Message-ID: From: Bryan Murphy To: "'djgpp AT delorie DOT com'" Subject: Weird Results Tesitng Allegro's Performance Date: Tue, 10 Jun 1997 23:49:00 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Precedence: bulk I threw together a quick test tonight to figure out exactly what kind of performance I could expect to get out of allegro for my computer. What I have is: Cyrix 686 P150+ Chip 16meg 70ns Ram (not EDO) 82430HX Chipset and most Importantly: Creative Labs 3D Blaster Now, not expecting to be wowed at all by my 3D Blaster (which I have full intentions on replacing soon, it sucks), I created a test program to figure out the max number of blits I could make each second. I was shocked by the results. Not at all what I expected. Here are the results under Windows 95: Tested: 640,480 for 20 seconds VGA Frames = 53.9 - Expected slow, but not this slow Vesa 1.0 Frames = 133.6 - Believable Vesa Banked Frames = 155.4 - A little faster, expected this Linear Frame Buffer = 114.1 - Whoa, did not expect this Why is the linear frame buffer SLOWER than the banked? Weird, so I tried it a few more times (only one shown): Tested: 640,480 for 20 seconds VGA Frames = 53.85 Vesa 1.0 Frames = 143.25 Vesa Banked Frames = 154.8 Linear Frame Buffer = 114.5 Yeah, it's definately slower. Very strange. Anybody know what can account for this? I suppose it's just crappy design on the part of Creative Labs. I do also have the most recent bios on the card. Ok, none of this really has anything to do with DJGPP and Allegro. So why am I posting this here? Well, with the Vesa 2.0 Banked Emulation mode, the bitmap was only blitting to the upper 20% of the screen, and the rest remained blank. Shawn, if you want me to help you track down another bug, I can give it a shot. I suspect, however, once again that it's just crappy design on the part of Creative Labs. I'll try it with UniVBE when I get a chance and report back if that fixes the problem.