ericmcg@pnet91.UUCP (Eric Mcgillicuddy) (10/09/90)
>>>>>>>>>>> from Doug Gwyn >GSOS is a 32- bit OS the likes of which Mac users will not see until >Jan. '91. Unfortunately it sits on top of an 8-bit FST. Operating systems are not measured in such terms. (If you think they are, then tell me how many bits UNIX "is".) The main problem with the ProDOS FST, whish is the only one that GS/OS supports at present that is suitable for general disk file storage, is that it has a limit of 32MB per volume (logical filesystem). That is due to the data format used in >>>>>>>>>>>>>>>> You are corect, but when I think of Prodos I think of 8-bits, the FST has nothing to do with word sizes. I do refer to GSOS as a 32-bit operating because The MacOS is referred to as a 24-bit operating system (perhaps incorrectly) with System7 providing the 32-bit version. In this context the OS has little or nothing to file management, but rather with memory pointers and block sizes. GSOS allows one to create memory blocks up to 4Gb anywhere within a 4Gb address space. This is what I meant by 32-bits. I believe this is true of UNIX, but there are so many versions that the 'standard' may be different, what is the largest memory that can be gotten with 'malloc()'? The 65816 can of course only access 24-bits of the 32 bit address space and the toolbox routines are a bit sloppy with the remaining 8, you can't guarantee that they will not be changed by a toolcall. But the potential is there and just needs someone to tap it. Anyone else have anything positive to add about the Apple II? UUCP: bkj386!pnet91!ericmcg INET: ericmcg@pnet91.cts.com
whitewolf@gnh-starport.cts.com (Tae Song) (10/16/90)
Huh, what?!? The 65816 can only address 24-bits directly. In the GS, Apple limited that to half which is 23-bits. 24-bits is 16M and 23 is 8Ms. Just how did you manage to get 4 GigaByte???
toddpw@tybalt.caltech.edu (Todd P. Whitesel) (10/17/90)
whitewolf@gnh-starport.cts.com (Tae Song) writes: >Huh, what?!? The 65816 can only address 24-bits directly. In the GS, Apple >limited that to half which is 23-bits. 24-bits is 16M and 23 is 8Ms. >Just how did you manage to get 4 GigaByte??? Simple. GS Applications are required to treat any pointer value as a 32 bit quantity, with the memory available to the current hardware starting at 0. The hardware can only access 24 bits / 16 megs but applications that follow Apple's guidelines (plus some of the tools will barf on non-zero high bytes) will still work if the high byte ever becomes significant. The Mac O/S did not formally place this restriction on applications until the '32 bit clean' coding standard was introduced, and any program that doesn't obey it will break under system 7. In this sense, GS/OS is a 32 bit / 4 gig operating system although there are plenty of instances where user code would break if it were located at a true four byte address. Todd Whitesel toddpw @ tybalt.caltech.edu
jh4o+@andrew.cmu.edu (Jeffrey T. Hutzelman) (10/17/90)
Todd,
Virtually all user code that uses long addressing would break
if moved to a system with a 32-bit address bus. The 65816 long
addressing modes use 3-byte pointers. That means that any code
that uses long addressing will have problems with anything that
resides above the first 16M of memory. Besides, who needs a system
with >16MB memory?
-----------------
Jeffrey Hutzelman
America Online: JeffreyH11
Internet/BITNET:jh4o+@andrew.cmu.edu, jhutz@drycas.club.cc.cmu.edu,
jh4o@cmuccvma
>> Apple // Forever!!! <<russotto@eng.umd.edu (Matthew T. Russotto) (10/18/90)
In article <Mb77_tG00awOA3Sh4z@andrew.cmu.edu> jh4o+@andrew.cmu.edu (Jeffrey T. Hutzelman) writes: >resides above the first 16M of memory. Besides, who needs a system >with >16MB memory? There's supposed to be a smiley there, right? I seem to remember the same statement being made by a certain prominent Apple person, except that it was about 16K, not 16M... -- Matthew T. Russotto russotto@eng.umd.edu russotto@wam.umd.edu .sig under construction, like the rest of this campus.
alfter@uns-helios.nevada.edu (SCOTT ALFTER) (10/18/90)
In article <Mb77_tG00awOA3Sh4z@andrew.cmu.edu> jh4o+@andrew.cmu.edu (Jeffrey T. Hutzelman) writes: >resides above the first 16M of memory. Besides, who needs a system >with >16MB memory? Hmm...UNIX types who might want to run a bunch of processes at once? You shouldn't be trying to dictate limitations like that; it's the lack of those limits that has made the Apple II what it is. BTW, I've worked with a machine with more than 16 megs. The CS class I took last semester used a Sequent Symmetry mainframe with 52 megs of RAM. (I will concede that the memory was split among six 386s, though, so each processor would have less than 16 megs tied to it, but who's watching? :-) ) Speed might be one reason: real memory will always be faster than virtual memory. If you think 52 megs is a lot of memory, I think the Connection Machine has either 2 or 4 _giga_bytes--I'm not too sure which. ----------------------------------------------------------------------------- Scott Alfter _/_ / v \ Apple II: Internet: alfter@uns-helios.nevada.edu ( ( the power to be your best! GEnie: S.ALFTER \_^_/
jh4o+@andrew.cmu.edu (Jeffrey T. Hutzelman) (10/19/90)
OK, maybe there should have been a smiley. It still holds that
expanding the address bus would mean a new chip with a whole lot of
2-byte opcodes.
-----------------
Jeffrey Hutzelman
America Online: JeffreyH11
Internet/BITNET:jh4o+@andrew.cmu.edu, jhutz@drycas.club.cc.cmu.edu,
jh4o@cmuccvma
>> Apple // Forever!!! <<