jdb@mordor.UUCP (John Bruner) (11/21/85)
>I didn't really want to get dragged into this, but a comparison of the sizes >of executables (using size(1), and only adding up .text) of the stuff in >/bin and /usr/bin on a 68k UNIX (Sun-2) versus a 286 UNIX SYS V shows that the >286 binaries are only 65% of the size of the 68k binaries. I think Brad's >argument *is* valid. While the 68010 may not achieve the code density of a 286, the difference you are seeing is greatly exaggerated. It would be much fairer for you to compare two System V ports. Sun's executables are much bigger than they should be. (As an experiment, try compiling and linking the program "main(){}" and see how big the executable is. Link it "-N" to avoid page-alignment of data and bss which makes text and data appear larger than they really are.) -- John Bruner (S-1 Project, Lawrence Livermore National Laboratory) MILNET: jdb@mordor [jdb@s1-c.ARPA] (415) 422-0758 UUCP: ...!ucbvax!dual!mordor!jdb ...!seismo!mordor!jdb
lotto@talcott.UUCP (Jerry Lotto) (11/22/85)
Who thinks that the 8088 was chosen for it's architecture? I would be surprised if that was the main consideration. The "task force" that designed the machine was concerned with time to market. At that time, I doubt IBM would have seriously entertained the possibility of establishing UNIX as a microcomputer standard and they did not want to cut the profitability of their minicomputer line with smaller compatible systems. The only thing to do was to write a new op sys or take a little known existing system and promote the heck out of it. -- Gerald Lotto - Harvard Chemistry Dept. UUCP: {seismo,harpo,ihnp4,linus,allegra,ut-sally}!harvard!lhasa!lotto ARPA: lotto@harvard.EDU CSNET: lotto%harvard@csnet-relay
zben@umd5.UUCP (11/23/85)
In article <544@talcott.UUCP> lotto@talcott.UUCP (Jerry Lotto) writes: >Who thinks that the 8088 was chosen for it's architecture? >I would be surprised if that was the main consideration. >The "task force" that designed the machine was concerned with >time to market. At that time, I doubt IBM would have seriously >entertained the possibility of establishing UNIX as a microcomputer >standard and they did not want to cut the profitability of their >minicomputer line with smaller compatible systems. The only thing >to do was to write a new op sys or take a little known existing >system and promote the heck out of it. What makes you think technical considerations had *anything* to do with it? Is our memory so short we forget the big cross-licensing agreement that IBM and Intel signed ten years ago? Intel got access to all IBM's magnetic bubble work (:-) and IBM got access to the 80xx design and masks. And I think a hefty piece of Intel too. The games IBM plays with its suppliers can be played with IBM too. Much nicer to have a tame almost-in-house supplier than to have to worry about Motorola or National Semi changing things out from under you... -- Ben Cranston ...{seismo!umcp-cs,ihnp4!rlgvax}!cvl!umd5!zben zben@umd2.ARPA