Ive just played the map, and its the best looking sof2 map I ever have seen, unfortunately u did not use any hint brushes, and I have seen some "loose" windows, I also didnt like the fog that much, the skybox was realy good tho.

I got 20/100 fps.
GF6800U / 1 Gig ram / 3200 XP AMD / 1280x1024

Plz, gimme the .map file I would like to make a competition version. ^L^

edited on May. 20, 2005 05:09 pm by ChrisZ

I know the original goal was to be quality, and you've achieved that, since your done, consider working with ChrisZ to improve it, as your next goal, it would be a real shame if you did all this work, for some comments in this thread...

As a show of good faith, I'm willing to lend my help aswell...

Make it playable, what do you say?

edited on May. 20, 2005 07:25 pm by PleaseMYOB

using fog and distancecull, should make the fps 10/20 higher. Using hint files should get some more FPS.

But try to make a smaller version like said before and keep the amoun of models to a minimum!

I even think its possible to make about 4 individual maps out if this. :)

Thanks all for appreciation and positive criticism !

All you said is true :D

I just have to repy that performance was not my objective, I said from the start.

This was my personal test of the... "SOF2 envirnment".

It prpbably could be made alot better than I did (max_pm is my very first mapping project), in the very nexy future the .map file will be released... it will be a challenge for mappers more experienced then me to make it something playable through the IP but I think this will be really a hard task men... at the moment, it's relaly fun and playable on a LOCAL 10/100 mbit network with several clients, tha's enough for me to say OBJECTIVE COMPLETED :D

PS: don't miss the RIDE-THE-ROVER (rover s the killingwhite ball) party by jumping on it to make a guided tour of the village! :D

PS2: in my P4 2.4 512 MB 10/100 mbit network it runs from 12 to 75 FPS... I don't think this will be tagged as UNPLAYABLE.. .. ok ok lte's say.. UNPLAYABLE THORUGH REMOTE CONNECTION then... :)



edited on May. 22, 2005 03:33 pm by massimagnus

For future referance only, by common standards, if you drop to 35 FPS or less, at any point, the map is unplayable ...

This is be entirely seperat from networking which is measured in "ping"...

Before you start your next project, I recommend reading some of these, they will shed alot of light on optimization...

Even JOR, SHOP & KAM5 go under 35 FPS in some cinrcumstances (lots of clients, open space, massive bombing).. so are these also to be tagged as UNPLAYABLE MAPS? I think no map ever made for SOF2 could be PLAYABLE if only this is the parameter that makes the difference between playable and unplayable :)

As I said, audio/visual has a price... I wouldn't trade off audio/visual for performance. But this doesn't mean I setup and built all without looking at performance.. all the caulkable was caulked, all alignable to grid was.. aligned to grid; all the non-solidable was tagged as non-solid and so on.

I spent the last 6 weeks optimizing, removing all the unnecessry, reducing entities. the problem is that Portmeirion is not based on a square grid of 128 gameunits and it is full of baroque stuff! That's why is it a so lovely place and not a suburbian deposit area (that perfectly fits the grid :D ).

Don't think I'm so stupid and unexperienced to have ignored what you suggest in your links. I'm a noob but.. an EXPERIENCED noob :D

How can you hint a so big openspace? And even if it could have sense hinting corners in the core of the village for gaining a bunch of FPS, what's the matter when the player is on the top of the bell-tower or far on the sea shore?

No hinting could be useful..

All is detail and most of the stuff is tagged non solid to set it apart form the BSP tree.

I can assure I've made all the possible to speed up things.

This map push the engine to its real limit: LARGE OUTDOOR AREAS.

So the mapper, starting a project, has only 3 choices if he doesn't want to make players sink down the 35 fps:

- to make a tiny map
- to make a large-but-watered and foggy map
- to make a big map full of caverns, rooms and passages

no one of the three option was interesting for me.

There are 1000s of those around and far better than max_pm to be played at 90 FPS online.

Everyway, thanks for the positive critisism, mates!

Your opinion, suggestions and notes are absoultely welcome.


edited on May. 24, 2005 01:20 am by massimagnus

As I said before, you don't have to defend yourself or your motives here, all is welcome...

Nobody is right/wrong, and you definatly learned alot to have this as your first map, so stop calling yourself a noob, nobodies buying it anymore...



On the same token, I've never seen a stock map drop under 35 FPS, even on my old geforce2MX when the game first cam out... And ofcourse it's not the only parameter to worry about...

And your absoluty correct, hinting will not help much if at all on your map... limiting access to the bell-tower and the far sea shore would, those are examples of why 1000's of maps adhere to those styles...

I still think that (vertical) hinting could be a great boost for performance, without any loss of quality, the fps in the tower will still be bad, but that would be one of the few spots imho...

I'm working on a lighter version.. I made several quick tests:
1) removed all the md3: result... a couple of FPS gain.. so MD3 are not the problem, all my models are NON SOLID shadered

2) removed all trains and triggered tricks: result NO GAIN AT ALL..

3) removed all speaker entities: result No GAIN AT ALL

This demonstrates that entities are not the problem.

Maybe FPS go KO for the complexity of BSP tree but I assure that all is caulcked and most of my additional shaders are NON SOLID.

You said vertical HINT will help, map is not hinted I'll try this (I to be honest I unseccessfully tried it in developing but probably was not in the right way.. ) but I dont konow how coul it help reducing BSP tree complexity the map is already partitioned in blocks by default blocksize that I changed to 2048. :?

DO you think that the ultimate useful test could be reducing drastically the farplane culling and fogging appropriately?

This helps to reduce tris drawing, but could it be useful to BSP tree too?

Keep in mind that there are more than 8000 brushes and 500 entities in this map

edited on May. 28, 2005 02:38 am by massimagnus

edited on May. 28, 2005 02:39 am by massimagnus