Table 4-4 shows the UFS, AFS, and FSEM results for different versions of Windows when 80 FPS are requested. Table 4-4. Average FPS/UPS rates for the AFS, UFS, and FSEM versions of WormChase when 80 FPS are requestedRequested 80 FPS | AFS | UFS | FSEM |
---|
Windows 98 | 75/83 | 70/83 | 81/83 | Windows 2000 | 20/83 | 18/83 | 60/83 | Windows XP (1) | 82/83 | 77/83 | 74/83 | Windows XP (2) | 75/83 | 68/83 | 83/83 |
The numbers send mixed signals and, in any case, the sample size is too small for strong conclusions. Nevertheless, I'll make a few observations: The use of additional state updates to keep the UPS close to the requested FPS is an important technique for giving the appearance of speed even when the rendering rate is sluggish. FSEM offers better frame rates than UFS, sometimes dramatically better. However, FSEM's benefits rely on MS Window's access to the graphics device via DirectDraw. The improvements on Linux, Solaris, and the Mac OS may not be so striking. AFS produces higher frame rates than UFS and may be a good choice if full-screen exclusive mode is unavailable. All the approaches supply good to excellent frame rates on modern CPUs (the Windows 2000 machine sports a Pentium II). Consequently, the best full-screen technique for a particular game will probably have to be determined by timing the game. Additional optimization techniques, such as clipping, may highlight the benefits of one technique over another. | I'd like to thank two of my students, Patipol Kulasi and Thana Konglikhit, who helped gather the timing data used in this chapter and Chapter 3. |
|
|