I agree, but implementing a new API is a pretty tough task and I don’t expect it to be perfect any time soon. Also they are trying to open up various systems which is awesome.
Don’t get me wrong, why some things take as long as they do baffle me. Hell someone made an AHR / Dynamic weather / Dynamic GI solution that works great cross platform for UE4 in half the time it took Unity to make a single SSR post effect. If they’d give me the source code I’m pretty sure I could speed things up exponentially too.
Point being, there’s many things to scratch your head about wondering why it’s taking so long. But this ain’t one of them…
I’m sure you’d be fine doing a substandard one that only worked on Windows, yes. Lets get real for a bit. People do post FX all the time. Do they do high performance or even good post FX? I’m not sure. I checked out this ocean thing in UE4. Ran like shit. Not interested in slideshows or hacks, but usable stuff I can use for real that don’t fold under many different conditions.
Low blow there Hippo, it was only partially my fault the game was so buggy.!
Thanks for the bitbucket link, I might start contributing to it and / or open sourcing shaders etc. in OpenGL and see if people can convert it into whatever Unity uses.
Edit: Point being if Unity need a hand in departments to speed things up, then I’m sure there’s many of us that would like to contribute. It’s just having something in place so we can…
P.S Although I’m happy with the direction Unity is going at the moment, things are guurrd!..
Excuse me hippocoder? That’s pushing it a little too far, I’m trying to have a real conversation about what I feel is a downward trend on Unity development cycles and your first reply to me was telling me that I have no understanding of programming or any actual development. I understand that (most likely) 80 to 90 percent of users of the Unity forums ( it being a “free” engine) may not have any actual development experience with large teams, but I’m not one of them.
This isn’t even a programming issue, like I’ve pointed out, it’s a project management issue.
Also, you’re telling me that 500 people are working on the engine? I may call a word starting with bull with that reply. Actual engine programmers are not in that range, so throwing around that number is a tad disgenious. And they’re all rockstars/top of the line programmers? I think other companies might attract the top .1% of the world’s programmers. Not that I’m saying that the staff at Unity is not up to par and not doing a great job, but to say that Unity has 500 of the world’s utter best working on the engine is stretching reality. Honestly. I’m not writing this up because I hate Unity. I’ve been using it for more hours than I could ever count in the last three years, and while yes, it’s my development engine of choice, there are troubling concerns that I’m voicing.
Telling me that everyone is wrong and I’m right is probably the most childish comment you could have mustered. Bravo.
(and I know you are a moderator on the forums, and this thread will probably be locked soon after I post my reply, but don’t treat me like a child.)
On that point, it’s not like “implement DX12” was not something that was PLANNED for a year ago or more. When things are planned, things go according to plan. The point is i do not even care about DX12. But it’s yet another point that goes into the “this may turn badly for all of us” column of why there is a troubling trend. I’m not the only developer that wrote about it on the forums provided by Unity, by far.
It’s still the same as when I looked at it before (like two weeks ago I think).
This feature is still on the roadmap for 5.3:
5.3: Graphics: Multi-threaded Rendering Phase 1
Although it does look like the dx12 part that will greatly improve performance is in 5.4. I agree though, 3-4 months is a long wait. Hopefully it will be worth the wait. I was going to implement it myself, but then I remembered, I’m not at that level yet. I’m chomping at the bit to get that yummy extra performance tho. Simply can’t wait.
Even though Windows 10 didn’t come out until end of July, it seems like it shouldn’t take them a year since Unity 5 was released to implement it. But in my case, beggars can’t be choosers. But I would like to see them get it implemented as soon as humanly possible.
Well frankly, claiming that throwing more programmers at it, does indicate that you don’t really know much about programming. This isn’t trying to hurt you or insult you, but an observation from me that is incorrect in your eyes. I don’t see the problem myself.
I’m not trying to upset you, merely draw attention to the fact that throwing more programmers at it won’t fix it in the short or long term - period. All that will occur is you get a huge backlog of people who haven’t got a clue how the Unity engine works (this eliminates short term), then finding something for them to do (which eliminates long term).
If a hire is to be done at this point, it needs to be very smart people not quantity of people. And very smart people are quite hard to find
As for locking threads, something bad should happen for a moderator to lock it. So far that hasn’t happened and you (or anyone else) shouldn’t pull punches. When I get involved in a topic, I am a user, just like you. With my own (right or wrong) opinion.
Then all I can suggest is jump ship, it put me back about a year but Unity wasn’t the correct tool for me to use. You live and learn…
It might give you some perspective, I used Unity properly again for the first time in a long time. It was so simple I was in utter bliss. Still have to use UE because of a custom framework bolted onto it and a couple of other reasons, but they are both great IMO. Depends on what you’re doing…
Sort of related question; Is there any point in promoting DX12 over Vulkan except for DX12 being ‘available’ right now? Afaik it’s supported (atleast the base of it) by all DX11 hardware, and it won’t be limited to windows 10. And the basic change from openGL → Vulkan seems close to the change of DX11 → DX12 as in; lower level, more threadable.
Disclaimer: never worked with openGL or directX directly.
DX12 benefits WP/Xbox users, possibly better performance on Windows. They should have almost identical performance on Windows, though we shall see how much AMD/Nvidia/Intel prioritize DX12 since most shops will go that route.
In developing software we’re often solving new problems. The only way to know how long it will take to solve a new problem is to time yourself actually solving it. Anything else can only be a guess, because the world is yet to supply software developers with (working) crystal balls. In large projects, you have guesses that rely on other guesses. Only one of them has to fail to accurately predict the future to effect the whole project’s time line. Effective management (more specifically, risk mitigation) can help a lot with that, but you can never completely eliminate it.
Here’s the thing, though: people will never be happy. “Everyone” (figuratively speaking) got upset about the GUI being pushed back. So Unity clammed up and stopped talking about delivery dates for anything. Then “everyone” (again, figuratively) got upset about that. So they implemented a public road map and now many people are upset when things get shifted around on that, even though they’re making it clear that it’s sharing plans for “possible” releases as opposed to concrete ones.
(I admit they could make it even more super duper abundantly clear. If it were me I think I’d be partial to a giant, blinking, red banner with the words “These dates represent estimates only. Do not use them for planning.”)
hi there, just jumping on board… I have some questions :
what is a directx12 video card? so you really need the latest nvidia stuff to see better performances ? or older but powerful cards will do ??
now, also the increase of performance is only for windows based platforms right?, like windows phone, win 10… so If I compile for android… then ?
EDIT, PS: I’ve seen the roadmap and there are so many interesting features that I wish to see ! I wish they could be already implemented
I must admit I had a similar experience when I notice it slipped.
Went from Yay DX12 for Xmas, to DX12 in time for Easter!
But think about it this has triggered a fundamental overhaul of the rendering sub-system in Unity, so hopefully we should all, regardless of platform and API, see some benefits.
Mobile platforms with Vulkan and Metal as well as DX12 platforms.
Medium to long term a cross-platform thing like Unity will need to support both, providing Vulkan isn’t a damp squib.
A lot of the underlying changes they are making to Unity will benefit both in theory - just because they are doing it for DX12 first doesn’t mean many of the changes (especially multi-threaded rendering) won’t also be useful for Vulkan. Note also the extensive changes they’ve been making to OpenGL in Unity, which in theory also gives them a future foundation for Vulkan support.
As for why Vulkan isn’t top priority, I would partly blame Apple for that. Because Apple have gone for their own Metal, its not clear that the OS X platform will get next generation OpenGL one day, as opposed to Apples own proprietary stuff. In the meantime desktop Metal support is going to at least bring the Mac up to somewhere around DX11-level feature set, with some of the performance improvements that people associate with DX12.
If Vulkan succeeds then I assume Unity will by then have a much better foundation to build support for it, but I don’t assume this success is whats actually going to happen. Especially where Windows is concerned - I would not assume that in practice Vulkan is going to be a popular route for those who want to remain on Windows 7. Resistance to Windows 10 will somewhat diminish over time, and depending on how well graphics card makers support Vulkan in their windows 7 drivers, it just might not pan out that way.
Also apple has longer term plans for it’s propriety metal. They can control closely what gets to be a native feature. For example PVR ray tracing is cheap and embedded, and metal can support that on an official basis if they so desire.
Also things like mac pro, and compute, they can figure advantages while keeping costs sane.
