuff occurs, when dev times are longer than they've ever been, with budgets creeping ever higher, it's that much worse to feel like a lot of experiences just aren't wins across the board. Especially as deep into this generation as we are, and with as much time as devs & engine makers have had to iron out issues. It feels like we may need to pump the brakes on the pace of research into graphics tech and rebalance towards optimization. Image clarity (native res, especially) continuing to fall further, with poor frametimes for a myriad of reasons, as the generation goes on doesn't feel the best.
I'm not just talking about shader compilation stages. There are games like Expedition 33 that barely have any pre-compilation stages (in the background on the main menu) and have almost zero comp stutter. It has to do with the way you use shaders and make your materials. It still has some small traversal stutter though.
It feels too soon to talk about UE6. It feels like UE5 barely got tapped, only a handfull of games really showed its potential.
NVIDIA’s RTX 50 “Blackwell” architecture has been a bit of a bore for us gamers. Apart from Multi Frame Generation, which has limited use-case scenarios, there isn’t much to be excited about. It is achieved using GPU-side Flip Metering. The optical field data is generated using AI models in the Tensor cores.
I reckon this is a pretty good approximation of what the first batch of next-gen consoles will be capable of, but, if this gen is anything to go by, there will be a lot of optimization to happen.
Can't wait to see what games look like in 2020.
I doubt next gen console with get Kepler GPU.
I guess we're talking GTX570/580 or even GTX680, which are expensive, too much for mass market consoles.
Maybe this kind of demo will run without AA and 30fps or something.
New consoles should let you upgrade your console. The way I see this happening though is allowing only SONY formatted graphic cards and Xbox graphics cards for each console. But then there is room for too expensive graphic cards and watered down graphics from the start. Hmmm
Sigh.......... I would luv Sony's PS4's GPU would be either
GTX640(GK106)$139 =GTX550Ti
GTX650(GK106) $179 =GTX560
1.5-2GB GDDR5 VRAM
8GB GDDR3 system RAM (C.Skill RipjawsX series)
Then Sony's next gen system should hold it's own well into the year 2021.