Photon Physics and Positional Spherical Mapping

Community Forums/Showcase/Photon Physics and Positional Spherical Mapping

TartanTangerine (was Indiepath)

(Posted 2005) [#1]

Bring em both together, add a bit of blitz max and you've got 2D rigid bodies with deforming textures...

This is a bit proof-of-concept for one of the new Indiepath Titles.

http://www.indiepath.com/tim/trubble.zip

big10p

(Posted 2005) [#2]

Looked really nice for the first 5 seconds, until there was >20 blobs and it slowed to a crawl. Good. organic feel to it, though. :)

TartanTangerine (was Indiepath)

(Posted 2005) [#3]

How slow did it go? what was the Millisecs reading? what is your spec?

big10p

(Posted 2005) [#4]

Well, the numbers aren't very stable but I get ~150ms with ~60 blobs.

Athlon 1Ghz, 256MB RAM, Geforce 2 GTS 32MB, Win98SE.

tonyg

(Posted 2005) [#5]

I only got as high as about 40 blobs which was taking about about 40ms with intenet connected.
30 blobs about 25ms
2700+, 9800Pro, 1G Ram, WINXP SP2

TartanTangerine (was Indiepath)

(Posted 2005) [#6]

Yeah I found it was taking just under 1ms per blob - need to optimise the phyiscs engine...

Grey Alien

(Posted 2005) [#7]

9ms 20 blobs, P4 3.2, 1GB RAM, Radeon 9800XT
How do you get more blobs?

TartanTangerine (was Indiepath)

(Posted 2005) [#8]

Heh you don't.

I've posted a new demo which is much faster. Click the mouse to make a new blob.

For me 20 (ish) blobs takes about 2ms.

p4 2.6Ghz, 1GB Dual-DDR Ram, Nvidia fx5200 Ultra.

Grey Alien

(Posted 2005) [#9]

weird, it's hard to stay around 20 blobs, but I get 3ms.

Mark Judd

(Posted 2005) [#10]

3ms with 20ish

P4 3.2Ghz 1.5GbRAM Radeon X300

Some of the blobs got stuck like the end of the slopes were penetrating (oo-err) the blob vertices :-

cheers

Funky Frank

TartanTangerine (was Indiepath)

(Posted 2005) [#11]

Yeah some blobs will stick, that's just the way it is. That won't happen in other versions as I'm impletmenting a point-in-poly and poly intersection routine, they'll pop when this happens.

SillyPutty

(Posted 2005) [#12]

this is very cool Tim

got 17ms with 20 blobs.

Picklesworth

(Posted 2005) [#13]

I get 5 million fps!
30 FPS with 180 blobs :(

Zenith

(Posted 2005) [#14]

Rofl,

http://electrodev.100free.com/insane.png

Hopefully the FPS is off? But it seems accurate when I get tons on screen.

4053499 FPS 0_O

TartanTangerine (was Indiepath)

(Posted 2005) [#15]

Actually thats the logic loop that it's reporting since the actual render is independent of cpu speed and monitor Hz. When the FPS dips below about 90 is when you start to see the actual true GFX render rate.

AdrianT

(Posted 2005) [#16]

on my athlon64 3200, geforce3 combo I need 230blobs on screen before I drop down to 30FPS

Thats impressive :) thainking of making something like that tar blob game that was so popular last year. Can't remember what it was called lol.

very cool squishy bubbles :)

TartanTangerine (was Indiepath)

(Posted 2005) [#17]

Heh, that game was called GISH. Na, I'm not going to make anything like that. I also think that this needs a lot more optimising before I move any further forward. If the ODE module for BMax is faster than this Proton code then I may continue with the development.

xlsior

(Posted 2005) [#18]

Nice... The bubbles are a bit on the square side though, and look crude -- no anti-aliasing?

Adding some smoothing and/or making them rounder would probably make it much easier on the eyes.