Server Help

[L.C.]

First I shall explain what I am planning or working on doing here (which involves testing before launching to make sure it would work in the first place).

I plan on making a duplicate of Jackpot SVS as Jackpot SVS (Bots), and it will feature 128 AI bots evenly spread out across the map.

Now this is the question I can't easily decide on (all the options seem pretty interesting in my opinion ): what kind of game mode should this run on? Here are the options I have:

AI versus Human = Co-op Survivor/Deathmatch Battle
2 Frequency War -- War of the Factions
8-Player Frequencies = Massive Gang War
Free for All = Utter chaotic, non-stop 24/7 action; deathmatch ultimatum (but with the ability for human players to form their own teams).

[Cheese]

ffa or ai vs human
_________________
SSC Distension Owner
SSCU Trench Wars Developer

[Samapico]

128 AI bots?
Good luck with that :/ I never was able to keep 1 running correctly lol
_________________
(Insert a bunch of dead links here)

[L.C.]

@Samapico Subgame's hardcoded maximum for players is 1024 and I proved it to myself and other people (some who were curious to see it themselves); with the help of a buddy (programmer) I created the configurations and spawns, which wasn't as much work as it would have been to manually (by hand) go and configure for 1024+ bots. I cannot say who "sponsored" or aided, but one of the SSC hosts (out of his own curiousity) ran my test again.

With all 1020+ idling in spectator mode, Subgame can actually handle it pretty well! Even my latency wasn't even effected. MERVBot ran at about 1% per client, and there were 23 MERVBot clients running. In our tests we did find some interesting bug/glitch regarding Public Arenas and the maximum number of players allowed in an arena. :X

In fact, I still have all the files from this test on my backup DVD that I have here. I even have screenshots somewhere with proof of me running everything on my own machine and being in SS.

The idea for a next-gen test would be to try and hit a 1020 to 1023 mark (leaving room for a human player so that they could witness it all) of AI bots and see how well Subgame could handle that.



I can almost guarantee you that Subgame will be able to handle it. As for running everything off the same machine -- well I'm running offline/LAN tests before I will take this live.

Phase 1 - Minimal Testing (Running a few AI bot cores to make sure they work as expected) [Complete]
Phase 2 - Full-throttle Testing (Running all 64 bot cores / 128 AI bots to see how it works out; I have 3 computers at my hand to use if I want to cluster the hosting) [Pending]
Phase 3 - Upload (Did I mention this whole package or setup is 360MB?) [Pending]
Phase 4 - Server setup (Includes CHMODing and clustering if necessary)
Phase 5 - Bootup of Serverware



Now depending on the offline/LAN test results, I will either host it all off of Warpath or cluster. If I need to cluster, I've got some additional servers through connections that I can get the bots clustered on (ie. hosting 16/32/48/64 bot cores per server while the zone will always be hosted off of Warpath). Hopefully clustering won't be necessary, but it's a possibility.

The reason I'm doing this is for the public curiousity of it. I want to see what other people thing, how people react to it, and how the gameplay would play out. It will only be available through warpath.hlrse.net directory server for now though (because I don't know if it would violate any rules). Massive multiplayer is the whole concept I'm trying to introduce...to a single arena.

At the moment I'm temporarily skipping Phase 2 and going to Phase 3 (which I will begin within an hour or two). Going to hook my PC up to 8mbit internet with a 64 KB/s upstream speed so that I can upload the RAR for this package (which fortunately weighs at 121MB over 360MB); I have to go anyway because my cousin is having a birthday (and I have to tag along with my family).

Right now the settings are setup for FFA gameplay.

[Samapico]

I'm not worried about subgame; I'm worried about the bots being laggy... But we'll see I guess

[Bak]

I think it's a very good idea... but I suppose being around subspace long enough you become somewhat of a pessimist:

You're missing phase 1.5: give up

that out of the way, it would work much better with fake players in asss, easier to manage too.
_________________
SubSpace Discretion: A Third Generation SubSpace Client

[L.C.]

Bak, your reply made no sense.

EDIT :: Fake players aren't the same as AI bots. Fake players just listen off of a loop/scripted loop. You should know what an AI bot is and I won't explain it. It cannot be fake players because that's crap compared to AI bots.

Anyway. Phase 2 failed. I might have to rework my plan, simply continue, or compromise some elements of the plan.



My computer couldn't handle it, even though it can handle 1024 idling MERVBot clients with 100% stability. Bjorn's bot isn't efficient (and perhaps it might also have a memory effect on Subgame, but it wasn't showing out during the testing).

I have an AMD Athlon XP overclocked by 1GHz (clock is 2GHz) and 1.5GB of PC3200 DDR RAM.

EDIT2 :: And please don't get into the business of telling me to use AS3 if you aren't willing to help. I would love to use AS3, but its current state is too poor for me to be able to do so (and of course it just looks like I'm simply shoving it out of my face simply because AS3 is AS3 rather than looking at the facts, which I already have looked at and is simply why I am doing this). I'm not going to use incomplete material and I'm not going to make a duplicate of that last AS3 thread I made.

[Cerium]

I think you need to do more research before you start telling people what does and doesn't work.
_________________
There are 7 user(s) ignoring me right now.

[L.C.]

Sorry about that. At the time of writing I was about to have a mental meltdown, so I made some rather rash, irrational and stupid comments/replies.

I'm fine now...after taking a nap to shake it all off.