4.3.1 is out now.

Hey everyone, 4.3.1 is out now with a bunch of fixes, check the release notes http://unity3d.com/unity/whats-new/unity-4.3.1 and download it here http://unity3d.com/unity/download

As always please keep discussions within this thread, thanks

 

Awesome! Looks like a lot of important bug fixes, thanks for all the hard work Unity folks! Downloading now...

 

Thank you.

 

Awesome. Downloading soon.

 

Downloading...

Ay caramba! 1.0 GB??? I missed the notice on 4.3
Perhaps the example scenes could be pushed to the Asset Store?

 

The example is in the Asset Store, but not every uses the Asset Store. The extra size is all the 2D and other enhancements but given what it does it's not a big surprise. Maybe they can keep one download for an example 2D and 3D project on the main download page but you can be pretty sure the included example was used in the regression tests Unity runs before a release and I doubt they want to break that.

 

Great 4.3.1 gets a Cubemark of...

35,212 (4.3.1) - Cubemark 4.3.1 zip (PC)
11,820 (4.3) - Cubemark 4.3 zip (PC)
33,765 (4.2) - Cubemark 4.2 zip (PC)

So it's 104.285% faster than 4.2 on my PC.

 

Great!

I noticed: "Graphics: Fixed real-time shadows with orthographic camera and baked occlusion data."

However, I was having a problem with lights not using realtime shadows and a perspective camera (they were flickering). Were there any fixes related to that?

I think you mean it's 4.285% faster (104% would mean that it got a score of about 67 000).

 

Hmm I don't know if you're right because I'm using a percentage and if 100% is classed as the baseline, then any improvement could be classed as > 100%.

For instance how would you rate 4.3's performance vs 4.2's?

 

Version 2's performance is 105% of that of Version 1.

Version 2 is 5% faster than Version 1.

Those two statements are equivalent. If you start speaking in relative terms (e.g. 'faster', 'bigger', 'greater', 'slower') then the 100% is already assumed. If you say 105% faster you actually mean 205% of the original measure.

 

4.3.1 is 1.04 x faster than 4.2.

1.04 is 104% which is the same as saying it's 4% faster, IMHO.

So I would say 4.3.1 is 297.9% faster than 4.3 or 2.979 times faster.

 

Forget about percentages for a second.

Object A is traveling 5m/s faster than Object B.

Is Object A moving at

1) 5m/s or
2) B.Speed + 5m/s?

 

Who cares! Its faster YAY

 

+1

 

That's awesome to hear, thanks for running the tests!

I've been using it today on osx 10.82 and all problems I was having with 4.3 seem cleared up (monodevelop and asset store issues mostly). It has rock solid so far, very happy about this. Good work Unity.

 

hey unity happy holidays thanks for the news release!

i didnt find out until now, missed out on 4 days of unity 4.3.1 happiness

be nice if when i start unity it would tell me there is a new release automatically. i found out by accident when i tried to import a free asset that just got released and it refused to when i look at why it was because it needed 4.3.1 and i was only 4.3.0 so i went to the unity site and surprise new release!

 

Fixed that for you.

1.04x faster is 2.04x as fast. Since I don't think you meant it was 104% faster, and really meant just 4% faster, the above is correct now.

 

Well... it depends... are you riding a dead or living horse?

 

Is anyone having trouble with undo ceasing to function? I'm not sure if it's 4.3.1 related or of it's somehow project related.

 

4.3.2 before Christmas would be good.

 

My team and me are using Unity 4.2.1. We have not tried out 4.3 yet because of bad experiences of another team in the company: their project did not convert correctly from 4.2.2 to 4.3 and left it in an state of crashing Unity when opening the project. In addition the new monodevelop version had serious issues regarding responsiveness. They returned to 4.2.2 the same day.

Having said that is Unity 4.3.1 worth it to try it out? In its release notes I can find nothing about project conversion issues or monodevelop.

 

You can check for all the fixes here (there's a lot) http://unity3d.com/unity/whats-new/unity-4.3.1#fixes

Generally we advise against anyone upgrading in the middle of a project.

 

The release notes say it fixes the OSX crashes/hanging of the debugger, but it doesn't

 

So we've been stuck on 4.15 (as 4.2 introduced a lighting bug we couldn't work around) and then I held off on 4.3 due to reports of bad build performance.

I'm testing 4.31 now, but I'm seeing considerably worse build performance than we were getting in 4.15.

Is anyone else experiencing this or know any likely suspects or workarounds? We're running deferred with lots of skinned meshes on screen (it's when the enemies show up that the framerate takes a hit) not sure what else is important... Hmmm.

 

I reported this in 4.3.0 but still unfixed on 4.3.1

- On Android. Splash screen set to AutoRotation Landscanep. Run app in portrait pose, then rotate to landscape when splash screen is showing. Splash screen will screw up.

- The editor does not close properly often. Have to kill the process

 

I had that constantly prior to 4.3, but haven't had it since (4.3/4.3.1). Maybe I've just been lucky, but I assumed it had been fixed somehow.

 

One of these days I'm going to actually listen to that advice, and life is going to improve dramatically

 

This is the best advice I've ever heard.

 

I think its common sense!

 

Not when theirs some shiny new feature blinding your senses

 

Still stuck on 4.1x. Come on Unity, polish this stuff up.

 

How big of a performance drop?

 

Does the DX11 GPU Skinning build option help?

 

Quite a few projects goes over a year. Mine goes over two and a half years now, and still not finished. Just curious. Do you really advice me to stay with Unity 2.6 while we have Unity 4.31 now?

 

i think that's where the generally term comes in.

If you feel you have the time, expertise and money to upgrade your project from 2.6/3.x to 4.x then go for it. Just understand that it might not be possible and that certain things aren't compatible. In mid project cases, we look at things and work out whether it's worthwhile doing the upgrade. Would it actually add to the project and do the gains outweigh the risks.....usually not.

At my work I'm responsible for introducing new tech to the team and build machines. When it's integrated it's then my job to investigate upgrading our older projects to the new tech, just in case a client wants an update. Most of the time i have little problem. Usually it's third party plugins that haven't been updated that are the biggest issue.

 

I understand your point. But backwards compatibility, means an easy way to upgrade your project from at least the latest version without too much hassle, is a must have for a game engine. You cannot bring out a new engine every three months that breaks all compatibility.

As told, a game project can easily go across several years. That`s why i am not so green with the general statement to stay away from upgrades during a project. A major release may break some things. Like the nasty hierarchy thing between 3.x and 4.x that forced me to run through all my hierarchies and turn them back on by hand one by one. But a Point Upgrades should be upgrade compatible in all cases. Like upgrading from 4.2 to 4.3. When this breaks the project then there is something wrong. Which can be called a bug then.

 

upgrade or not to upgrade....keeping a backup of your project and give a try!
you will soon find out

 

apologies for an imprecise metric, but it was around 60fps to 45fps, I was just comparing builds with a FPS counter, so I don't have ms breakdowns. Pretty significant however!

DX11 and GPU skinning helped close the gap, but didn't get me back to where I was with 4.15 which worried me. Plus building the 4.15 version with DX11 also gives me a slight speed bump (widening the gap) but we've been avoiding it for compatibility reasons. Basically comparing like with like, 4.31 is running worse than 4.15. I was curious if this is consistent with anyone else's observations.


As for not upgrading, on projects involving many people or short time spans it makes sense to freeze your engine, especially if you have a tech team making extensive modifications, but for indie developers who are focused on building content, there's a lot less reason to stick to an old version of Unity. So many features get added, old features get depreciated, bugs get fixed, new platforms get added...

We are only using a few pieces of middle-ware, but when they get bugs fixed or have vital features added, they get added against the current version of Unity. Sometimes they try to maintain compatibility for as long as they can, but there's a practical limit to this.

And for those of us who are getting into self publishing on the new consoles, we have to keep up to date with Unity or we're going to have widely incompatible projects. Even 4.15 to 4.2+ is a pain, but imagine 3.5 to 4.2+!

 

Yes that is significant and worrying. Unity fixed at least one performance drop introduced in 4.3 and I was worried they may miss one. Maybe they did.

 

Not when we need the bug fixes. I'm not too concern about shiny new features. In fact, the "new" features like occulusion are hardy production ready. Maybe the recent improvement will be better.

 

Still need fix for iOS performance issues, already some strange postings in iOS section of forum.
Increased fps on development build ticked, "AssistiveTouch" on iPhone 5S, fps drops , ipad issues, issues with io7 and icons.
iOS7 memory and thread issues.

Please next release be a good mobile bug fix, don't care about new features.

 

If you haven't already be sure to use the Issue tracker(http://issuetracker.unity3d.com/) search and vote if it's an existing issue or create a new issue if you've found a new bug.

 

Completelly agree.

I think Unity have to forget adding new features for a while and fix all the bugs.