Sorry your browser is not supported!

You are using an outdated browser that does not support modern web technologies, in order to use this site please update to a new browser.

Browsers supported include Chrome, FireFox, Safari, Opera, Internet Explorer 10+ or Microsoft Edge.

AppGameKit Classic Chat / AGK 2 Official Development Blog

Author
Message
Marl
12
Years of Service
User Offline
Joined: 19th Nov 2011
Location: Bradford, UK
Posted: 23rd Apr 2014 07:01
Networking (which already is) and Background loading (would be nice) are the only uses I can think of for multi-threading in Tier 1.

Everything else can be achieved with state machine style programming, providing you have only one main loop and no dead stops within the state code.
Digital Awakening
AGK Developer
21
Years of Service
User Offline
Joined: 27th Aug 2002
Location: Sweden
Posted: 23rd Apr 2014 09:37 Edited at: 23rd Apr 2014 09:39
Multithread loading of media would be awesome for the future. Perhaps commands for adding things to load into a queue and then being able to check how many items is left in that queue would be good enough?

Quote: "I've sort of settled on Digital Awakenings idea to have a second type of string using single quotes, so that double quoted strings behave as they do now, but single quoted strings can have escaped characters."


Maybe the opposite of PHP but it will look the cleanest

Quote: "22nd April
Finalised the function call procedure that allows recursive functions with arrays, types, passing by reference or value, and added the ability to handle local arrays in functions. The next problem we're working on is how to handle multi-dimensional arrays, either flattening it into a single large array or having arrays of arrays which opens up interesting future uses. In particular what happens when either of these are dynamically resized, considering it has to be backwards compatible with existing functionality. Happy Easter BTW!"


Good times ahead of us

Naphier
13
Years of Service
User Offline
Joined: 2nd Oct 2010
Location: St Petersburg, Florida
Posted: 23rd Apr 2014 09:43
+1
Quote: "having arrays of arrays which opens up interesting future uses"


@Rick/Paul - when the initial post is updated would you guys mind posting a new message to the thread which states it's been updated? Otherwise it is difficult to notice when you update the first post

Digital Awakening
AGK Developer
21
Years of Service
User Offline
Joined: 27th Aug 2002
Location: Sweden
Posted: 23rd Apr 2014 12:46
Which is why I quote it when I notice But it would be nice if Rick made update post so we don't have to look at the first post all the time.

Markus
Valued Member
20
Years of Service
User Offline
Joined: 10th Apr 2004
Location: Germany
Posted: 23rd Apr 2014 19:48
wow, cool new features
my wishes, can you add foreach .. in .. too?

AGK 108 (B)19 : Windows 8.1 Pro 64 Bit : AMD Radeon HD 6670
Digital Awakening
AGK Developer
21
Years of Service
User Offline
Joined: 27th Aug 2002
Location: Sweden
Posted: 24th Apr 2014 08:37
I would love foreach

Santman
12
Years of Service
User Offline
Joined: 15th Sep 2011
Location: Inverness
Posted: 24th Apr 2014 21:01 Edited at: 24th Apr 2014 21:05
I've been watching the comments and I'd have to say that I agree with Marl - why on earth would you want threading control, especially on a mobile game? In reality, background image loading would be nice....but even that I can't see how it's overly needed other than for animated loading screens? Background networking is in, but has anyone made a game where you start playing level 1 while the others download yet? The interactions I've had with downloading involve a progress bar anyway, background downloading isn't overly needed really on mobiles either. I once used DBPro and Matrix utils to create a background streaming GTA style system, but it was ridiculously complex to coordinate, and only worked when messing around with shared memory spaces also. For example, a city streaming in the back ground as you move around - but the threaded function must be able to access and manipulate the same screen/memory space as the main app or you get pauses as models are created from memory banks etc etc. The level of background AI needed for these style projects is also intense - calculating and creating new NPC characters, animating them, scaling models - DBPro had addons to do this but I still don't think anything overly advanced came out of it??? But on a mobile game, how in all seriousness is going to be able to do this? Please don't take this as offensive, if you could use this great...but the bottleneck for mobiles is usually the rendering pipeline and very few "AAA" mobile titles even feature this style of thing. Actually, only GTA does that I can think of. Lol. Unless you mean true threading of the engine so that is, but lets not forget some devices are phones & tablets, they still need to do networking, messaging, system ops, memory management, gps etc etc in the background. They ARN'T PCs who can dedicate 4 cores to a game.

And for:each? I can't think of a single instance of this being useful. I use this all the time in VB in my day job, but in AppGameKit what is it you want to cycle through? Sprites? Arrays? In AppGameKit we create and assign all these ourselves, not like addressing a VB control in a set we didn't specify - in what case would a for:next loop be on any use?

I'd prefer to see core things like the IDE/compiler being worked on (the IDE is pretty inadequate for large projects, lets be honest - even simple things like saving bookmarks with the file would be enormously beneficial), optimizing the newer android 4+ features (image loading speeds massively jumped when Paul looked at newer ways of doing that which I believe were based on newer android libraries???), voice recog, a broadcast function that worked first time instead of every second, PC support needs a massive overhaul - res setting, actual full screen modes and not stretched/emulated, the ability to use a separate buffer like DBPro (the clear screen command still seems dysfunctional), compiling to a different icon than the AppGameKit one, ability to use another media location for PC perhaps?, a shortcut that isn't within an "AGK" subgroup (nothing like a bit of forced advertising!!), def an inbuilt deploy to android routine.....

I actually think half of these requests are more suited to DBPro and PC only development. Perhaps if more elaboration could be given so to what threading etc is needed for???
Naphier
13
Years of Service
User Offline
Joined: 2nd Oct 2010
Location: St Petersburg, Florida
Posted: 24th Apr 2014 21:36
You make really good points. I honestly can't see myself using threading other than for a loading/waiting animation. And this would best be done at the java/obj c level so that you can call a new thread for the loading anim and dismiss it when you're ready.
Other threading is, as you say, likely unnecessary and overly complex.

As far as things go for for-each. The only thing we could use this on is arrays (unless I'm mistaken). And that can just be done with traditional for-next. Since v2 will be able to pass arrays by reference there's really no need for this, but it is convenient. I'm really hoping that Paul also builds in an array size command. Some other stuff like c++ vector lib commands (size, insert, swap, and please sort!). But those things I can do myself, albeit less efficiently than C++ can and right now I have to write a specific function for every durned array to do so, but passing by ref will fix that.

Anyway, for I tend to agree with you on these. I'd rather see fixes and performance boosts over any new functionality. Especially something as complex as multithreading.

Markus
Valued Member
20
Years of Service
User Offline
Joined: 10th Apr 2004
Location: Germany
Posted: 25th Apr 2014 08:56 Edited at: 25th Apr 2014 08:57
@Santman,Paul
yes, i want cyle through lists.
agk have no lbound(blah[]) or ubound(blah[]) or Count(blah[]),
always i need save array count somewhere and it needs a index variable, it nerves. foreach is faster / easier to write.
foreach a in blah[]
Update(ref a) < byref and change also values in udt element in array
next
maybe the pre compiler can convert this to: (means comething in c++)
for i4711 = 0 to blah[].count-1
Update(ref blah[i4711])
Next

AGK 108 (B)19 : Windows 8.1 Pro 64 Bit : AMD Radeon HD 6670
JimHawkins
14
Years of Service
User Offline
Joined: 26th Jul 2009
Location: Hull - UK
Posted: 25th Apr 2014 15:10
Using threads is, I think, a very advanced programming technique and in innocent fingers capable of killing the machine. Even PureBasic spends most of its Help pages on this with warnings.

As a kind old man I would advise not thinking about this for T1.

T2/Pascal developers can do it as they wish.

Neither DirectX or OpenGL are fully thread-safe. Watch out!

-- Jim - When is there going to be a release?
Santman
12
Years of Service
User Offline
Joined: 15th Sep 2011
Location: Inverness
Posted: 26th Apr 2014 00:36
Markus, I wasn't suggesting you had to justify yourself to me in any way, sorry if that's how I came over. Lol. All I meant was it may better help TGC and everyone else to understand users biggest issues/needs.

As for the for:next, I'm still missing it. Don;t get me wrong, if it makes things easier that's great and I'm all for it...I'm just not seeing the advantage. In terms of what the computer does, for:next and for:each in an array are exactly the same things. We also have to tell it how big an array would be, so how would a count help? Even a repeat:until loop can be used to go through them, which can be faster if checking for one specific element than a for:each as it would always check the entire array.

I'm just curious as to how arrays could be handled faster as I have one 315'000 big in three dimensions and I manipulate it in all kinds of ways without any speed issues...but more speed is always better.
Naphier
13
Years of Service
User Offline
Joined: 2nd Oct 2010
Location: St Petersburg, Florida
Posted: 27th Apr 2014 01:23
Just so AppGameKit is aware:
AdMob is deprecating the standalone SDK in August.
On the site to download the SDK latest version (6.4.1 -- AppGameKit uses 6.1.0)
https://developers.google.com/mobile-ads-sdk/download#downloadandroid
Quote: "Deprecated. On August 1, 2014, Google Play will stop accepting new or updated apps that use the old standalone Google Mobile Ads SDK v6.4.1 or lower. You must upgrade to the Google Play version of the Mobile Ads SDK by then."


I also hope that AppGameKit will take this opportunity to implement other functionality from Google Play services like leaderboards, achievements, cloud data, google sign in, etc.

xGEKKOx
AGK Master
15
Years of Service
User Offline
Joined: 7th Dec 2008
Location: Italy
Posted: 27th Apr 2014 02:40
AGK Team???
And where are they?
Are they real?

I'm waiting some bugs fix that i would done in 2 min.
I need to be calm and have patience.

Long life to Steve!
bjadams
AGK Backer
16
Years of Service
User Offline
Joined: 29th Mar 2008
Location:
Posted: 27th Apr 2014 19:28
I have to agree with Gekko.

Most of the problems reported by users take less than 1 hour to fix, so fixing a few of these every day, and then releasing a regular update with 10-15 fixes will keep a lot of users happy
Naphier
13
Years of Service
User Offline
Joined: 2nd Oct 2010
Location: St Petersburg, Florida
Posted: 27th Apr 2014 20:48
Arbrakan
13
Years of Service
User Offline
Joined: 10th Oct 2010
Location: Geneva
Posted: 29th Apr 2014 19:42
Quote: "
28th April

The interpreter is now pretty much feature complete and multidimensional arrays are represented as arrays of arrays allowing each dimension to be resized individually. Work returns to the parser and making it output bytecode compatible with the new interpreter."


Digital Awakening
AGK Developer
21
Years of Service
User Offline
Joined: 27th Aug 2002
Location: Sweden
Posted: 1st May 2014 09:53
I've been working on the PC screen settings for RTA and encountered a few things.

First off, when altering the windowed resolution the position of the window is affected by the height of the titlebar, or it doesn't take the task bar into consideration. I use a resolution that maximizes the windows height, it is placed so there's space above the window but it goes into the task bar. The size is perfect to have the window from the top of the screen down to the task bar. If the window is small enough to fit on screen and you set the resolution to the same as the startup resolution. You will notice that after setting the resolution it is positioned lower than where it was before.




Here are two snippets that shows up some inconsistencies in framerate. My laptop has a framerate of 60 but I sometimes get 70 FPS. This makes games appear choppy. It also happens more frequently in fullscreen. I have noticed this before when working on RTA but this is a more thorough test. These are commented but also print info on the screen so you can easily compare your results to mine.

Windowed


Fullscreen


Clonkex
Forum Vice President
13
Years of Service
User Offline
Joined: 20th May 2010
Location: Northern Tablelands, NSW, Australia
Posted: 1st May 2014 16:51
Quote: "My current project. There is some screens."


Wow, looks nice.

Arbrakan
13
Years of Service
User Offline
Joined: 10th Oct 2010
Location: Geneva
Posted: 1st May 2014 22:28
Quote: "Wow, looks nice."


Thanks!,
Hope I can share a video soon.

Paul Johnston
TGC Developer
21
Years of Service
User Offline
Joined: 16th Nov 2002
Location: United Kingdom
Posted: 1st May 2014 23:37 Edited at: 1st May 2014 23:44
Quote: "Here are two snippets that shows up some inconsistencies in framerate."


SetVSync and SetSyncRate are mutually exclusive, setting one overrides the settings of the other so you should only use one in your code. The 70 fps happens if you turn VSync on but for some reason the platform doesn't support. VSync would normally limit the fps to 60 so AppGameKit sets a fallback of 70 fps using a timer to stop the framerate being unlimited.

Quote: "AdMob is deprecating the standalone SDK in August."


Thanks for letting me know, I'll see if we can get something in place before then.

Quote: "when altering the windowed resolution the position of the window is affected by the height of the titlebar"


Thanks, I've modified the window position to account for the taskbar like the initial size code does.
Harlequin
14
Years of Service
User Offline
Joined: 22nd Mar 2010
Location: Melbourne (AU)
Posted: 2nd May 2014 05:22
Considering pledging, but so I know what to expect, I was wondering are stretch target features like the terrain system implemented yet?

Ultimately I guess how complete is v2? (ignoring the bugs as it is an Alpha)

thanks

It's late, I'm tired, out of coffee and now I'm wet.
Ancient Lady
Valued Member
20
Years of Service
User Offline
Joined: 17th Mar 2004
Location: Anchorage, Alaska, USA
Posted: 3rd May 2014 20:59
Harlequin, the first post in this thread is being regularly updated to show what they are currently working on. Based on that, it looks like they are working on the second stretch goal. So, I don't think they have started on the terrain system yet, which is stretch goal 3.

Quote: "The interpreter is now pretty much feature complete and multidimensional arrays are represented as arrays of arrays allowing each dimension to be resized individually."

Pretty cool!

The separate dimension resizing is an interesting feature.

Cheers,
Ancient Lady
Clonkex
Forum Vice President
13
Years of Service
User Offline
Joined: 20th May 2010
Location: Northern Tablelands, NSW, Australia
Posted: 6th May 2014 04:42
Harlequin vs Paul Johnston


F I G H T!!


...what? You don't get it? ...same avatar...? But... it IS funny- ....what? I'll- I'll just....be- be going now....

Harlequin
14
Years of Service
User Offline
Joined: 22nd Mar 2010
Location: Melbourne (AU)
Posted: 7th May 2014 03:45
Thanks AncientLady, that was the conclusion I had, but wanted to see if there was any more detail on what stage AppGameKit was up to.


I like lava lamps, but not sure I would want to go one-on-one for them and risk distracing Paul from AppGameKit dev work

It's late, I'm tired, out of coffee and now I'm wet.
xCept
21
Years of Service
User Offline
Joined: 15th Dec 2002
Location:
Posted: 7th May 2014 18:19
Quote: "6th May

We're still working on the parser to make sure it's backwards compatible with existing code, which has some interesting edge cases that need handling. For example when calling a gosub from within a function the gosub loses access to the function local variables and gains access to non-global variables outside the function. Hopefully we won't run into many more unusual cases this week and have something up and running soon. Cheers for being so patient while we work on this very deep part of AppGameKit V2!"


CAN ANYONE GET V2 PLAYERS TO COMPILE AND RUN FOR IOS OR ANDROID? ARE YOU ABLE TO RUN ALL OF THE V2 SAMPLES SUCCESSFULLY ON IPHONE/ANDROID?

I have had no success getting V2 players to build or run V2 features (shaders, render to image etc) and would much rather have working players than months of updates to the default IDE. Especially since it is indicated full support has been added to iOS/Android as of Alpha 3. (Sorry for shouting, still haven't gotten a single response about this.)
Naphier
13
Years of Service
User Offline
Joined: 2nd Oct 2010
Location: St Petersburg, Florida
Posted: 8th May 2014 00:09
I just found an interesting performance issue when switching over to AGK10821 for dFenz. I was messing around with my monitor setup and GPUs and turned off AMD's GPU Graphics Overdrive. WOW! AppGameKit games run at a max of 30 FPS with vsync on and if you have lots of physics shapes then the game engine just drops FPS until it dies.
I can't imagine that a lot of end-users turn on AMD OverDrive by default. I hope the performance in v2 is a lot better because if anyone downloaded an AppGameKit game and had that experience it would be terrible. Anyway, turning it back on cured everything after I was searching for hours for the reason of the seemingly random crash. The biggest issue came when I had a large enemy sprite on the screen.

Clonkex
Forum Vice President
13
Years of Service
User Offline
Joined: 20th May 2010
Location: Northern Tablelands, NSW, Australia
Posted: 8th May 2014 07:27
@Naphier:

I would never buy an AMD device. I bought a laptop (AMD A8 processor with dedicated AMD graphics chip) a couple of years ago (my first laptop!), and while I loved it, and I have a lot of great memories associated with it (BF3 campaign in an 800x600 window at barely 25fps!), I had such terrible experiences with AMD's software that I would never again go to them. Nvidia and Intel FTW!

Naphier
13
Years of Service
User Offline
Joined: 2nd Oct 2010
Location: St Petersburg, Florida
Posted: 8th May 2014 08:26
That's nice for you, but Intel graphics are some of the worst benchmarks, nvidia is overly expensive and AMD has 20% market share (Intel is 60 something and nvidia 16%). Whatever the market share is it doesn't matter. Software needs to be compatible with major hardware. I have had great experiences with AMD (though I'd love an nvidia card). I've played 1,000s of games with my AMD setup and never have experiened such huge drops in frames. The AppGameKit engine needs stabilizing.

Clonkex
Forum Vice President
13
Years of Service
User Offline
Joined: 20th May 2010
Location: Northern Tablelands, NSW, Australia
Posted: 8th May 2014 09:02
Quote: "Intel graphics are some of the worst benchmarks"


Which is why I would use an Intel CPU with an Nvidia GPU.

Quote: "nvidia is overly expensive"


I would disagree with that.

Quote: "AMD has 20% market share"

Quote: "nvidia 16%"


Really? Wow. Wait, do you mean share in GPUs or CPUs?

Quote: "Whatever the market share is it doesn't matter. Software needs to be compatible with major hardware."


I do get your point. My post was intended to insult AMD, not say that your problem is not legitimate. While AMD's drivers tend to be bloated and unreliable (in my experience), and are almost certainly not correctly handling some feature of OpenGL that is obscure but allowed that AppGameKit uses, AppGameKit definitely needs to run no matter the hardware. So if AMD's drivers don't work properly, then unless the next update of the drivers fixes that, AppGameKit needs to be the one to fix it by working around :/

bjadams
AGK Backer
16
Years of Service
User Offline
Joined: 29th Mar 2008
Location:
Posted: 8th May 2014 11:26
intel + nvidia is the only good combination for gaming.
JimHawkins
14
Years of Service
User Offline
Joined: 26th Jul 2009
Location: Hull - UK
Posted: 8th May 2014 13:42
I have two computers here which have been running with AMD processors and graphics boards for several years without fault. The one I'm using now pumps out 120Hz video to an LG monitor for about 16 hours a day and has NEVER failed.

Another one has Intel-i5 and Nvidia graphics with 3D Vision. It's more inclined to screw up than the AMD machines.

There are four fast Samsung laptops with i5 processors and AMD graphics and they work perfectly.

Most of the Nvidia boards I have had over many years have failed a lot faster than the ATI/AMD boards.

To do a proper test would only be to change graphics boards in the same computer and then compare frame-rates. Of course, they'd have to be comparable boards!

-- Jim - When is there going to be a release?
Markus
Valued Member
20
Years of Service
User Offline
Joined: 10th Apr 2004
Location: Germany
Posted: 8th May 2014 14:11
@Paul

please can u say what curl lib version u use and update this lib in agk 1 or 2?

AGK 108 (B)19 : Windows 8.1 Pro 64 Bit : AMD Radeon HD 6670
Paul Johnston
TGC Developer
21
Years of Service
User Offline
Joined: 16th Nov 2002
Location: United Kingdom
Posted: 8th May 2014 15:45
Quote: "please can u say what curl lib version u use and update this lib in agk 1 or 2?"


We use cURL only on Android, and it is using 7.25.0. I'm a strong believer in "if it ain't broke don't fix it", so I won't be updating it unless there is something wrong with the current version.

Quote: "I just found an interesting performance issue when switching over to AGK10821"


Please can you provide a project that behaves at noticeably different speeds with different versions, or go into detail about what caused the slow down.

Quote: "I have had no success getting V2 players to build or run V2 features"


I have built and run the player on iOS and Android, albeit briefly, what problems are you having exactly? Does the Android player in AGK/Players/Android work?
Markus
Valued Member
20
Years of Service
User Offline
Joined: 10th Apr 2004
Location: Germany
Posted: 8th May 2014 16:05
@Paul,
ok, i just asked because there was many fixes in this lib.
maybe it solves "xGEKKOx" problem.
7.25.0 - March 22 2012
7.36.0 - March 26 2014

AGK 108 (B)19 : Windows 8.1 Pro 64 Bit : AMD Radeon HD 6670
Clonkex
Forum Vice President
13
Years of Service
User Offline
Joined: 20th May 2010
Location: Northern Tablelands, NSW, Australia
Posted: 8th May 2014 17:28
Quote: "To do a proper test would only be to change graphics boards in the same computer and then compare frame-rates. Of course, they'd have to be comparable boards!"


I only spoke from personal experience (bad with AMD, good with Intel/Nvidia), so it's not scientific by any means

I find Intel's CPU design to be better than AMD's. Intel goes for fewer, more powerful cores, while AMD prefers a greater number of less powerful cores. To me, it seems Intel has the advantage over AMD. Very few programs can actually use more than 4 cores, and when they can't, having 8 or 16 slower cores isn't going to help anyone. Most programs are optimised for quad-core architecture.

xGEKKOx
AGK Master
15
Years of Service
User Offline
Joined: 7th Dec 2008
Location: Italy
Posted: 8th May 2014 17:58 Edited at: 8th May 2014 18:04
Why you use cURL on Android, instead the internal Java HTTP request?
I used it in the past and it worked fine and really easy.

I remember it was really similar to the OBJ C one.
Maybe you can do a JNI function with the java internal HTTP.

This is the link where i saw it LINK

This link show the POST code LINK

Long life to Steve!
Paul Johnston
TGC Developer
21
Years of Service
User Offline
Joined: 16th Nov 2002
Location: United Kingdom
Posted: 8th May 2014 19:23
Quote: "Why you use cURL on Android, instead the internal Java HTTP request?"


Because at that point in development we didn't have a solid connection between Java and the NDK native code. JNI is a suitable solution for that but since our HTTP code uses threads I'm not sure if it would work in this case. So far we have not had any problems with cURL, and unfortunately I was not able to replicate any of your problems when connecting to the same server.
Naphier
13
Years of Service
User Offline
Joined: 2nd Oct 2010
Location: St Petersburg, Florida
Posted: 8th May 2014 19:54
@Paul - The project is messy, but I've also tried it out with 2 other projects with the same result, so either I'm doing something inherently wrong or we've got a performance issue. I will try working on an example project that I can share with you over the weekend. I wonder if this will also point to the reason why so many folks have an issue with AppGameKit on Windows 8. We're doing Windows 8 tests today so that we can try to narrow down what commands we're having issues with as well. Also, please don't change the cURL in Android. It works so well right now. However, I did notice when developing Wordspionage that sometimes if you re-use an HTTP connection that you might get the response from a previous connection even though it had been closed, and reopened. Actually that was only on iOS, so it could be completely unrelated. However, iOS has some quirks in the HTTP commands that I've pointed out in the bug forum.

@everyone else, let's start a new thread if you all want to discuss the merits of AMD vs Intel vs nVidia. I enjoy the discussion, but I feel this thread will quickly be hijacked if we keep up.

Timshark
16
Years of Service
User Offline
Joined: 30th Jun 2007
Location: Oslo, Norway
Posted: 9th May 2014 01:02
@Paul

Don't want to sound greedy, you have given us spine support already, and development looks great....

But is there any chance that you will update the Spine support with the new ffd (free form mesh deformation) in the future? When all the other work is done?

I never want what I know.
Naphier
13
Years of Service
User Offline
Joined: 2nd Oct 2010
Location: St Petersburg, Florida
Posted: 9th May 2014 09:02
@Paul - I can't, for the life of me, reproduce the frame rate drop issue. I've tried a thousand things and I must have just got lucky and fixed something I did wrong. Sorry for the false alarm. I swear it was consistent

Clonkex
Forum Vice President
13
Years of Service
User Offline
Joined: 20th May 2010
Location: Northern Tablelands, NSW, Australia
Posted: 9th May 2014 09:54
Quote: "@everyone else, let's start a new thread if you all want to discuss the merits of AMD vs Intel vs nVidia. I enjoy the discussion, but I feel this thread will quickly be hijacked if we keep up."


Indeed, hijacked threads are not good. I think the discussion was good but not really worth continuing.

Quote: "I can't, for the life of me, reproduce the frame rate drop issue. I've tried a thousand things and I must have just got lucky and fixed something I did wrong. Sorry for the false alarm. I swear it was consistent"


Haha You probably just restarted your computer

Naphier
13
Years of Service
User Offline
Joined: 2nd Oct 2010
Location: St Petersburg, Florida
Posted: 9th May 2014 09:56
xGEKKOx
AGK Master
15
Years of Service
User Offline
Joined: 7th Dec 2008
Location: Italy
Posted: 9th May 2014 19:14
@Paul

What about the timeout parameter in the function at least?

Long life to Steve!
Tapewormz
21
Years of Service
User Offline
Joined: 15th Sep 2002
Location: Winnipeg, Mantoba, Canada
Posted: 10th May 2014 01:53
Spine support sold me on AppGameKit 2. Any word if there will be support for Spriter?
Digital Awakening
AGK Developer
21
Years of Service
User Offline
Joined: 27th Aug 2002
Location: Sweden
Posted: 10th May 2014 15:02 Edited at: 10th May 2014 15:07
Paul:

Quote: "Thanks, I've modified the window position to account for the taskbar like the initial size code does. "


Great


Quote: "SetVSync and SetSyncRate are mutually exclusive, setting one overrides the settings of the other so you should only use one in your code. The 70 fps happens if you turn VSync on but for some reason the platform doesn't support. VSync would normally limit the fps to 60 so AppGameKit sets a fallback of 70 fps using a timer to stop the framerate being unlimited."


I have tested my code on a my brand new gaming PC and it has a lot less 70 FPS errors. Why is the fallback 70 FPS and not 60 FPS? 70 FPS is very likely to cause stutters, which happens on both my PCs. This also happens when I manually set the syncrate to 70. Since there are 2 draws every 1/6 syncs, everything will stutter.

This code gives me 70 FPS for no reason on both my laptop and my gaming PC. When the syncrate is default or set to 60 and vsync is set to 0 I get 70 FPS. This doesn't happen if the syncrate is 120.



bjadams
AGK Backer
16
Years of Service
User Offline
Joined: 29th Mar 2008
Location:
Posted: 10th May 2014 19:40
@Tapewormz there won't be Spriter support natively unless a third party comes up with some code for that.

I prefer Spriter myself, but luckily for me other sdks do support Spriter natively
Paul Johnston
TGC Developer
21
Years of Service
User Offline
Joined: 16th Nov 2002
Location: United Kingdom
Posted: 12th May 2014 16:14
Quote: "What about the timeout parameter in the function at least?"


It's still on my list of things to do, but I don't think it will work as well as you hope. As you said iOS does not have any problems and it also has a 6 second time out, so unless your server responds differently to iOS and Android requests it sounds more like the device you are using is having problems with a firewall or something.

Quote: "Any word if there will be support for Spriter?"


Not for version 2, but after the kickstarter goals are done we can come back and look at it.

Quote: "Why is the fallback 70 FPS and not 60 FPS?"


I didn't want it competing with the vsync, if they were both at 60 fps sometimes the timer might kick in and ruin the vsync timing.

Quote: "setsyncrate( 60, 0 )
setvsync( 0 )
"


These are overwriting the settings of each other. Try this instead
xGEKKOx
AGK Master
15
Years of Service
User Offline
Joined: 7th Dec 2008
Location: Italy
Posted: 12th May 2014 19:00
@Paul
I don't use AppGameKit for HTTP.
I use OBJ C, this is why i have no problem on iOS/MAC.

Long life to Steve!
Digital Awakening
AGK Developer
21
Years of Service
User Offline
Joined: 27th Aug 2002
Location: Sweden
Posted: 12th May 2014 22:01
Paul:
Thanks for the explanations. I've made it to a single option where the user can use vsync or specify 60 or 120 FPS, or even unlimited. 60 FPS is the default. So I won't use both commands at the same time anymore.

The Zoq2
14
Years of Service
User Offline
Joined: 4th Nov 2009
Location: Linköping, Sweden
Posted: 13th May 2014 16:35
Quote: "I don't use AppGameKit for HTTP.
I use OBJ C, this is why i have no problem on iOS/MAC."


Can't you use an external C++ library on android to then?

Say ONE stupid thing and it ends up as a forum signature forever. - Neuro Fuzzy

Login to post a reply

Server time is: 2024-05-07 02:06:29
Your offset time is: 2024-05-07 02:06:29