Jump to content

Photo

Zelda Classic 2.53.0 Beta Release 2 (Official)


  • This topic is locked This topic is locked
118 replies to this topic

#16 ZoriaRPG

ZoriaRPG

    ZC Contributor & Wiki Editor

  • Members
  • Gender:Unspecified
  • Location:Prydon Academy

Posted 12 July 2017 - 10:02 AM

If you're going to nm ot run the old stuff, I highly suggest changing the name. It's a different game at that point and not based off of Zelda 1.
Just saying. (not necessarily a bad thing)
-James

It's the other way around. Olde versions of ZC won't be able o run 2.54, 2.55, and 2.60 quests. My apologies for any confusion.

2.53.0 will run all older quests, as will future versions. 2.50.0, 2.50.1, 2.50.2, and 2.50.3 can also run 2.53.0 quests. Future ZQuest versions will require a version of ZC greater than, or equal to their version ID to play quests made in that ZQuest build.

My goal is both to prevent feature-locking future ZC versions, and to encourage the userbase to use the latest version of ZC, thus reucing the support overhead.

Edited by ZoriaRPG, 12 July 2017 - 10:14 AM.

  • Nightmare likes this

#17 Nightmare

Nightmare

    Original ZC Tester

  • Members
  • Real Name:James
  • Gender:Unspecified
  • Location:Jackson, NJ

Posted 12 July 2017 - 10:38 AM

It's the other way around. Olde versions of ZC won't be able o run 2.54, 2.55, and 2.60 quests. My apologies for any confusion.
2.53.0 will run all older quests, as will future versions. 2.50.0, 2.50.1, 2.50.2, and 2.50.3 can also run 2.53.0 quests. Future ZQuest versions will require a version of ZC greater than, or equal to their version ID to play quests made in that ZQuest build.
My goal is both to prevent feature-locking future ZC versions, and to encourage the userbase to use the latest version of ZC, thus reucing the support overhead.

I still don't think a rebranding would be a bad idea at this point. Zelda Classic has been around forever, and we really haven't had a terribly new product. A new name, a fresh coat of paint, is not a farfetched idea after 17 years.

Also, I know numbers are a strange thing, but the stuff you're having me test and things makes me feel more like 3.0 than 2.55+

Not trying to be offensive here, just constructive. Apologize ahead of time if I'm too direct or forceful.

-James

Edited by Nightmare, 12 July 2017 - 10:41 AM.


#18 Saffith

Saffith

    IPv7 user

  • ZC Developers
  • Gender:Male

Posted 12 July 2017 - 10:49 AM

On the one non-virtual Windows system I have access to, I tried starting the first quest and just standing there. In 2.50.2, it runs at about 1670 FPS. In this build, it's about 995 FPS.
 

This is not a 'new feature' release. 2.53.0 exists to stabilise 2.50.2 and 2.50.3 into one unified body, as a final stable 2.50.x version.

I'm still pretty unclear on this. This is an official release, not your own fork? Why is it 2.53 instead of 2.50.3? I thought the only reason for changing the version number was that change to the combo solidity checking, which seems to have been reverted.
Misc.>About still says 2.50. And it looks like you removed ZELDA_VERSION_STR...?

#19 Avataro

Avataro

    Quest maker xD

  • Members
  • Real Name:Robin xD
  • Gender:Male
  • Location:Germany

Posted 12 July 2017 - 11:10 AM

I thought this was 2.50.3.

 

(About the FPS think i might get to it later xd)



#20 ZoriaRPG

ZoriaRPG

    ZC Contributor & Wiki Editor

  • Members
  • Gender:Unspecified
  • Location:Prydon Academy

Posted 12 July 2017 - 02:49 PM

This is official, and approved by (at least) DD.

 

2.53.0 is compiled from the code at GitHub.com/ArmageddonGames/ZeldaClassic , branch 2.50.x.

 

Th 2.53 desination came about through discussions on Skype, here with the userbase, and on AGN.

 

It effectively replaces 2.50.3.


Edited by ZoriaRPG, 12 July 2017 - 02:53 PM.


#21 Avataro

Avataro

    Quest maker xD

  • Members
  • Real Name:Robin xD
  • Gender:Male
  • Location:Germany

Posted 12 July 2017 - 04:28 PM

But 2.50.3 never released. We only got it's release candiate. I think it'd be more appropriate to call this 2.50.3.

 

Well, 2.53 has at least one advantage. It's faster to type..


Edited by Avataro, 12 July 2017 - 04:32 PM.


#22 DarkDragon

DarkDragon

    Junior

  • ZC Developers
  • Gender:Unspecified

Posted 13 July 2017 - 01:43 AM

Is the performance difference based on both executables build from source using the same toolchain?

 

Or is there a possibility that the two were built with different compilers/optimization flags? ZoriaRPG, did you post executables build in debug mode, or release?

 

If the comparison is not apples to apples I'm not worried, especially since the only change that could really have affected performance at all is upgrading the Allegro library.


ZC randomly switched my windows theme to windows 7 basic once, but it was a one-time thing.


Unfortunately this is unfixable. Allegro uses the ancient and long-deprecated DirectDraw API which is bound to act a little idiosyncratic on modern versions of Windows.

I still don't think a rebranding would be a bad idea at this point. Zelda Classic has been around forever, and we really haven't had a terribly new product. A new name, a fresh coat of paint, is not a farfetched idea after 17 years.

I agree in principle, but this release is intended to just be a bugfix release, to provide you with the most stable platform for building quests that we currently have available. It's a tidying-up of the previous 2.50 betas, and doesn't include any new features we have been working on.


Edited by DarkDragon, 13 July 2017 - 01:40 AM.


#23 ZoriaRPG

ZoriaRPG

    ZC Contributor & Wiki Editor

  • Members
  • Gender:Unspecified
  • Location:Prydon Academy

Posted 13 July 2017 - 04:05 AM

Is the performance difference based on both executables build from source using the same toolchain?
 
Or is there a possibility that the two were built with different compilers/optimization flags? ZoriaRPG, did you post executables build in debug mode, or release?


Release. CMake config: No PCH, non-Unity; MSVS CPP compiler v9, (2008). It is of course possible that GCC is better optimised in some regard.

( I would never dump out a public beta built as a debug compilation. :/ )

@Saffith: What framerate do you have for 2.50.3RC1, using the same VM?

Edited by ZoriaRPG, 13 July 2017 - 04:08 AM.


#24 Gleeok

Gleeok

    It's dangerous to dough alone, bake this.

  • ZC Developers
  • Real Name:Pillsbury
  • Gender:Male
  • Location:Magical Land of Dough

Posted 13 July 2017 - 04:41 AM

The performance of this build is indeed very poor. I'm seeing about a 2x slowdown from 2.50.2 here. Time to break out the profiling tools!


[edit] Yeah.... ZoriaRPG: Compile 2.53 (with the same settings) and link with the old allegro4.2 lib [and not 4.4] instead. You should see it (if for some reason you don't can you upload that here to make sure).

#25 ZoriaRPG

ZoriaRPG

    ZC Contributor & Wiki Editor

  • Members
  • Gender:Unspecified
  • Location:Prydon Academy

Posted 13 July 2017 - 06:36 AM

The performance of this build is indeed very poor. I'm seeing about a 2x slowdown from 2.50.2 here. Time to break out the profiling tools!


[edit] Yeah.... ZoriaRPG: Compile 2.53 (with the same settings) and link with the old allegro4.2 lib [and not 4.4] instead. You should see it (if for some reason you don't can you upload that here to make sure).


Gleeok, is there any chance you would be willing either to stream the process that you go through in figuring this out; or record and upload it? I do not need live commentary, or to chat during a stream. I merely want to watch you work it out, so that I can learn how you handle it, for the future.

Edit. Flidd, I replied without refreshing. Allegro 4.2 is just plain out, as it is the pure source of half a dozen critical bugs that this fixes; and future stuff completely relies on it. If the performance drop is purely caused by 4.4, and there is no way to rectify that (still using 4.4) , its just a bitter pill to swallow, and move on for now.

Future builds are destined to have even greater system requirements too. Hell, jman was recetly talking about using HW rendering in the future, which would make it nearly impossible for me to run ZC at full speed on any system that I own. :(

The only thing that I can suggest, might be to try the 4.4.3 version of allegro that is only on GitHub, and hope that it addresses performance issues. I believe that we are using 4.4.2, but I have no clue what was actually changed between the two, as there is no basic changelog entry for 4.4.3, and I would need to review all the GH commits for it.

I'm open to suggestions, but if people can't run this because it uses ag4.4, and runs too slow, then why are they running stuff like Windows 10 on their HW?

Edited by ZoriaRPG, 13 July 2017 - 06:44 AM.


#26 DarkDragon

DarkDragon

    Junior

  • ZC Developers
  • Gender:Unspecified

Posted 13 July 2017 - 08:42 AM

Again before making conclusions I would like to see numbers for executables built with the *same* compiler.



#27 Zeldaman23

Zeldaman23

    Darkie & Dhane

  • Members
  • Real Name:Bradley Dotson
  • Gender:Male

Posted 13 July 2017 - 11:09 AM

I like to report few problem with ZC.   Not really seriously problem.  3 things. 

 

1.  I am using my PSP and Joy2Key because I don't want use circle pad.  Even I disable circle pad, it didn't work for d-pad.   There are few times force left d-pad but I made quick mash button like combine d-pad button.

 

2. http://prntscr.com/fvahaw I am not sure problem with 2.53 or quest itself.   There is bug on this.   After I kill all enemies and got key, I don't want go around on third floor.  So I went back to black hole, but somehow, I am standing on hole that shouldn't do that.   So I use leaf to fly over and see what happen to land on hole, it disappear.  I though it froze, but it not.  I use leaf to show you where you are at.  There is way back to normal.   You need fall down to hole.   But not texture hole make you fall down to second floor.   Even I went outside, it still disappear.   I went back inside. Still disappear until I fall down hole, it back to normal.   But I have to go third floor to get boss key.  

 

3. http://prntscr.com/fvanoy  Is this normal way even you get staff, it still effect to talk rock block?

 

2 & 3


Edited by Zeldaman23, 13 July 2017 - 11:10 AM.


#28 Gleeok

Gleeok

    It's dangerous to dough alone, bake this.

  • ZC Developers
  • Real Name:Pillsbury
  • Gender:Male
  • Location:Magical Land of Dough

Posted 13 July 2017 - 05:23 PM

Gleeok, is there any chance you would be willing either to stream the process that you go through in figuring this out; or record and upload it? I do not need live commentary, or to chat during a stream. I merely want to watch you work it out, so that I can learn how you handle it, for the future.

Edit. Flidd, I replied without refreshing. Allegro 4.2 is just plain out, as it is the pure source of half a dozen critical bugs that this fixes; and future stuff completely relies on it. If the performance drop is purely caused by 4.4, and there is no way to rectify that (still using 4.4) , its just a bitter pill to swallow, and move on for now.


It would be pretty pointless to do that. You can just google for c++ profiling tools instead. :P
Now maybe live fixing of non-performant code and optimizing in general might be useful, but why do that when it could be a simple fix. Diagnose first, then fix if needed.

It looks to me like the allegro44 dll might need to be recompiled. It seems it is spending way too much time doing color conversions compared with the previous library we've used but I'm not entirely sure why with the dll in the trunk. Could be any number of things: Maybe they messed up the CMAKE for Visual Studio and an optimization was turned off; maybe something similar that's causing the L2 to get thrashed, or, because of aliasing issues we're getting fucked with load-store hits that weren't there before. idk.

This is why I suggested you to recompile with allegro4.2 (just a few quick changes) with the other exact same settings and toolchains/binaries.

#29 Saffith

Saffith

    IPv7 user

  • ZC Developers
  • Gender:Male

Posted 13 July 2017 - 06:48 PM

Allegro 4.2 had color conversion and blitting routines written in assembly, which were disabled in 4.4. That may well be a major factor. It looks like all the code's still there; maybe it'd be possible to re-enable it.

#30 Gleeok

Gleeok

    It's dangerous to dough alone, bake this.

  • ZC Developers
  • Real Name:Pillsbury
  • Gender:Male
  • Location:Magical Land of Dough

Posted 13 July 2017 - 07:48 PM

Oh no! *shock and amazement* They broke the library when they switched to CMAKE! *totally surprised gasp* What are the odds!?!? *complete disbelief and not blinking anymore* How could this happen??? *cries until dies of dehydration*

Do I think I'm too good for "I told you so?" ... Nope. Not even close. :D :D :D


0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users