Fusion 2.5 Microstuttering (moved from Fusion 3 status thread)

Quote from Julian82

Thanks Danny, don't take my comment as an attack in any way. I just honestly thought looking into this from a technical perspective was not your part of the business. Cheers

No offence taken Julian Never underestimate my role at Clickteam. We are all a big unit when it comes to ensuring the runtime is optimized and smooth and believe me when I say this, Yves is the biggest advocate of making time to tweak the performance of the runtime. Yves likes a challenge and likes to do whatever he can to ensure it's optimum and it's running as smooth as possible.

Simon has a better handle on this than me as he has been much more proactive with you guys in this thread and he's also an advocate of ensuring the runtime runs as smooth and as fast as possible. It is a big topic here at CT but I can only reiterate what Simon has said thus far.

Quote from Danny

No offence taken Julian Never underestimate my role at Clickteam. We are all a big unit when it comes to ensuring the runtime is optimized and smooth and believe me when I say this, Yves is the biggest advocate of making time to tweak the performance of the runtime. Yves likes a challenge and likes to do whatever he can to ensure it's optimum and it's running as smooth as possible.

Simon has a better handle on this than me as he has been much more proactive with you guys in this thread and he's also an advocate of ensuring the runtime runs as smooth and as fast as possible. It is a big topic here at CT but I can only reiterate what Simon has said thus far.

Thanks, than there is nothing more to demand at this point:) As I said, nobody demands miracles, we just wanted a confirmation that (reasonable) effort will be put into this.

Quote from Simon

not least of all that the affected users are very few in number out of the hundreds of thousands of Fusion users.

just jumping in to disagree with this point XD

i think all users are affected by this, but most, either... can't see it, don't care about it, or can see it but don't bother to voice there opinion.

if the test julian did was done by every fusion user then i'm sure we'd see its a very large percentage of users experiencing it.

& even if they take no action themselves, wen they release their games, there's a good chance the public will let them know about it, as we all know the internet is not so forgiving lol >:)

anyways do what u can guys i just hav to jump on any time i see anyone try & state that it's not really a thing XD

With respect, if users can't see it, they're NOT affected by it. Similarly if they don't care, whilst they may see it then it's not an issue for them. If they don't bother to voice their opinion, especially in light of a very prominent thread about the issue, then chances are it's not really a significant issue to them on the whole.

I'm not saying there isn't an issue there, just that it affects only a tiny fraction of our user base that we know of and as such, we have to weigh up how much priority and resources we dedicate to this.

I'd actually say you're right, Simon. Probably most Fusion users don't care about the issue. But the few professional ones active in this forum do for sure and at least I hope we made our points pretty clear. It's for sure your choice as a company if you still want F2.5 to compete as a serious gamedev tool or if you put all effort into F3 now. Anyway, if you still want bigger commercial games being made in your existing engine, then you should embrace involved users like us a little more, even if we're few and demand a lot. We don't do this to bother you, we do it for good reason. The quality bar has risen drastically in independent game development in the last years, so what Shin probably wanted to stress is that it would be totally naive to believe that your player base would not be bothered by an issue like Microstuttering once they actually paid for your game. That subjective sensitivity to Microstuttering is IMO pretty dangerous, as it's an issue a developer may not be able to spot himself before he gets it served cold in negative reviews.

I really cannot understand how you guys from clickteam can think of not fixing this issue, its the most core/fundamental thing to have a smooth working engine.

So if a core issue like the microstuttering is not on your priority list, please tell us Clickteam -> what are your prioritys than?

Quote from Simon

With respect, if users can't see it, they're NOT affected by it. Similarly if they don't care, whilst they may see it then it's not an issue for them. If they don't bother to voice their opinion, especially in light of a very prominent thread about the issue, then chances are it's not really a significant issue to them on the whole.

I'm not saying there isn't an issue there, just that it affects only a tiny fraction of our user base that we know of and as such, we have to weigh up how much priority and resources we dedicate to this.

fair enough, wat u say is true. (& yes julian worded it much better than i cud lol)

i'm just saying that wen u put out a game to the mass public, there will be many people like us in this thread out there.

for example, say the next big mario or fifa game launched with this same micro-stuttering in, the internet would go mad & wud tear it to shreds! XD

( & if they then if it came out that they were programmed in clickteam fusion 2.5 .......gulp! GAME OVER XD )

anyways, i kno u guys all work super hard & i hav full faith in all the clickteam staff. do wat u think is best

The point is we do have a smooth and stable engine, but a handful of users are having issues which remain very subjective. I for one am unable to recreate the issues and I have tried on quite a range of machines, as have other members of staff.

However, you are putting words in my mouth and totally overreacting! I didn't say we're not attempting to deal with it, just that if at this late stage in Fusion 2.5's lifespan an issue which affects a minute percentage of our user base is going to require massive time and resource cost, that we have to balance how we deal with it. And people wonder why we sometimes don't like to converse with users! Sheesh!

Yves very clearly said:

Please login to see this attachment.

I don't know how we could be any clearer!!!

I ran an empty Fusion application with v-sync on, and alternated between Display Modes and got vastly different frame time outputs.

Please login to see this picture.

Can somebody confirm on their end? Did you already investigate this, [MENTION=12521]Julian82[/MENTION] ?

[MENTION=5383]chrilley[/MENTION]: Nope, but that DirectX8 output looks like it should be, very interesting, I’ll check today on my setup. Cheers

[MENTION=5383]chrilley[/MENTION]: Now look at this. Not one frame misses the call on Direct X8. You've made a very important finding I think. The problem seems to be related to the Direct X9 runtime only.

Please login to see this picture.

edit: And this is my game on Direct X8. I think this clearly confirms that Microstuttering is a Direct X9 runtime issue. Unfortunately I can't use Direct X8 as essential shaders don't seem to work and I get a lot of strange issues with texture clipping also.

Please login to see this picture.

Thanks Simon and Danny for the replies.

I understand where you are coming from, and it's good to hear that Clickteam are looking into it as much as they can.

It seems to be less subjective than you claim though, when quite a few people are posting these graphs in an attempt to show objective evidence of the problem.

I haven't followed this thread closely, even though I've been always affected by micro-stuttering.

However, I've noticed you guys are using FRAPS Bench Viewer (or that's what it says)? I'll give it a go today and post results here at some point.
I'll try some other games as well to see how it works out for me and if it's a reliable way of measuring stuff.

Maybe it has been already asked that people that are not having this issue; Simon, Yves and everyone else to do the same and post their results here?
Would be great to see no issues for you guys, and maybe then we could start comparing our set-ups? Maybe it's something super simple in the end...!

I'd honestly love to get this solved, I'm not sure if we're just super unlucky few having this, or if there's more to it...

EDIT:
Empty Fusion 2.5 frame, no debugger and vsync on. 60fps.
Please login to see this attachment.
i5-4670K 3.40GHz - GTX 780 - Windows 10 64bit

Hello [MENTION=5109]Simon[/MENTION],

i am sorry if my comment was a bit to harsh for you.

But you have to consider my (the users) point of view, because when i research, the stuttering issue was reported long time (years) ago but never fixed - and thats the point for me. Now all that comes back to you, because its such a big problem for the users that you should never have ignored that for so long!

Than when your new software is around you say things like "Fusion 2.5 late stage of lifespan - stuttering may not be fixed for numerous reasons...". I mean how should people feel now when you say something like that? You dont fix that issue for years and than you put out a comment like that, .... really? (That was also the reason for my previous post.)

I have invested lot of my money, around 900 Euros (Fusion 2.5 + all Exporters) into you Software, thats not cheap, for that price you actually could expect that things getting fixed properly, regardless if there is a new version of Fusion being released or not.

Although I find it a good approach to get objective measurements of this issue, I would be cautious about using only one measurement tool. Fraps is installing its own function hooks. The hooks are simply JMP operations assembled ad-hoc at the beginning of Direct3D API functions. It may have a small or big influence on the results, especially if you compare with other games rendering with different APi like OpenGL.
Fraps is certainly a good tool but it may be not so suitable for some of the configuration comparisons.

I am not saying that there is no isssue though. Just that it is not so obvious and involving other parameters than Fusion rendering.
Debugging such issues can be really time-consuming and may involve changing third party drivers like rendering dlls.

In any case, I think taht Mathias' approach is worth the try.

Quote from conceptgame

Although I find it a good approach to get objective measurements of this issue, I would be cautious about using only one measurement tool. Fraps is installing its own function hooks. The hooks are simply JMP operations assembled ad-hoc at the beginning of Direct3D API functions. It may have a small or big influence on the results, especially if you compare with other games rendering with different APi like OpenGL.
Fraps is certainly a good tool but it may be not so suitable for some of the configuration comparisons.

I am not saying that there is no isssue though. Just that it is not so obvious and involving other parameters than Fusion rendering.
Debugging such issues can be really time-consuming and may involve changing third party drivers like rendering dlls.

In any case, I think taht Mathias' approach is worth the try.

Good points, especially as Chrilley's finding suggests that this problem may very well be hooked to the Direct X9 runtime. X9 is pretty outdated now, my system is on 11, for example, and who knows if X9 is still works well with modern setups. Googeling the topic in genral a lot of threads pop up where people had updated their Direct X and older games started to stutter. Also, it seems GM2 (which runs good), is using both recent Direct X and Open GL versions, depending on the runtime.

Quote from Manuel

Hello @Please login to see this link. ,

i am sorry if my comment was a bit to harsh for you.

But you have to consider my (the users) point of view, because when i research, the stuttering issue was reported long time (years) ago but never fixed - and thats the point for me. Now all that comes back to you, because its such a big problem for the users that you should never have ignored that for so long!

Than when your new software is around you say things like "Fusion 2.5 late stage of lifespan - stuttering may not be fixed for numerous reasons...". I mean how should people feel now when you say something like that? You dont fix that issue for years and than you put out a comment like that, .... really? (That was also the reason for my previous post.)

I have invested lot of my money, around 900 Euros (Fusion 2.5 + all Exporters) into you Software, thats not cheap, for that price you actually could expect that things getting fixed properly, regardless if there is a new version of Fusion being released or not.

Manuel, we didn't ignore it. Yves has already clearly stated that he looked at the code previously and could not find anything in there which could be causing the problem. He is looking into it again, but there is no guarantee we can find anything this time either. There's no certainty that it's actually even Fusion's fault - there are plenty of examples of games developed outside of Fusion which exhibit the same problem too. This is still a marginal issue, whether or not you wish to hear that. As I have previously said I've tested this on the following machines:

[TABLE='class: grid, width: 100%']
[tr][/tr]

[tr]
[td][/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]CPU[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]GPU[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]RAM[/td][/tr]

[tr]
[/tr]

[tr][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Desktop PC[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Intel Core i7-6700L (4 cores, 8 threads) @ 4 GHz[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]GeForce GTX 980 with 4GB GDDR5[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]32GB DDR4[/td][/tr]

[tr]
[/tr]

[tr][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Mac Mini 2011[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Intel i5-2410 (2 cores, 4 threads) @ 2.3 GHz[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Intel HD 3000 with 288MB shared RAM[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]2GB DDR3[/td][/tr]

[tr]
[/tr]

[tr][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Acer Aspire 5750G Laptop[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Intel Core i5-2410M (2 cores, 4 threads) @ 2.3 GHz[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Intel HD 3000 + NVIDIA GeForce GT 540M - 2 GB[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]4GB DDR3[/td][/tr]

[tr]
[/tr]

[tr][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Linx 10 Tablet[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Intel Atom Processor Z3735F (4 cores, 4 threads) @ 1.33 GHz (1.83 GHz turbo)[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Intel HD for Atom Z3700[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]2GB LPDDR3[/td][/tr]

[tr]
[/tr]

[/TABLE]

I point blank cannot reproduce the microstuttering on any of these machines and they represent a very varied mix of specs. I've tried on Windows 7 and 10 and even done tests with CPU load simulation running in the background. I've tried on monitors, tablet displays, laptop display and even a DLP projector! I just cannot get a result which matches the experience some people are reporting. Danny similarly has done various tests and cannot see the problem either. So it is not a universal problem. Don't forget this is done with a test MFA made by Volnaiskra, amongst other things.

As conceptgame also correctly states the methodolgy of using Fraps may be flawed, even to the point of introducing delays/error itself. People here are assuming this is Fusion's fault, and that if it is we can fix it - neither thing is certain. We are doing work to try to determine what the case is.

Please just give Yves some time to look at this, but we can make no promises.

[TABLE='class: grid, width: 1502']
[tr][/tr]

[tr]
[td][/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]CPU[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]GPU[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]RAM[/td][/tr]

[tr]
[/tr]

[tr][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Desktop PC[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]Intel Core i5-4460 (4 cores) @ 3.2 GHz[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]GeForce GTX 750 with 1GB GDDR5[/td][/tr]

[tr]
[/tr][tr][td][/td][/tr]

[tr][td]8GB DDR3[/td][/tr]

[tr]
[/tr]

[/TABLE]

I can only get my two laptop specs when I have my laptops, which will be in a few days.

I gotta say, I'm super confused about this. I think there are too many things going on at once.
I think we need super simple .mfa that will stutter for some people and not for some. Then, we could measure the stuttering using the FRAPS tool for example (just to see if our outputs differs)!
At first, we need to check the example without FRAPS running in the background as we don't know if that can cause stuttering as well.

After that, we should measure using FRAPS and post the results here. We should be seeing very different kind of outputs that way. It would clear things up!

By the way, I'm getting some micro-stuttering from using deltatime, I will need to take a closer look, though, as it can be a problem with scrolling rather than stuttering. You know, value going from 0,xxx to 1,xxx -> and jumping that pixel, appearing to stutter.

However, at the moment we have too many "yes", "no", "ifs" and "buts".

ok, somebody make mfa with different settings. We will make own testings and attach results.