Unreal Engine 5 = Game Changer

It's completely possible today.

https://www.newegg.com/sabrent-rocket-nvme-4-0-1tb/p/1Z4-00H3-00016

 

But at least your blank project iOS build is under 100MBs

 

Is this equivalent to the SSD in the PS5?

 

PS5 is 5.5 GB/sec. Sabrent's SSD is 5 GB/sec. System memory caching common files can help make up the difference.

 

To be honest, this thing to me doesn't look that different compared to Rise of Tomb Raider, because for whatever reason high fidelity does not have much impact on me, most of the time. Like all that extra detail doesn't even register.

Except GI.

The GI is cool and something I'd want.

The polygon cruncher tech, however, is more useful for larger studios with big budgets and high fidelity assets. It will probably take huge amount of disk space as well.

But like I said, GI is cool, and something I'd want. Unity is phasing out Enlighten, and are promising to replace it with a "robust solution" and I'm not finding a demo of said "robust solution" anywhere.

 

Man, that is a great value - ~$400 for 2TB. Might pick one of those up myself! I didn't realize they were already so cheap.

I'd need to go back and watch Cerny's talk, but I believe there was more to the PS5's SSD than the R/W speeds.

 

Compression was one of the features but anyone wanting to port a game to more than just PS5 won't likely build their game around it.

 

I don't think so.

I personally would prefer if Unity ditched C# completely and provided C++ framework, as cross compilation can be done through IR assembly produced by clang/llvm. However, this is a very extreme approach which will be liked by very few people aside from me.

So a second "closer to metal" C++ framework would be good.

I don't know why people insist on using C# for games. But at least this is not javascript, I guess.

 

I'd be willing to bet anyone making a multi-platform game in the near future won't be build their game around players having fast SSDs, either.

 

Main point is - we're freaking out right now, because we're comparing next-gen Unreal to current-gen Unity. We haven't seen next-gen Unity yet.

So maybe better to hold off trying to compare just yet. If the result is the same after this year's Unite, then this reaction would make more sense. But it's probably a bit premature today.

 

Yes and no. You mentioned that the PCIe 4 SSD was a good price at $400 for 2TB, but if you can live with a little less performance (3.5GB/sec instead of the 5GB/sec) you can go down to $200 for 2TB, $100 for 1TB, and $50 for 512GB.

 

I have to say, that UE5 demo was mighty impressive. I do however echo some of the thoughts here - I do not want to switch to C++ and the dreaded Blueprints system. I'm happy with Unity but would honestly love some of the visual fidelity of UE5 in the future in Unity HDRP.

 

scares me people saying Unity should give up the graphics race, hopefully this is not their mindset

 

When you watch the interview with the Epic bosses and Geoff Keighley, the Epic guys answer to "how did you achieve that" is always "because the new hardware allows us to".
For example at 12:30. Keighley asks "What is Nanite and why is it so great?" The Epic guy answers: "It is possible because we have a new generation of hardware, that really allows us to put loads and loads of trinagles at the screen"

 

Abandoning HDRP to focus everything on URP would have advantages.

 

Yeah we have. Just look at all the preview packages.

 

Nono, I mean we haven't seen them go out of stage and say "Look, here's what we can do on the PS5!".

Man...lot of people in this thread working hard to make themselves feel terrible.

 

Keep in mind every single time they've shown off a fantastic demo it was with the caveat that it required something that was outside the limits of what the vast majority of developers have access to.

 

I suspect that's more of a generalization. Many of their demos have involved internal code written just for the demo. I would bet the same in probably true of this Epic demo.

But let's say it is true. What's your point?

 

Kraken is a hardware version of Bink and can be used to decompress anything, effectively making the SSD throughput much faster than 5GB/S - basically it's a total game changer, will probably lead to regular SSDs adopting some kind of native hw decompressor? Not good for all files but for media? very...

edit:
http://www.radgametools.com/oodlecompressors.htm (software version)

 

B--b-b-uuuut I like my volumetric fog and lots of lights with shadows lol.

 

Epic Games creates demos showing off functionality that they intend to bring to developers. Unity creates demos showing off functionality that is not available to developers.

 

We just don't have enough data at say something like that with confidence. The main missing piece is that we don't know much about the code in this demo. Or previous Epic demos. It's also highly fallible to assume things about people's intentions. That's...not very reliable. I'm not sure the point here.

 

It's really not much of a generalization at all. A lot of the features in Unity demos never make it into the engine, in fact. It's not just a compatibility thing.

 

On the contrary we do have enough data to say something like that with confidence. We've had numerous past threads where we discussed the demos created by Unity, and how some of the functionality required was never made available outside of the team that developed the demo.

It's not "Look, anyone can do this with our engine". It's "Look, only we could do this with our engine".

 

nono, Ryiah had qualified the statement with "outside the limits of what the vast majority of developers have access to."

That's subjective. No way to actually know that without generalizing about developers, or without having the missing code to know how unobtainable it would be to write.

What we can say with confidence is "Unity never released most of that code." That's just a fact.

 

Unity typically discusses the technology they had to create, and it's very much code that can't be written without access to the source license. Source licenses for this engine are insanely expensive and simply out of the reach of the vast majority of developers.

 

If that's true then it is factually unobtainable.

It's not useful, however, in a comparison to Epic's demos. It's likely they write custom code for those that they don't release as well. We just don't know that information.

Why did you bring this up again? Honestly asking. I'm not...so good with understanding emotion or intent sometimes.

 

You made the point that we haven't seen anything from Unity concerning the PS5. I'm saying that it's completely irrelevant because anything they show won't be achievable by just anyone. Meanwhile Epic Games may show off a very incredible demo with very expensive work done on it but the features it uses will be available to everyone.

 

Now *that* is definitely subjective and unknowable. I think that might be doomsday-like thinking - jumping to conclusions because something bad and surprising happened. The outcome of today doesn't have to be bad. We just don't know yet.

 

You've only been here for a year. Some of us have been here for many years, and every demo has been the exact same outcome as the previous demos. We're shown features that we never get access to. Past experience says it will happen again and to date it has never been wrong.

 

In any case the take-away I want people to have is not the engine or the console, but the workflows.

This removes the idea of workflows. You just make. You add insane amounts of beautiful art and it's lit all the time perfectly with infinite bounce GI.


Why would I not want that for Unity too? And scalable from mobile (baked) to next gen console (dynamic).


I would like to be able to have great workflows in Unity instead of glancing at a neglected branch of hlod on github, while PLM falls back to grandpa CPU, and not dread porting from HDRP all the way down to URP.

Frankly, UE5 won this round, for me in terms of where things are at the moment because UE5 is removing work from the developer while Unity is adding it.

 

Lol, I have have been here for many years. This username is already a year old, huh? But when a user joined forums isn't really relevant in this case, I don't think.

 

We all need to switch over to Beeflang: https://www.beeflang.org/
Basically C# for games, unmanaged, performant, not as gnarly looking as C++ (and I say that as someone who learned programming with C++, but C# is definitely C++'s more attractive cousin.)

EDIT: /s for anyone who thought I was being serious.

 

Stage, you mean Unite? Those things where Unity shows a bunch of not ready stuff and the crowd goes “yeah, that’s not gonna be ready for at least another 5 years”.

 

Choosing a language with little to no real presence is never a good idea. UnityScript and Boo might have survived had they been JavaScript and Python, but instead they were no names that very quickly died to the sheer popularity of C#.

 

We don't know when the Epic tech will be ready either. I believe it's often common for tech demos to show off things that aren't ready immediately.

I believe someone posted the Unreal 4 tech demo that Epic was showing in 2009. It's a good reference point. Some of the key features in it were released much later, or in some cases, never. And it's a good thing they weren't, if they weren't ready.

 

We've been told when it will be available. Early 2021 for preview release. Late 2021 for full release.

https://www.unrealengine.com/en-US/blog/a-first-look-at-unreal-engine-5

No, because Epic didn't say anything about it being ready at the time they showed it off.

 

I think the quote there (unless I missed it) is:

"Unreal Engine 5 will be available in preview in early 2021, and in full release late in 2021."

 

Blacksmith demo had a lot of unique code changes way above average unity programmer's grade.

Adam demo was using, I think, either Cinemachine or Timeline long before it was available to the users.

Then there was that pyramid thing demo.
Right, Courtyard.

If my memory serves me right, in this case they Forked standard shader just for this demo specifically. Once again, way beyond normal programmer's abilities (Standard shaders is a fairly large codebase and hard to untangle).

So, in case of unity demos it is common for them to use one of a kind stuff, that is not technically part of the engine, but was made for the demo specifically, and will likely break half a year later.

No. Unless you really want to pay premium rates for the only Beef programmer in western hemisphere or something.
The language has to be popular enough so there's a pool of programmers available.
Beef will have its chance in quarter of century or now, I suppose, if it ever takes off the ground.

 

Correct. The technology we're seeing is Unreal Engine 5. Nanite and Lumen are among the major features of it.

 

So, we don't really know when these features will be ready. We know when it will be in preview. People here seem to be pretty down on the idea of preview features.

And we have a late 2021 window for a full release. Things that far away aren't always so reliable. They certainly haven't been with Unity, and I *believe* some Unreal releases in the past.

 

I'd say DOTS / HPC# is the sweet spot in terms of ideal game programming languages for me, so I don't really think we should be looking at new alternatives. It's basically C# (good human-readable syntax), but is a million times easier to manage your code complexity than anything object-oriented, and the job system is a joy to use.

Programming language & well-designed APIs (even though not always perfect) is definitely the main strength of Unity when compared to Unreal, I think

 

I'll believe it until they fail to make it. Unreal has been far more reliable with their releases than Unity.

 

Unreal is also free until 1million $ revenue per title.

 

This is what I'd be more worried about if I were Unity. This effectively makes Unreal free for Unity's target market.

 

I agree with this for it has really hit the mark

 

Just want to support this comment. I think it's on the money (though I think maybe it's tough for ECS to be appreciated by folks who don't use it regularly).

 

wait per title? I thought it was per entity...if per title is true...then I have very little faith that Unity can match it...

 

And Epic Game Services being free also makes all of Unity's recent acquisitions less good from a business standpoint. ChilliConnect, DeltaDNA, Multiplay, etc etc.... now unity will be hard pressed to charge for these services to recoup the cost of the acquisitions and development.

Also RIP Playfab. or should i say Pay-fab, am i right?