Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!van-bc!ubc-cs!alberta!edson!news From: jpenne@ee.ualberta.ca (Jerry Penner) Newsgroups: comp.sys.apple2 Subject: Re: Stellar 7 re-release Message-ID: <1990Dec15.073350.20404@ee.ualberta.ca> Date: 15 Dec 90 07:33:50 GMT References: <9012120353.AA26625@apple.com> <3137.27692c59@miavx1.acs.muohio.edu> <41464@ut-emx.uucp> Sender: news@ee.ualberta.ca Organization: University of Alberta Electrical Engineering Lines: 57 In article <41464@ut-emx.uucp> ifar355@ccwf.cc.utexas.edu (David H. Huang) writes: >In article <3137.27692c59@miavx1.acs.muohio.edu> rlcollins@miavx1.acs.muohio.edu (Ryan 'Gozar' Collins) writes: >>A 1Mhz GS faster than a 4.77MHz IBM, this I find hard to believe. If it is [stuff about 4.77 MHz ibms being fast] > >In _Programming the 65816_ by Lichty and Eyes, there is a comparison of >the speeds of a 4Mhz 65816, a 4.77(?)Mhz 8088, a 4Mhz 6502 and a 8Mhz 68000. >The tests were conducted by running an assembly language version of the >sieve of erastosthenes. I forgot the exact results, but the 65816 was more >than 2 times faster than the 6502, and about 3(?) times faster than the 8088. >The 68000 was 4 times faster than the 65816. > >If someone has the book, maybe they could post the exact results. If my memory >is correct, though, it would seem that a 1Mhz 65816 would be faster than a >4.77Mhz 8088. > >David Huang | From the book "Programming the 65816" page 298 Sieve of Eratosthenes calculating the primes from 1 up to 16383 ten iterations of the program. 65816 @ 4 MHz 1.56 seconds 6502 @ 4 MHz 3.12 8088 @ 5 MHz 4.0 8086 @ 8 MHz 1.90 68000 @ 8 MHz 0.49 65816 @ 8 MHz 0.78 (so says the book) The time for the 8088, 8086, 68000 were from Jan '83 BYTE according to the book. I'm assuming BYTE meant 4.77 MHz for the 8088 when they said 4.0 seconds, but we'll give them the benefit of the doubt @ 5 MHz (just makes it slower :-) The book also notes that since you're coding in assembler, anything goes. I would say their code is pretty tight. ------------------ Now, using the Sieve to determine a computer's speed is rather bogus. Most people don't run around using Sieves all day. But, for stuff like text processing, a 4.77 MHz PC is slower than an apple IIe at 1 MHz. Try doing DIR on the PC side by side a IIe doing CAT. The PC is slower. For stuff doing multiplications, the PC will probably win. The 8088 & 8086 had multiplication instructions, though they took over 100 clocks to execute. Now the 386s and 68030's can do mults & divs in ~1-4 clock cycles. Also note that on 8088's it took 4 clock cycles to do a memory operation; the 6502 & its cousins do those in 1 clock. -- ------------- Jerry Penner alberta!bode!jpenne Edmonton, Alberta, Canada