Well, that idea of moving away from Quads appears to have sped things up but I haven't noticed a marked improvment ( hopefully he sent the right updated) client file). All i can tell you is that I cap my frame rate at 70fps and it used to drop to 20fps ( exactly ) with a 50particle explosion. Now with the blast cut to 5 particles the drop is from 70 to 40fps. Which is still significant. BUT one of our other test machines reported the same drop to exactly 20fps even though is was a PII 450 with 256Mb and 32Mb card (unknown brand). So it's odd how it stops at 20fps on both machines.
But the effect is more noticable when the particles blend/overlap, almost as if the transpareceny rendering is what is slowing it down. 4 sachel charges using sprites down't casue any issues normally for me. But then using 3d explosions with the current standing of 5 particles each the frame ratedrop to 20fps for a brief period. But if you spread the packs out and detonate and the blast effects don't intersect then the drop in frame rate is as if only one was blown up. If you need any more details feel fre to ask. I'll remember to add you name to our credits' thnx list when this is released next year - even if it doesn't help. ----- Original Message ----- From: Miguel Aleman <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Sunday, May 12, 2002 12:38 AM Subject: Re: [hlcoders] triApi speed problems > How about some more information about the frame rates of the slower system. > Does it run over 50fps with the particles on screen? > > ----- Original Message ----- > From: "Chris 'Tal-N' Blane" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Saturday, May 11, 2002 6:11 PM > Subject: Re: [hlcoders] triApi speed problems > > > > Read the other post for more details, why would this effect one system of > a > > higher spec 1.1Ghz 512Mb Ram with the same gfx card but not a far slower > > one? > > > > > > ----- Original Message ----- > > From: Miguel Aleman <[EMAIL PROTECTED]> > > To: <[EMAIL PROTECTED]> > > Sent: Sunday, May 12, 2002 12:11 AM > > Subject: Re: [hlcoders] triApi speed problems > > > > > > > On some videocards, Quads are extremely slow due to the way they are > > > calculated. Use triangle strips instead of quads when possible (like in > a > > > particle engine). > > > > > > ----- Original Message ----- > > > From: "Chris 'Tal-N' Blane" <[EMAIL PROTECTED]> > > > To: <[EMAIL PROTECTED]> > > > Sent: Saturday, May 11, 2002 6:03 PM > > > Subject: Re: [hlcoders] triApi speed problems > > > > > > > > > > Just a comment from me. I'm the prioject leader, this is the particle > > > system > > > > kindly donated to us by friends in Hostile Intent. Just the engine > mind > > > you, > > > > not the implementation. The effect ISN'T consistant. On a PIII 350 > with > > > GF2 > > > > MX it runs fine, yet on some other systems ( mine in particular) which > > is > > > a > > > > higher spec and the same gfx card and drivers the frame frate drops by > > > about > > > > 50fps when a single one of these explosions occours. At first it use > to > > > > occour even when the blast couldn't be seen but we've gotten around > then > > > so > > > > that while there's a slight drop it hardly noticeable. This suggested > to > > > us > > > > that it' wasn't graphics related since the effect wasn't on screen and > > > > wasn't even in the same leaf or the rendering pipeline. > > > > > > > > But the point is that we need to know if it's a duff particle engine > > which > > > > isn't properly handling the effect when blending occours or if the > > > functions > > > > are being called quite right. > > > > > > > > Any hints or suggests are most welcome not matter how small, we've > done > > as > > > > much as we're aware of to avoid this serious frame rate issue. > > > > > > > > > > > > ----- Original Message ----- > > > > From: Frazee <[EMAIL PROTECTED]> > > > > To: <[EMAIL PROTECTED]> > > > > Sent: Sunday, May 12, 2002 12:00 AM > > > > Subject: [hlcoders] triApi speed problems > > > > > > > > > > > > > This is a multi-part message in MIME format. > > > > > -- > > > > > [ Picked text/plain from multipart/alternative ] > > > > > I have a particle system for my mod, which is really quite nice. > > Problem > > > > > is, as soon as multiple particles start blending it slows down. For > > > > > instance, with explosions I have cut it down to 5 particles (yeah, > 10 > > > > > polys). But that is still slow on some systems. Not to mention when > > you > > > > > combine explosions. I have optimized it all I can, and decided to > come > > > > > to you guys. One thing I didn't try is 2 triangles rather than the > > quad. > > > > > if that will help I will try. Can you just tell me how to get the > most > > > > > juice out of this thing? Thanks. > > > > > -- > > > > > > > > > > _______________________________________________ > > > > > To unsubscribe, edit your list preferences, or view the list > archives, > > > ple > > > > ase visit: > > > > > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > > > > > > > > > > > > > > > > > _______________________________________________ > > > > To unsubscribe, edit your list preferences, or view the list archives, > > > please visit: > > > > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > > > > > > > > > > > > > _______________________________________________ > > > To unsubscribe, edit your list preferences, or view the list archives, > > please visit: > > > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > > > > > > > > > _______________________________________________ > > To unsubscribe, edit your list preferences, or view the list archives, > please visit: > > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > > > > > _______________________________________________ > To unsubscribe, edit your list preferences, or view the list archives, please visit: > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
