Sunder demos [-complevel 9]

[doomdaniel95]

Aw man, does anybody know if any of these runs are on YouTube or not? I guess my problem here is that for some reason I can't seem to play back too many Sunder demos.. they just desync for me. Would anybody know any possible reasons for that? Sorry if this should belong in techie questions or somewhere else.

[Ribbiks]

I know on the dsda page there are a bunch of demos marked as having been recorded on "sunder_old.wad" (anyone have a copy of that btw?), were those the ones that were desyncing?

[Memfis]

If there is a comment like "plays back with x.wad" on DSDA then x.wad should be inside the zip that you download when you click on the wad name.

[Ribbiks]

oh yeah, guess I'm just stupid, seeing as dsda is where I think I downloaded sunder from in the first place -.-

[doomdaniel95]

The ones I have trouble with are ones that don't have any comments under them such as "plays back with ...". They don't have those. I wonder if it's my old version of prboom-plus I'm using (2.5.0.2), which I don't bother to update since it works for me. I also tried redownloading the WAD from DSDA when I noticed that problem, but that didn't help either. It's very strange. I can play back demos on most other WADs (though I have trouble with a couple of cchest and cchest2 demos too). Like in the cchests, blob's demos will play back fine for me. However, I can't play back Tatsurd-cacocaco's Cchest1 MAP01 Tyson. It just desyncs. So do most of these Sunder demos I try to watch.

[Urban Space Cowboy]

doomdaniel95 said:

The ones I have trouble with are ones that don't have any comments under them such as "plays back with ...". They don't have those.

Be more specific. Which particular demo(s) is/are giving you trouble?

[Phml]

I think Sunder uses ZDoom extended nodes for several of the later maps, support for which was only added in 2.5.0.7. Map11 and Map14 won't run at all for me in 2.5.0.6.

[Cacatou]

MAP14 UV-Pacifist in 0:17

su14p017.zip

[doomdaniel95]

Urban Space Cowboy said:

Be more specific. Which particular demo(s) is/are giving you trouble?

All of Ribbiks' demos give me trouble. Odysseyofnoises' (Light_Speed's) MAP03 UV-Speed gave me trouble (but I saw that one on YT so it was ok). MAP05, MAP08, MAP12, and MAP13 maxes that are up (all of them) give me trouble. It's so weird.

[Ribbiks]

the following demos of mine were recorded without specifying complevel 9 because I was retarded and these were the first demos I've recorded and I'm sorry:

map01_pacifist
map02
map06_uvspeed
map09_pacifist
map10_pacifist

though I actually just tested, and verified that I can play all of these demos back without desyncing with complevel 9, using prb+ 2.5.1.4, and the version of sunder available from the dsda. So honestly I'm not quite sure from where your issues arise.

[Grazza]

Ribbiks said:

play all of these demos back without desyncing with complevel 9

You don't specify a complevel for playback except in exceptional circumstances. In most cases (such as this) the attempt to force a complevel that is incompatible with the lmp file will be ignored and prb+ will use the info from the lmp file itself.

[Ribbiks]

interesting, how is lmp/cl incompatibility determined? I could swear I've had cl2 demos desync by specifying cl9, or vice versa maybe. does it have something to do with 'extended' demo formats?

[vdgg]

An lmp file has a byte responsible for compatibility

http://www.doomworld.com/vb/showthread.php?s=&postid=809670#post809670

Also, after you started a playback you can quit PrBoom+ and check stdout.txt and find a line like one of these

G_DoPlayDemo: playing demo with Ultimate Doom/Doom95 compatibility
G_DoPlayDemo: playing demo with boom v2.02 compatibility
G_DoPlayDemo: playing demo with Doom/Doom2 v1.9 compatibility
G_DoPlayDemo: playing demo with early DosDoom compatibility

I used -complevel for playback in less than 30 demos I ever watched.
Also, don't be sorry about your first demos, I started with messed up complevels as well (in 2009) :)

[4shockblast]

Sunder Map 05 Pacifist in 0:11.

su05p011.zip

[Springy]

Map 02 no monsters in 0:57. Thought I'd give this another shot today and get 0:55 but kept getting 0:57 again so I give this up for now.

su02o057.zip

[Veinen]

Hello. I made a Map04 UV-Max in 14:05. It's bad. Goodbye.

su04-1405.zip

[Ribbiks]

I wonder why oku's 13:57 never made it onto the dsda. ohwell. Either way nice work :) certainly not bad.

[j4rio]

It's for old version, there's additional invul in last room.

[Veinen]

But wait! It gets worse. Facepalm-worthy platforming section... First attempt after I posted that 14:05 heh.

Map04 max in 13:52

su04-1352.zip

[4shockblast]

Map 01 UV-Speed in 1:42.

su01-142.zip

[doomdaniel95]

4shockblast said:

Map 01 UV-Speed in 1:42.

Only in Sunder do you see a UV-Speed with 95% kills, haha. Very nice run man!

[Khaoscythe]

Ribbiks said:

Map 12: The Zealous Machine, UV-Speed in 8:17


I kept dying on my UV-max attempts, so I made this in the meantime. I hate the invul placement in the last room, for some reason I always fall down there -.-

Ugh, as much as I dislike bumping old threads...

This can surely be optimized to ~3:30, although recording the actual demo is quite taxing (between maximizing damage on the 4 Cybers, the careful weaving through the Rev/HK/Caco mob as well as the damn block bridges).

The entire Rocket Launcher scenario (picking it up, etc.) can be skipped completely, you do not need to waste minutes killing the Revs. You can practically reach the chamber with the 4 Cybers at ~0:45, exiting that room by ~2:00 and reaching the final chamber by ~2:35.


Excuse the completely random post... just came back from an excruciatingly long hiatus. I'm still a ghost though, going to attempt optimizing Map 13 somehow. ><

[dew]

It is perfectly okay to bump threads in Doom Speed Demos. It is actually encouraged, because we'd merge new threads with the old ones anyways.

[Ribbiks]

Khaoscythe said:

This can surely be optimized to ~3:30, although recording the actual demo is quite taxing (between maximizing damage on the 4 Cybers, the careful weaving through the Rev/HK/Caco mob as well as the damn block bridges).

The entire Rocket Launcher scenario (picking it up, etc.) can be skipped completely, you do not need to waste minutes killing the Revs. You can practically reach the chamber with the 4 Cybers at ~0:45, exiting that room by ~2:00 and reaching the final chamber by ~2:35.

That's cool to hear. I recorded that demo spitefully because I just previously had a run that would've been a faster uv-max but I missed one monster. That map is a nightmare when it comes to keeping track of wandering baddies and straggling teleporters

I hadn't spent much time on route planning or anything I kinda just winged it and posted the first exit I got. It'd be cool to see such an improved run if you find the patience to record it :D

On a side-note Oku mentioned that he thought that map was also possible to pacifist. I gave it a few tries but I can't for the life of me manage to get through the hordes in the second to last room -.-. Would be fun to make a proof of concept tas demo maybe if it's doable.

[DeletedUser]

For MAP10 is their a specific rule for how many imps you must kill on the ledges, such as all the ones which are clearly visible must die for a max to count whilst the ones in the corner which are hard to reach are excluded? I noticed both j4rio and Okuplok exited with 7 monsters remaining yet I exited with 3 monsters on a practice run.

[dew]

Going by the rules, that invalidates their runs as maxes and makes your run the record despite being slower. Exceptions to this rule exist (pl27, mm11), because it's possible but extremely annoying to kill all the monsters in them. I don't have direct experience with trying to kill the imps on the map in question, but from watching the demos (and the demonstrative video on youtube) I assumed it's next to impossible to kill them. Was it a fluke occurence, or do you have a reliable (and sane) way to kill the extra 4 imps?

[j4rio]

the amount you kill is pretty much random depending on how many deaf imps get woken up by splash damage

[dew]

I understand that, but how reliable is it to wake the "extra" imps? Is there perhaps a spot from which you could consistently get more kills?

[j4rio]

I believe it's based on movement of imps. If they make favorable formations, splash wakes up more of them.

[Phml]

Are we talking about the imps hidden in the trees? Never tried to speedrun this, but I could swear I completed the map with all kills several times. It could be because I let the PEs infight so there's plenty of lost souls getting stuck near walls for convenient splash damage.