Blighted Moonbase IV - 100% this map for 200 bucks

[Gaia74]

24 minutes ago, OniriA said:

Cacoward for worst wad 2018?

yea !! xD

[therektafire]

If only it was released in 2017 it could have gotten mockaward, but I don't think that is a thing this year iirc :(

[Player177]

21 hours ago, therektafire said:

edit: crap and it looks like the newest 64 bit version of gzdoom won't run on the other computer which I sometimes use to play wads and used to use to make maps on which has a slightly better cpu but an older GPU that only has opengl 3.1 and the minimum requirement is 3.3, and there is no 3.1 version for 64 bit, only 32 bit... ooofff. What was the last 64 bit version that supports opengl 3.1 so I can try it on this one 

In this case I don't even know how notable the difference between 32 bit and 64 bit can be. Just tried both GZDoom 3.5.1 versions on Dual Core E5800 3.2 OpenGL 4.5 with bloody moonbase. Changed different settings. Indiscernible. Was looking at frames per second.

 

Meanwhile, I took some helpful tips from comments and sort of wanted to check closely one conspicuous  feature of the map which among others was defined as a negative component that greatly reduces framerate.  Here they are - constantly mentioned decorative things such as pool of blood and similar stuff. I removed all things type: 80, 24. There were like over 8000! One on top of another. The result is shown on video that I'm frigging tired to mold. I took into account the author's style of argumentation and also lately it is essential at times to prove something somehow so that there are less suspicious probabilities.
By the way, recorded on an old intel core duo T9800 in software mode. On the left screen you see the current map edition and on the right you see the same map only without 8000 floor decorations. The difference is significant!

 

 

 

Pan, if you are interested in this, although I understand the pain. How to remove all this fast in DB2: choose Things mode, then Find Replace mode, search Thing Type and Find. Then just close the window and push Delete. Maybe this method will help with something else. In general, it was a demonstrative experiment. Therefore, we can save money instead of buying modern i7, i9 or 12 cores CPU's 😆 

[Novaseer]

I feel like, with an offer like that, there's going to be the classic MAP15 issue of a secret sector being unreachable, or alternatively have a monster in an inaccessible area, and thus a 100% would not be possible.

[Pan]

5 hours ago, Kapanyo said:

Yeah, just splitting off different "areas"' walls instead of all of them should suffice. I've gotta say, I'm still floored that there exists a map that looks like a Vanilla WAD yet brings normal systems to their knees.

 

Awesome! So it clearly is an issue of too many decorations. I’ll work on this on this as it seems quite easy to implement. I’ll be back in a day or so. Thanks Player 177!

 

I will deleted all things 80 and 24, tidy the map up and post a new link. 

[OniriA]

Can someone  upload a demo of a full playthrough on Youtube including all kills, items etc. just for sheer entertainment?

 

Just to be done with it all (as I dont feel like playing it either), and after that just turn our back to this and ignore the thread altogether.

 

 

 

 

 

 

 

[therektafire]

Ok, so, I decided to be really helpful for a change and instead of simply pointing out issues decided to include pictures of debug puke with them to illustrate my point :D Maybe Pan is a visual learner and can learn from this despite his extreme stubbornness which is even worse than mine. Note that the graphics debug tools only work in the GL renderer so that is what I used to get this info. OK, so, look at this screenshot:

 

 

As you can see I have basically all rendering disabled and am getting 60 FPS in console, out of the console I get 84-85 FPS.

 

Now look at this screenshot here

 

 

As you can see, with just rendering the WALLS the FPS goes down to a consistent 26-27 out of the console. And if you look at the render stats you can clearly see it is trying to render OVER 9000!!! walls. And I can assure you you definitely can't see over 9000 walls from where I am pointing which is the direction where the most walls are rendered in the whole map. Adding floors decreases the FPS to 15-16, interestingly if I disable walls and just leave floors it goes back to 26, the same framerate I got when I just had walls alone as well.

 

Now here is the interesting part, and the part that shows without a doubt that A) you have way too much decorative shit in the map, and B) the map is built poorly. So, lets see what the framerate is when I try to render the sprites as well as the walls...

 

 

5... 5 FPS. Now, what do you think will happen if I disable everything *except* for sprites?

 

 

still ONLY 8!!! Disabling walls and floors jumped the FPS from 16 to 84 when sprites were off, but when they were on it did absolute FA. And this on your "less blood" map by the way... Also, another interesting thing to note is that in the screenshot with the walls and sprites being rendered and in the one with only the sprites, notice how the amount of sprites being rendered is THE EXACT SAME despite there BEING WALLS THERE! This also 100% undeniably proves there are no visblocking linedefs in the whole map which is what is causing everything to be rendered at once. This is why lots of Doom (and quake as well for that matter) maps are corridor based and don't have huge wide open areas with no non-entity world geometry that blocks visibility anywhere.

[Pan]

25 minutes ago, therektafire said:

Ok, so, I decided to be really helpful for a change and instead of simply pointing out issues decided to include pictures of debug puke with them to illustrate my point :D Maybe Pan is a visual learner and can learn from this despite his extreme stubbornness which is even worse than mine. Note that the graphics debug tools only work in the GL renderer so that is what I used to get this info. OK, so, look at this screenshot:

 

 

As you can see I have basically all rendering disabled and am getting 60 FPS in console, out of the console I get 84-85 FPS.

 

Now look at this screenshot here

 

 

As you can see, with just rendering the WALLS the FPS goes down to a consistent 26-27 out of the console. And if you look at the render stats you can clearly see it is trying to render OVER 9000!!! walls. And I can assure you you definitely can't see over 9000 walls from where I am pointing which is the direction where the most walls are rendered in the whole map. Adding floors decreases the FPS to 15-16, interestingly if I disable walls and just leave floors it goes back to 26, the same framerate I got when I just had walls alone as well.

 

Now here is the interesting part, and the part that shows without a doubt that A) you have way too much decorative shit in the map, and B) the map is built poorly. So, lets see what the framerate is when I try to render the sprites as well as the walls...

 

 

5... 5 FPS. Now, what do you think will happen if I disable everything *except* for sprites?

 

 

still ONLY 8!!! Disabling walls and floors jumped the FPS from 16 to 84 when sprites were off, but when they were on it did absolute FA. And this on your "less blood" map by the way... Also, another interesting thing to note is that in the screenshot with the walls and sprites being rendered and in the one with only the sprites, notice how the amount of sprites being rendered is THE EXACT SAME despite there BEING WALLS THERE! This also 100% undeniably proves there are no visblocking linedefs in the whole map which is what is causing everything to be rendered at once. This is why lots of Doom (and quake as well for that matter) maps are corridor based and don't have huge wide open areas with no non-entity world geometry that blocks visibility anywhere.

 

That is extremely interesting. Thank you.

 

From what I gather, this map is irredeemably cursed. There are no blocking visibility lines as you say,and this is to allow player free access ( as well as the monsters) to every corner of the maze. Deleting lots of the decorations may improve the performance, but including blocking lines will totally change the flow of the map. The map is currently largely based around monsters being able to follow you across huge distances without getting blocked.

 

this map is fundamentally satanic in its laggy design, and that can't be fixed without a total rework of the map, which would hopefully lift this curse. 

 

Id rather continue working on my new map project and banish this accursed, blighted map to the darkness of space where it belongs.

[Ancalagon]

Guys we're getting there, keep trying!

 

39 minutes ago, Pan said:

From what I gather, this map is irredeemably cursed. There are no blocking visibility lines as you say,and this is to allow player free access ( as well as the monsters) to every corner of the maze. Deleting lots of the decorations may improve the performance, but including blocking lines will totally change the flow of the map. The map is currently largely based around monsters being able to follow you across huge distances without getting blocked.

 

Other people can explain with more technical details than me, but "blocking visibility" in this case we're talking about the engine. If a wall doesn't go from the floor to the ceiling it doesn't matter if you as a player can't see behind the wall, the engine can "see" and it will still render everything that's behind the wall, this is the point we have been trying to make, you don't need to alter the gameplay of the map in any way, at all, just change how the walls are built.

[Archi]

Some more investigation.

Here we have 1078 sectors. All those sectors have the same floor and ceiling height and same flats. Basically the same sectors, but for some reason they aren't joined. Joining them removes around 7 thousand unnecessary lines. I've also tried to removes huge chunks on the north, that's just me experimenting. Usually you would want to get rid of static 0 height sectors simply because they don't block rendering visibility and look bad on automap.

 

Joining sectors, removing 8 thousand of decorations, 7 thousand lines raised performance on my pc from 9 fps to 19. It might be even playable at this state, but I did all the work the quick and dirty way.

[Pan]

45 minutes ago, Ancalagon said:

Guys we're getting there, keep trying!

 

 

Other people can explain with more technical details than me, but "blocking visibility" in this case we're talking about the engine. If a wall doesn't go from the floor to the ceiling it doesn't matter if you as a player can't see behind the wall, the engine can "see" and it will still render everything that's behind the wall, this is the point we have been trying to make, you don't need to alter the gameplay of the map in any way, at all, just change how the walls are built.

 

And my point is changing the walls will alter the gameplay. You can already see in the sections of the maze that are enclosed, the cacos can't fly over it and are instead forced to come down to maze-level. They are way easier to deal with in this way because they can't surround you from above and harass you from multiple angles at a distance. I could lower the entire sky and make all the walls in the map impassable, but it would become a totally different map. If I need to go to that much effort then I'd rather just work on a new map and apply the lessons.

[Pan]

1 minute ago, Archi said:

Some more investigation.

Here we have 1078 sectors. All those sectors have the same floor and ceiling height and same flats. Basically the same sectors, but for some reason they aren't joined. Joining them removes around 7 thousand unnecessary lines. I've also tried to removes huge chunks on the north, that's just me experimenting. Usually you would want to get rid of static 0 height sectors simply because they don't block rendering visibility and look bad on automap.

 

Joining sectors, removing 8 thousand of decorations, 7 thousand lines raised performance on my pc from 9 fps to 19. It might be even playable at this state, but I did all the work the quick and dirty way.

If you'd like to send me the .wad then I'll take a look and implement your changes. Joining sectors is easy. I didn't realise it would effect the frame rate so dramatically. 

[Archi]

I've made some bad decisions while trying to improve the map so I broke it. But you can fix it by yourself pretty quickly, select everything, then use Find & Replace Mode and then within selected find the ones with the floor height -392, ceiling height 128 and floor flat GRAY1 and ceiling flat F_SKY1.

[therektafire]

42 minutes ago, Pan said:

 

And my point is changing the walls will alter the gameplay. You can already see in the sections of the maze that are enclosed, the cacos can't fly over it and are instead forced to come down to maze-level. They are way easier to deal with in this way because they can't surround you from above and harass you from multiple angles at a distance. I could lower the entire sky and make all the walls in the map impassable, but it would become a totally different map. If I need to go to that much effort then I'd rather just work on a new map and apply the lessons.

Well you wouldn't necessarily need to change *all* the walls, you would just need to change some strategically picked ones that would divide the level into easily manageable chunks in terms of drawing/rendering/whatever you want to think of it as. Like for example the outer edge walls of the maze are a prime candidate for this since the player is almost never outside of there and all the action is supposed to take place inside so it would be perfectly ok to make them 1 sided especially since they are taller than the player already. Of course the other most optimal picks won't be so easy to find, you would need to take the time to run around to every part of the maze and spin around at every angle with the profiling tools on to see how much stuff is being drawn from any one position and try to make a configuration that minimizes unnecessary drawing while still allowing for enough of the original vision of the encounter design, aside from removing redundant sectors like @Archi suggests there isn't much else you can do to optimize this kind of map that I personally can think of, well I mean aside from making it in a format that can be run in faster ports lol. Then again I'm not exactly the greatest mapper myself, maybe Dragonfly or Dobu could give you some better suggestions 

[Pan]

58 minutes ago, therektafire said:

Well you wouldn't necessarily need to change *all* the walls, you would just need to change some strategically picked ones that would divide the level into easily manageable chunks in terms of drawing/rendering/whatever you want to think of it as. Like for example the outer edge walls of the maze are a prime candidate for this since the player is almost never outside of there and all the action is supposed to take place inside so it would be perfectly ok to make them 1 sided especially since they are taller than the player already. Of course the other most optimal picks won't be so easy to find, you would need to take the time to run around to every part of the maze and spin around at every angle with the profiling tools on to see how much stuff is being drawn from any one position and try to make a configuration that minimizes unnecessary drawing while still allowing for enough of the original vision of the encounter design, aside from removing redundant sectors like @Archi suggests there isn't much else you can do to optimize this kind of map that I personally can think of, well I mean aside from making it in a format that can be run in faster ports lol. Then again I'm not exactly the greatest mapper myself, maybe Dragonfly or Dobu could give you some better suggestions 

The optimisation fixes that need to be implemented will be more time consuming and will completely change the map. It's ultimately no worth it. I have found joining the sectors makes little noticeable difference to frps. Decorations and sprites are the major culprit. Try playing the map on a no monster run. I get a noticeable increase in smoothness.

 

If someone wants to beat this map then the prize is there. i7 computers appear to handle this map at a more-than playable level.

 

It is a real shame that there doesn't appear to be a way to convert UDMF into more open map formats.

 

GZ Doom should think about it's platform limitations for their next update. Registering a big map of uncomplicated pixels and sprites should be a main feature of any Doom platform (Pr Boom can do it). Whats the point of having a complicated platform like GZ doom if you can only experience it in small rooms and not on a grand scale? It seems GZ Doom is suffering from optimisation problems, not my map. UDMF (Universal Doom Map Format) is a misnomer. There's nothing universal about it.

As a previous poster mentioned, it is staggering to see how limited GZ Doom really is. It is disappointing because there's a lot about gzdoom that I like. It's like having an 8gb ipod and you want to put 16gb of songs on it. I feel a bit like that at the moment.

[therektafire]

11 minutes ago, Pan said:

I have found joining the sectors makes little noticeable difference to frps.

Ugh here we go again, and I thought we were actually making progress here :/ It might not make a difference for *you* but for Archi it did obviously, they said they got 10 more fps on average, that's definitely not "hardly any difference", that's a considerable difference. So you should join them anyway and update the map so that it will run better for people like me and archi who aren't so blessed to have an in your words "i7 crapbook"

[Pan]

37 minutes ago, therektafire said:

Ugh here we go again, and I thought we were actually making progress here :/ It might not make a difference for *you* but for Archi it did obviously, they said they got 10 more fps on average, that's definitely not "hardly any difference", that's a considerable difference. So you should join them anyway and update the map so that it will run better for people like me and archi who aren't so blessed to have an in your words "i7 crapbook"

 

Done! If it’s just a joined sector+ lower decorations heavy map that you’re after then give me a day or so. It doesn’t make a difference to my frps, but if helps you to compete then I’ll do it. 

 

I might add an extra archvile tho, because supposedly all the Doom Gods play on low spec laptops ;-) its a big prize so I think it’s justified..

 

the prize will stay at 200. Extra archvile highly likely (you’ll manage. Ancalagon finished the map with plenty of ammo leftover).

 

ill be back!

[Gaia74]

9 minutes ago, Pan said:

because supposedly all the Doom Gods play on low spec laptops ;-)

that hurt me :"v

[Pan]

if you think that hurt, wait until you get deep into Blighted  Moonbase 

 

i still don’t think Ancalagon has fully recovered from the experience..

[Nine Inch Heels]

1 hour ago, Pan said:

As a previous poster mentioned, it is staggering to see how limited GZ Doom really is. It is disappointing because there's a lot about gzdoom that I like. It's like having an 8gb ipod and you want to put 16gb of songs on it. I feel a bit like that at the moment.

The main problem is: You have no experience with building large and/or very monster intensive maps. People told you how that manifests itself in the form of performance issues they experienced, and you were in denial about it from the get go.

 

Moreover, you have been dishonest about how good the map runs on your system(s). You said both your laptops can run the map perfectly fine, and it turns out that you yourself actually noticed that certain changes make the map run better for you. BUT: If the map ran perfectly fine for you, you wouldn't be able to notice anything to begin with.

[Pan]

50 minutes ago, Nine Inch Heels said:

The main problem is: You have no experience with building large and/or very monster intensive maps. People told you how that manifests itself in the form of performance issues they experienced, and you were in denial about it from the get go.

 

Moreover, you have been dishonest about how good the map runs on your system(s). You said both your laptops can run the map perfectly fine, and it turns out that you yourself actually noticed that certain changes make the map run better for you. BUT: If the map ran perfectly fine for you, you wouldn't be able to notice anything to begin with.

That's true. But it was smooth enough for me to play comfortably and I didn't know I was using a comparatively high-end machine.

[Pan]

4 hours ago, therektafire said:

Ugh here we go again, and I thought we were actually making progress here :/ It might not make a difference for *you* but for Archi it did obviously, they said they got 10 more fps on average, that's definitely not "hardly any difference", that's a considerable difference. So you should join them anyway and update the map so that it will run better for people like me and archi who aren't so blessed to have an in your words "i7 crapbook"

Try this out: http://www.mediafire.com/file/k28ovyd8yzvfxc6/Blighted_Moonbase_IV-_Remastered.wad/file

Blood sprites reduced by 75-80%, blood pools replacing sprites. All sectors joined. Large chunks of background cut out + distorting distant light textures replaced.

I've added archviles and aracnorbs as your punishment. They have been added, sparingly, to the end fights. You wouldn't notice a difficulty spike unless you make it to the end stages of hell and the desert. Ancalagon's run inspired the slight tweaking of hell, which was a bit too easy for him, and the desert, which was definitely too easy for him. Everything else remains the same, including the removal of Player 177's hack of the final boss fight.

 

If you get improved FRPS then please let me know. If this version turns out to run better then it will replace the main map as the new competition map. I will wait for feedback

[Bryan T]

I get the same fps as before, 24.  Not sure why you need someone else to tell you if it runs better.

[Pan]

24 is playable I reckon ;)

[Guest Unregistered account]

I don't want to spark an argument, but if GZDoom can handle enormous open vistas filled with line horizon effects, particles, massive swarms of enemies and projectiles, 3D floors and 3D models, as can be seen in various combinations in many popular GZDoom WADs, with barely a hitch on most average devices, I think it's ludicrous to brush off these problems and blame it on the port when your map takes advantage of none of these features and looks closer to a Vanilla 90's PWAD, yet manages to run like sludge for multiple people.

This is entirely a problem with how the map is constructed, not the port nor peoples' computers.

 

Anyway, you've heard it already from multiple people, I won't bang on about it. Good luck with your future projects!

[wintertowns]

[Aquila Chrysaetos]

So I grabbed the "remastered" version and looked at it.

I know I won't be listened to here, but I might as well say it anyway.

I couldn't tell what was optimized. All I knew was that it was unplayable.

I performed all the other optimizations first before deleting the walls to reduce the rendering load and I noticed very little improvement. It was only when the walls were deleted that I got 25-30 fps (GZDoom 2.4.0, GL Renderer, 1024x768 resolution) and 40-60 fps (ZDoom32, software, 1024x768).

Now that I can play the wad, I think I can safely say that Ancalagon will probably be the only one to try this, not that it matters, because the goalposts will just be moved again anyway.

Alternatively, someone like MrZzul/Killer5 or Bloodite Krypto could do this, not that I'd expect them to want to.

[Pan]

10 hours ago, wintertowns said:

 

 

 

 

 

 

 

 

 

 

 

[therektafire]

Just when I thought this thread couldn't get any cringier...

[Ancalagon]

Doing these pics took around as much time as it would take to fix the map.