🚀 go-pugleaf

Hey all,

I had the oportunity the test a Quadra 700 with AUX (Apple Unix), X11 and
gcc 2.2.2 against a 486 33MHz with Linux 0.99, X11 and gcc 2.2.2. I would
gues that this is fair and comparable test, and the closes enviroment you can
get for this two maschines. The program I tested was an MPEG decoder for X11.

Now the results,

with display

Quadra 700 with AUX 3.0 and gcc -O2  5.25 frames/sec.
486 33MHz with 256KByte secondary cache Linux 0.99 and gcc -O2  6.1 frames/sec.

without display (to mesure the raw processor power)

Quadra 700 with AUX 3.0 and gcc -O2 6.0 frames/sec.
486 33MHz with 256KByte secondary cache Linux 0.99 and gcc -O2 7.87 frames/sec.

The Free MPEG controler was Version 1.2 and the file flight.mpg.

The conclusion:

The 68040 and 486 are really close in performace for the same clockrate. This
isn't unpredictable because Motorola and Intel say that its processors have 1.3
cycles per instruction over all performace. This is shown in the second benchmark
her you see that the 486 33MHz is 1.32 times faster than the 68040 25MHz.(but the
486 gains it performace only with the secondary cache were as the 68040 doesn't
have one). The first benchmark shows that the videohardware in the Quadra is
better than the et4000 SVGA ISA card. If you use a localbus card this will give
you a better performace, but melt the price differnce between 486 and Mac Quadra
700. The influence of the secondary cache in 486 is the next thing I like to find
out. So wait for someting more, and also math performace is something to test.

Rainer

email: menes@statistik.tu-muenchen.de

P.S: Comparing Windows 3.1 against System 7.1 for the Mac is a differnt question.
I personally love the Mac system more than Windows 3.1 and a PC, for objective
and subjective resonse.