Mail Archives: djgpp/1993/08/22/03:17:24
Some of the benchmark results really makes one wonder how performance can differ so much.
First, here is the benchmark result for a 386/40 With an ATI Ultra
Plus, ISA bus at 13MHz.
CPU & Video Benchmarks 1.0 (C) 1993 DJ Delorie. Distribute freely.
CPU: 386, 40 MHz Video: ATIUltraPlus
All results are Kbytes per second, s=system RAM, v=video RAM, mode 0x13
bits s->s s->v v->s v->v s-> v-> ->s ->v
8 5,386 2,307 1,365 965 6,665 1,412 6,305 2,608
waits 3.4 13.3 25.3 37.4 0.0 22.3 1.3 10.3
8 /8 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
16 9,634 4,613 2,679 1,930 13,331 2,825 13,289 5,217
waits 4.3 13.3 25.8 37.4 0.0 22.3 1.0 10.3
16/8 1.7 1.9 1.9 2.0 2.0 2.0 2.1 2.0
32 18,307 6,039 3,357 2,298 26,662 3,496 26,313 6,520
waits 4.7 22.5 43.6 65.6 0.0 39.7 1.0 19.5
32/8 3.3 2.6 2.4 2.3 4.0 2.4 4.1 2.5
Looking at the s->s column, these results place the 386/40 at a better
performance than some 486 processors, even 66DX2 ones. But a group of
486 machines also have up to twice the s->s speed.
The s->v of 6MB/s puts the machine/card just below the 486dx66 machines
(with local bus?). Running the ISA bus at 13MHz seems to do wonders for
this benchmarks performance.
Some machines appears to be lousy performers. For example, one
486sx/25 appears to have higher performance in s->s/16 than s->s/32.
Could someone write something about how the different CPUs should
perform. For example, what should the s->s speed be for 0WS for a 386
cpu, a 486 CPU and a 486DX2 CPU? How is it with the theoretical speed
for the ISA bus, and the LB?
--
-- Per Lindqvist pgd AT shakti DOT bbt DOT se
- Raw text -