Update to Unity 5 and Multithread?

[Dark_Ansem]

They could always switch to the Unity 5 beta...

[Elerond]

They could always switch to the Unity 5 beta...

 

That is something that one should never do. Because there is reason why Unity's developers call their product beta, instead of released version. In commercial products it is always wiser use stable versions than those under testing and fixing.

 

Beta versions are good if you want test new features and get familiar with them, but they are always too unpredictable to use as platform for big commercial products especially when there is stable version on hand that offers most of the features and capacities that said beta version offers.

[Luckmann]

it's not a fair comparison: times have changed, and Obsidian is more competent than the other developers.

 

Also, not all engines are equal. My (admittedly limited) understanding is that Unity is made to be fairly upgrade-friendly, not breaking old functionality and so on.

[anameforobsidian]

If the problem with the Linux build is resource loading, and the new version of Unity will fix that, that doesn't sound entirely trivial.  But, I'm not a professional and they are.

Edited January 10, 2015 by anameforobsidian

[Sanquiz]

It will be sad if this game sit on one core, this will limit the game.

 

Divine Original sin have this problem, the game sit on 1 core, right now they are updating the engine to support multithreading.

 

 

I just check:

Wasteland 2 use all my quadcore. All cores working at the same time, thats means POE is going to have no problems, also the game is 32 Bits, max use of memory is 4 GB, windows show thats is using like 1-2 GB.

Edited January 17, 2015 by Arturo Sanquiz

[Dark_Ansem]

It will be sad if this game sit on one core, this will limit the game.

 

Divine Original sin have this problem, the game sit on 1 core, right now they are updating the engine to support multithreading.

 

 

I just check:

Wasteland 2 use all my quadcore. All cores working at the same time, thats means POE is going to have no problems, also the game is 32 Bits, max use of memory is 4 GB, windows show thats is using like 1-2 GB.

 

aaargh don't say that it's SOOO bad luck!

[shmerl]

Wasteland 2 is being proted to Unity 5 by inXile (and will come out this summer). I wonder if Obsidian is interested in doing a similar thing for Pillars of Eternity. Understandably it's not a minor task, but everyone can benefit from better performance.

[Dark_Ansem]

Wasteland 2 is being proted to Unity 5 by inXile (and will come out this summer). I wonder if Obsidian is interested in doing a similar thing for Pillars of Eternity. Understandably it's not a minor task, but everyone can benefit from better performance.

 

I hope so

[Sabin Stargem]

I figure that it would be an ideal step for releasing a GOTY or complete edition of POE.

[bonarbill]

Wasteland 2 is being proted to Unity 5 by inXile (and will come out this summer). I wonder if Obsidian is interested in doing a similar thing for Pillars of Eternity. Understandably it's not a minor task, but everyone can benefit from better performance.

 

Same.  I know people are going to say it's "a lot of work," but the differences in performance will benefit everybody,e specially for us people who are gaming on Linux where unity performance has it problems.

Edited May 8, 2015 by bonarbill

[Dark_Ansem]

I figure that it would be an ideal step for releasing a GOTY or complete edition of POE.

 

But not forgetting current users\backers as well!

[Elerond]

When I read problems that Red Thread Games have had with porting Dreamfall Chapters: The Longest Journey to Unity 5 (because its bugs and because it does somethings differently), this topic come back to my mind and fact that Obsidian decided not to upgrade PoE, at least yet to Unity 5.

 

 

Unity 5 has introduced a huge number of new bugs to the game and to our production pipeline, including, but not limited to:

• Missing and faulty lighting and shadows. (The lighting system has also seen fundamental changes. This will eventually make things look better, but it’s not working as intended yet.)
• Missing and faulty collision. (Due to fundamental changes to the collision system, Propast pavements have become insurmountable obstacles, while walls often offer no resistance whatsoever.)
• Random game logic hiccups. (We're testing and trying to figure out what's broken.)
• Random crashes on load. (We're trying to figure this one out as well.)
• Crazy, exploding scenery!
• (And many, many other issues — too numerous to get into here.)

We’ve spent a lot of time dealing with these fundamental engine changes, updating scripts, lighting and shaders, finding workarounds to editor and build bugs, testing and re-testing all previous episodes — along with wrapping up Book Four.

 

This situation has become unsustainable, so today I made the difficult decision to temporarily halt work on Book Four in order to focus completely on finishing the port.

The reason for this is simple: without the first three episodes and core game engine working with Unity 5.2, we can't release episode four. Our episodes don’t stand alone: they’re part of one growing and evolving game, sharing assets and code.

 

Switching engine versions is always risky, but our tests back in August were positive. We remained cautiously optimistic about our deadlines through September. Unfortunately, as of today, Unity 5.2.2 remains buggy and the expected fixes have not arrived, leading to our current predicament. Since we can’t wait for Unity to release version 5.3, we have to muddle through with what we currently have.

 

https://www.kickstarter.com/projects/redthread/dreamfall-chapters-the-longest-journey/posts/1390566

[Dark_Ansem]

but they still did, right? the benefits outweighed the bugs.

PoE uses such a heavily modified engine, anyway...

Edited October 22, 2015 by Dark_Ansem

[Elerond]

but they still did, right? the benefits outweighed the bugs.

PoE uses such a heavily modified engine, anyway...

 

They still try to port, if you read that update you see that it is about them delaying episode 4 in future because they need to put all their resource in porting previous episodes and core game to unity 5. So we can only hope that benefits from the porting will be enough to justify work they put in the porting or they just made very poor decision that cost them quite lot money.

Edited October 22, 2015 by Elerond

[Aurelio]

Read the earlier update on Kickstarter about Dreamfall Chapters: The Longest Journey

 

So there are reasons to do the port but it takes time. I guess Obsidan would also need 6 months - 1 year to do just the port and I'm not sure if all benefits would be the same. Maybe it makes more sense to do the switch, when they create a PoE 2 game or what ever the name will be.

 

Perhaps they could upgrade PoE to the latest 4.6.x version. The latest PoE update is still on the patch level of mid may. On the other hand most Unity3D fixes seem more related to mobile devices.

 

https://unity3d.com/unity/qa/patch-releases

 

----

The reason for this is our long-gestating port of the game from Unity 4.6 to Unity 5.x (we're currently using version 5.2).

This upcoming port is important for several reasons:

• To improve frame-rate and overall performance.
• To reduce memory-related crashes.
• To improve our internal workflow with Unity's new 64-bit editor.¹
• To open up for ongoing and future improvements to lighting, shaders and effects, and further performance and stability fixes.
• To facilitate for ports to new platforms, including PlayStation 4.²

Although switching engines mid-production is a risky thing to do, during our initial investigation into the engine switch we were quite optimistic about the timeline. The game appeared to run well in Unity 5 and there were no massive issues with our own tools or third-party plugins. In addition, this felt like the perfect time to switch and move the production over to an improved pipeline before wrapping up the final two episodes.

 

Doing the actual port has proved more time-consuming and challenging than expected, however — and of course this is the case, because it always is. Still, we did not anticipate this work to delay the release of Book Four.

 

Why a delay? Due to the way our game is constructed, we've had to port everything. We couldn't just start with the fourth episode and patch the rest later: the whole game needs to run in the new engine. This has resulted in a ton of work for the artists, along with plenty of bug-testing and -fixing. Things like collision, lighting, plugins, shaders work differently — and with every major engine upgrade, there are new bugs, some of which we can't fix on our own.

 

Will the port be worth the wait? We believe so! This was a necessary switch, something we had to do sooner or later, both for obvious performance reasons (we're very aware of the issues some players have had) and to allow porting to consoles and other platforms. How much a performance boost players will see depends, of course, on their specific hardware and OS version. Hopefully it will make a significant difference, particularly to those of you who have struggled running the game.

 

¹ Working on locations such as Europolis and Marcuria using the 32-bit editor has been an exercise in frustration. Our artists and designers experienced regular crashes and extremely slow performance. The move to Unity 5 has massively improved our workflow.

² A lot of you have been waiting for the PS4 port, and this is not something that was possible with v4.x of Unity. The game is currently running well on our PS4 dev kits, though we don't yet have a concrete release date in mind. We're prioritising Book Four and PC/Mac/Linux first — but PS4 is next in line!

[sibakruom]

InXile first announced they were working on their port of Wasteland 2 in February, so 6 months seems like a bare minimum. Frankly, unless Obsidian decide to make it coincide with a console version as well, I don't see an upgrade to Unity 5 happening for PoE. There's simply no way it can be a worthy investment of time and money.