I have just added multitexturing (normal texture + lightmap) to my program, and it looks fine but it gets much slower (about 50% of the the former speed, I guess). As I use a gf 4 ti 4200 I thought it should be just as fast as with only one texture? Is there anything special one has to known? (I used glAcviteTextureARB(…), glMultiTexCoordARB(…) etc.).
got it fixed. thx to anyone who would have answered .
You should check the optimal size of lightmap! I had 256x256 lightmaps & later I decided to check the optimal size. I added code that allowed me to import lightmaps of size 256x256, 128x128, 64x64, 32x32, 16x16, 8x8 and dinamycaly switch among them. After small test I found out that !!!16x16!!! are allmost identical to 256x256! 8x8 was noticably worse. But after remaking my lightmaps I got + 8-10 fps on GF2GTS in app that runs at ~26avg, so eliminating swapping to RAM I guess.
BTW, in some cases you could use 128x2 lightmaps somtimes 32x32 is needed.
at the moment there is only one large lightmap that covers the entire area and is about 10000 x 10000 texels, but the program does not seem to slow down at all with multitexturing enabled.
i have the same problem with the frame rate.
maybe you could post your solution ??
i simply had too much opengl calls, i called glBindTexture… before every single texture coordinate passed. what is the problem with your program?
I had a lot 256x256 textures + cubemap 512x512*6 so on 32 MB GF2GTS it was bottleneck, now I have GFFX5200w128MB on table, I’ll check what it’ll change when I’ll get home