Pictorials

Dark Basic
Speed Testing

Hardware:

Processor: Pentium II 350Mhz
Ram: 192M SDRAM(100 Mhz)
Harddrive: Maxtor 10G 10kRPM
Graphics: Viper 550 16M AGP

Software:

Dark Basic Version 1.04
Windows 98b

Criteria:

Display Mode 640x480x16
All Textures 512x512x8

Camera Range Test

Heres are some visual results from my Dark Basic Speed Testing. The pictorial over to the left is a render scene with 500 objects 200 Shown | 300 Hidden (error in display).

With a camera range of 3000, 30 FPS was achieved with camera in static position.

A significant FPS gain was achieved when camera range was set to minimum range of 255 with needed objects in view. In a static position 48 FPS was achieved.

TEXTURE OBJECT vs. The COLOR OBJECT Command

In this test, I discovered significant FPS gains when objects were textured vs colored. To the right, this scene is rendered using the COLOR OBJECT command. All colors were generated random. However, with the use the one color, no FPS gains were noticeable, leaving us with the result of 13 FPS.

Textured objects resulted in a significant gain of 33 FPS. This may be dependent on the video card. This suggest the use of solid colored textures if solid colors are needed.

Texture Size and Filter Test

In these test I experimented with texture size, filtering and scaling. The texture quality is porportional to the size of the texture use. The image to right displays a series of textures mapped to 3D Plain Objects. Sizes 64x64, 128x128, 256x256, and 512x512 respectively. The Filter Flag was set to 1 for each object.

Smaller textures 64x64 or less result in severe Pixelation when scaled regardless of Filter Flag. If such effect is desired. I is recommend to set the Filter Flag to 0. The smoothest scaling can be achieved with larger Textures of 256x256+

Large Level Object Test

In this test I experimented with possibility of the loading one large level object. With an average rendering 50 FPS! I can see potential uses.