mel (04/07/83)
I ran the sieve.c program that came with BDS C on the Rainbow and on a Morrow
Micro Decision 4mhz Z80 system (about half the cost of the Rainbow) and on the
8088 side of the Rainbow with the Computer Innovations C86 compiler.
Morrow Z80 Rainbow Z80 Rainbow 8088
cc 9.6 sec 12.9 (1.3) 60.2 (4.7)
ld 11.3 14.1 (1.2) 35.2 (2.5)
load 1.7 2.7 (1.6) 2.7 (1.0)
run 15.1 17.0 (1.1) 17.9 (1.05)
total 37.7 46.7 (1.2) 116.0 (2.5)
object size 4K 4K 14K (3.5)
The Rainbow Z80 seems to be about 20% slower than the Morrow, and the Rainbow
8088 seems to be a whole lot slower, except perhaps for arithmetic. The total
compile and load time for a large program (othello.c) was 82.0 sec for the Z80
and 319.8 (3.9) for the 8088. Mel Haas , houxm!melstevel (04/09/83)
#R:houxm:-36900:ima:16900002:000:854
ima!stevel Apr 8 10:44:00 1983
This seems to be more of a reason to aviod the Computers
Innovations C86 compiler rather than a comparison of the 8088
side of the Rainbow to anything. The 8088 loaded and ran things in about
the same amount of time. But the compile times are horendous.
Looked at another way the numbers are.
Morrow Z80 Rainbow Z80 Rainbow 8088
cc + ld 20.9 sec 27.0 95.4
load + run 16.8 19.7 20.6 (1.0)
total 37.7 46.7 (1.2) 116.0 (2.5)
While the Rainbow is still not impressive it shows that slow
software can make a snail out of any machine.
Steve Ludlum
decvax!yale-co!ima!stevel, ucbvax!cbosgd!ima!stevel, research!ima!stevel
PS: BDS C is a really good piece of software maybe in a few
years something as good will be available for the 8088.