andy@pro-palmtree.cts.com (Andy Stein) (06/17/91)
While it may be nonstandard and not fully addressable, isn't the Apple IIGS's memory limit 17 Meagabytes? If one has a ROM 03 GS, and Applied Engineering's RamKeeper and two GS RAM Pluses with 8 Megabyes each, isn't the total memory capacity of the computer 17.125 Megabytes, with the ROM 03's 1.125 Megabyte on board memory? Maybe I'm totally offbase, but that's what AE advertised.
q4kx@vax5.cit.cornell.edu (Joel Sumner) (06/17/91)
In article <1991Jun17.065317.13965@clark.edu>, andy@pro-palmtree.cts.com (Andy Stein) writes: > While it may be nonstandard and not fully addressable, isn't the Apple > IIGS's memory limit 17 Meagabytes? If one has a ROM 03 GS, and Applied > Engineering's RamKeeper and two GS RAM Pluses with 8 Megabyes each, isn't > the total memory capacity of the computer 17.125 Megabytes, with the ROM > 03's 1.125 Megabyte on board memory? Maybe I'm totally offbase, but that's > what AE advertised. Sorry, but this is incorrect. The 65816 processor has an address space of 2^24 or 16 million, 771 thousand, + a few. Therefore, the most memory the processor can directly address is 2^24 bytes. Some of this memory, though, is reserved by Apple and some of it is already in use. The main memory that came with the original Apple IIgs was in addresses $00/0000-$01/FFFF and $E0/0000-E1/FFFF (Adds up to 256K). The 128K of ROM is mapped to $FE/0000-$FF/FFFF. When you add memory via slinky cards in the Apple RAM Expansion slot, it gets mapped from $02/0000 on up. You would think that you still have 15 megs or so to expand though. This is not the case. When Apple built the IIgs, it reserved the memory above $08/0000 for Apple purposes (ROM expansion, etc...) This you can only have actual RAM from $00/0000-$7F/FFFF or 8 megabytes. Therefore, the maximum RAM you can have (without some sort of trick that they may invent in the next 5 years) is 8 Megs. None of this is including the 64K dedicated to the DOC. I am not sure if that is included in the memory map or not. For more information, look in the Apple IIgs Hardware Reference. It is probably in the firmware reference too. And I am sure that Matt and Dave will both jump on this post and clear up any inaccuracies I have written. -- Joel Sumner GENIE:JOEL.SUMNER This .sig may not be used q4kx@cornella.ccs.cornell.edu q4kx@cornella for public viewing or q4kx@vax5.cit.cornell.edu q4kx@crnlvax5 rebroadcast without the .................................................... express written consent The impedance of absolutely nothing is 377 ohms. of major league baseball.
gwyn@smoke.brl.mil (Doug Gwyn) (06/17/91)
In article <1991Jun17.065317.13965@clark.edu> andy@pro-palmtree.cts.com (Andy Stein) writes: > While it may be nonstandard and not fully addressable, isn't the Apple >IIGS's memory limit 17 Meagabytes? No, it is architecturally incapable of accessing more than 8MB of main-memory RAM. There is also a ROM space (containing the firmware and room for other ROM), but only a tiny number of sites are going to be able to exploit the additional ROM space. The on-board RAM lies within the 8MB space.
stuckey@ux1.cso.uiuc.edu (Anthony J. Stuckey) (06/17/91)
andy@pro-palmtree.cts.com (Andy Stein) writes: > While it may be nonstandard and not fully addressable, isn't the Apple >IIGS's memory limit 17 Meagabytes? If one has a ROM 03 GS, and Applied >Engineering's RamKeeper and two GS RAM Pluses with 8 Megabyes each, isn't >the total memory capacity of the computer 17.125 Megabytes, with the ROM >03's 1.125 Megabyte on board memory? Maybe I'm totally offbase, but that's >what AE advertised. Except that I think you can aonly address 16 MB. The GS does use 24 bit addressing, which gives a theoretical maximum of 16 MB. I would be real surprised if 16 MB proved to be exceedable. -- Anthony J. Stuckey stuckey@ux1.cso.uiuc.edu
$CSB205@LSUVM.BITNET (Mark Orr) (06/18/91)
|From: Joel Sumner <vax5.cit.cornell.edu!q4kx@CU-ARPA.CS.CORNELL.EDU> |Subject: Re: Apple IIGS's Memory Limit |Therefore, the maximum RAM you can have (without some sort of trick that |they may invent in the next 5 years) is 8 Megs. None of this is including |the 64K dedicated to the DOC. I am not sure if that is included in the |memory map or not. Well I suppose you could develop a RAM card with a pointer location that would BANK SELECT. But we don't want to get into that, do we? |Joel Sumner GENIE:JOEL.SUMNER This .sig may not be us |q4kx@cornella.ccs.cornell.edu q4kx@cornella for public viewing or |q4kx@vax5.cit.cornell.edu q4kx@crnlvax5 rebroadcast without the |.................................................... express written consent |The impedance of absolutely nothing is 377 ohms. of major league basebal -------------------------------- "We were having a great time, but ! Mark Orr ! Macintosh wasn't selling that well. ! $CSB205 @ LSUVM.SNCC.LSU.EDU ! The Apple II was paying our way." ! @ LSUVM.BITNET ! - Guy Kawasaki -------------------------------- From his book: The Macintosh Way
PKBRANDON@MSUS1.MSUS.EDU (06/18/91)
I _believe_ that the 65816 microprocessor in the GS can address 16meg. However, the GS is set up to allocate 8meg to RAM and 8meg to ROM, so to use more than 8meg you need to set a a ROM disk (the RamKeeper does this). The GS Ram+ can hold up to 6meg; 8meg with a piggyback board that may or may not exist at a reasonable price. Can't see any reason to stick 16meg in there, 'tho; my experience has been that hard drives are more reliable and cost-effective (I've got a Ram Keeper in my office machine; a CMS hard drive in my home machine). The big advantage of the Ram Keeper is the faster boot and a couple of meg can handle that. ------------------------------------------------------------------ --- Paul Brandon Psychology Dept Mankato State Univ --- --- PKBRANDON@MSUS1.MSUS.EDU Mankato, MN 56001 --- ------------------------------------------------------------------
gwyn@smoke.brl.mil (Doug Gwyn) (06/18/91)
In article <1991Jun17.100712.5532@vax5.cit.cornell.edu> q4kx@vax5.cit.cornell.edu (Joel Sumner) writes: >None of this is including the 64K dedicated to the DOC. I am not sure >if that is included in the memory map or not. No, the DOC RAM is in an address space of its own, not directly accessible to the CPU.
taob@micor.ocunix.on.ca (Brian Tao) (06/18/91)
stuckey@ux1.cso.uiuc.edu (Anthony J. Stuckey) writes: > Except that I think you can aonly address 16 MB. The GS does use 24 bit > addressing, which gives a theoretical maximum of 16 MB. > I would be real surprised if 16 MB proved to be exceedable. How about memory plugged into one of the seven "normal" slots? Would GS/OS recognize that as a RAM disk? For example, a card could be built which supports 1- and 4-meg SIMM's, but into a bus slot instead of the memory expansion slot. With perhaps a GS/OS driver, it could be used as a large RAM disk. Add a battery-backup option, and you can get 32 megs of fast, internal storage (for big $$$, obviously). Wasn't Harris Labs going to do something like this?
MQUINN@UTCVM.BITNET (06/18/91)
On Mon, 17 Jun 1991 15:47:08 GMT Anthony J. Stuckey said: >andy@pro-palmtree.cts.com (Andy Stein) writes: >>Engineering's RamKeeper and two GS RAM Pluses with 8 Megabyes each, isn't >>the total memory capacity of the computer 17.125 Megabytes, with the ROM >>03's 1.125 Megabyte on board memory? Maybe I'm totally offbase, but that's >>what AE advertised. > >Except that I think you can aonly address 16 MB. The GS does use 24 bit >addressing, which gives a theoretical maximum of 16 MB. >I would be real surprised if 16 MB proved to be exceedable. The '816 can address up to 16MB of -memory-. Remember, that part of that memory is ROM, not RAM. The -practible- limit of RAM is 8MB. I assume the upper 8 megs are reserved for future expansion ROM and/or user upgradable ROM. If you have more than 8MB of RAM installed, you'd have to patch the memory manager use anything over the 8MB, but then, you still couldn't access it completely with current applications. For example, I doubt you'd be able to JML >8MB etc. without jumping to a ROM routine. I'm sure someone else could explain it much better than I'm doing now because I don't completely understand it myself. >-- >Anthony J. Stuckey >stuckey@ux1.cso.uiuc.edu ---------------------------------------- BITNET-- mquinn@utcvm <------------send files here pro-line-- mquinn@pro-gsplus.cts.com
gwyn@smoke.brl.mil (Doug Gwyn) (06/18/91)
In article <PmTP44w162w@micor.ocunix.on.ca> taob@micor.ocunix.on.ca (Brian Tao) writes: > How about memory plugged into one of the seven "normal" slots? Would >GS/OS recognize that as a RAM disk? If its slot firmware made it look like a block device, yes. There have in fact been Apple II RAM disk cards available. However, since access to their contents is radically different (and much more cumbersome) than direct access to main memory, they are no substitute for main memory RAM.
rhyde@koufax.ucr.edu (randy hyde) (06/19/91)
Of course, the Apple IIe uses a 6502 with a 16-bit address bus. That would limit it to 64K. Didn't stop people from putting a megabyte into it. Yeah, it used bankswitching. The GS could use that too. That's probably what AE is alluding to, some sort of bank switching scheme. I wouldn't know though, I haven't looked closely at their stuff in a couple of years. *** Randy Hyde
gray@ibis.cs.umass.edu (Lyle Gray) (06/19/91)
In article <16430@smoke.brl.mil> gwyn@smoke.brl.mil (Doug Gwyn) writes: >No, it is architecturally incapable of accessing more than 8MB of >main-memory RAM. There is also a ROM space (containing the firmware >and room for other ROM), but only a tiny number of sites are going to >be able to exploit the additional ROM space. > >The on-board RAM lies within the 8MB space. I seem to recall (from the original owner's manuals) that the RAM limit was 8.25 Mb, not 8 Mb. The remaining addressable space (up to 16 Mb) was reserved for ROM expansions. -------------------------------------------------------------------------- Lyle H. Gray Internet (personal): gray@cs.umass.edu Quodata Corporation Phone: (203) 728-6777, FAX: (203) 247-0249 --------------------------------------------------------------------------
MQUINN@UTCVM.BITNET (06/20/91)
On Tue, 18 Jun 1991 20:20:04 GMT Lyle Gray said: >>The on-board RAM lies within the 8MB space. > >I seem to recall (from the original owner's manuals) that the RAM limit was >8.25 Mb, not 8 Mb. The remaining addressable space (up to 16 Mb) was reserved >for ROM expansions. It's actually 8.388608MB, but that's considered '8MB'. Just like, 1k is really 1024 bytes and 1MB is really 1 million plus 24kb. >-------------------------------------------------------------------------- >Lyle H. Gray Internet (personal): gray@cs.umass.edu >Quodata Corporation Phone: (203) 728-6777, FAX: (203) 247-0249 >-------------------------------------------------------------------------- ---------------------------------------- BITNET-- mquinn@utcvm <------------send files here pro-line-- mquinn@pro-gsplus.cts.com
MQUINN@UTCVM.BITNET (06/20/91)
On Tue, 18 Jun 1991 18:09:14 GMT randy hyde said: >Of course, the Apple IIe uses a 6502 with a 16-bit address bus. That would >limit it to 64K. Didn't stop people from putting a megabyte into it. >Yeah, it used bankswitching. The GS could use that too. That's probably what >AE is alluding to, some sort of bank switching scheme. I wouldn't know though, >I haven't looked closely at their stuff in a couple of years. >*** Randy Hyde ...but that would make the new memory available only to a few applications, like with the //e. Only Appleworks, and a few odds and ends programs could use it directly, while the vast majority of programs could only use it as a RAM disk, although, that would still be better than not having it. ---------------------------------------- BITNET-- mquinn@utcvm <------------send files here pro-line-- mquinn@pro-gsplus.cts.com
andy@pro-palmtree.cts.com (Andy Stein) (06/21/91)
In-Reply-To: message from gray@ibis.cs.umass.edu
OK, I had to ask this. JUST WHAT THE HECK CAN ANYONE DO WITH 8
MEGABYTES OF ROM?!? Didn't Apple have any pragmatism when it designed the
Apple IIGS? I mean, wouldn't 12 Megs of RAM and 4 Megs of ROM have been a
better allocation of system memory? AppleWorks GS doesn't even consume 4
Megs, so that could be put in the ROM, and Apple would have one killer
machine. Could you imagine AppleWorks GS in ROM? Oh, no. It might sell
then.
.swrhood@pro-gsplus.cts.com (Robert Hood) (06/21/91)
In-Reply-To: message from MQUINN@UTCVM.BITNET Re: Theoretical memory limit on the GS - 8 MB? 16 MB? 17.125 MB? I'm shocked. In all of this thread I've seen, everyone has mentioned only what is directly accessible. Has everyone forgotten the 128K 8-bits? The address space on a 6502/65C02 is only 64K - surely someone could figure out a way to do something similar with the GS, no? ---- ProLine: rhood@pro-gsplus | "Wherever you go...there you are." Internet: rhood@pro-gsplus.cts.com | -- Buckaroo Banzai UUCP: crash!pro-gsplus!rhood | Wanted: An unZIPper for a II! ARPA: crash!pro-gsplus!rhood@nosc.mil | If you have one, let's chat!
unknown@ucscb.UCSC.EDU (The Unknown User) (06/21/91)
In article <1991Jun20.221622.561@crash.cts.com> rhood@pro-gsplus.cts.com (Robert Hood) writes: >In-Reply-To: message from MQUINN@UTCVM.BITNET >Re: Theoretical memory limit on the GS - 8 MB? 16 MB? 17.125 MB? >I'm shocked. In all of this thread I've seen, everyone has mentioned only >what is directly accessible. Has everyone forgotten the 128K 8-bits? The >address space on a 6502/65C02 is only 64K - surely someone could figure out >a way to do something similar with the GS, no? Yes, I guess you could do something similar, but -WHY-? If more memory were needed (I think 8 megs is a HELL of a lot, and even if a MultiFinder GS ever existed, I doubt you'd ever fill that much), go through the bother of building in VIRTUAL memory, which makes more memory invisible to the application.. That is, the operating system handles dealing with more memory, rather than the program doing bank switching crap. -- /unknown@ucscb.ucsc.edu Apple IIGS Forever! unknown@cats.ucsc.edu\ |WANT to help get ULTIMA VI //e or GS written?-mail me. CHEAP CD info-mail me.| \ It's a Late Night World.... Of Love /
-Rich-@cup.portal.com (Richard Sherman Payne) (06/21/91)
>On Tue, 18 Jun 1991 20:20:04 GMT Lyle Gray said: >>>The on-board RAM lies within the 8MB space. >> >>I seem to recall (from the original owner's manuals) that the RAM limit was >>8.25 Mb, not 8 Mb. The remaining addressable space (up to 16 Mb) was reserve d >>for ROM expansions. > >It's actually 8.388608MB, but that's considered '8MB'. Just like, 1k is reall y >1024 bytes and 1MB is really 1 million plus 24kb. >>-------------------------------------------------------------------------- >>Lyle H. Gray Internet (personal): gray@cs.umass.edu >>Quodata Corporation Phone: (203) 728-6777, FAX: (203) 247-0249 > This is a little mixed up as I understand things. The addressable space is 2^23= 8,338,608 bytes. Since 1K = 1024 bytes, we have 8,388,608/1024 = 8,192 KB. By common usage, this is called 8MB. Now 8.338608 MB would be 8.338608 * 1024 = 8,589.93 KB, which is about 400K over 2^23 bytes. The error seems to be dividing by 1000 to convert from 8,338,608 in converting from bytes to MB. A common mistake to those not computer literate. But I am surprised to see it here. Rich -Rich-@cup.portal.com
gwyn@smoke.brl.mil (Doug Gwyn) (06/21/91)
In article <1991Jun20.221622.561@crash.cts.com> rhood@pro-gsplus.cts.com (Robert Hood) writes: >I'm shocked. In all of this thread I've seen, everyone has mentioned only >what is directly accessible. Has everyone forgotten the 128K 8-bits? The >address space on a 6502/65C02 is only 64K - surely someone could figure out >a way to do something similar with the GS, no? I can tell you're not a programmer! As soon as a different mechanism has to be introduced to access such "extended" memory, programming to use it becomes MUCH harder. You might as well resort to RAM disk or even hard disk for the extension, which of course is already possible as things now stand.
jpenne@ee.ualberta.ca (Jerry Penner) (06/21/91)
In article <1991Jun20.221622.561@crash.cts.com> rhood@pro-gsplus.cts.com (Robert Hood) writes: >In-Reply-To: message from MQUINN@UTCVM.BITNET > >Re: Theoretical memory limit on the GS - 8 MB? 16 MB? 17.125 MB? > >I'm shocked. In all of this thread I've seen, everyone has mentioned only >what is directly accessible. Has everyone forgotten the 128K 8-bits? The >address space on a 6502/65C02 is only 64K - surely someone could figure out >a way to do something similar with the GS, no? Sure, but what's the point of using bank switching? First of all, rather few of the installed base of GS's have more than 1.5 or 2 megs of RAM; we're not even pushing the 8 Meg barrier yet so why worry about more. Second, bank switching would slow the machine down further and it's not fast enough as is. Yup, we could probably come up with a really nifty bank-switching scheme, but we don't need to yet so we won't. -- Jerry Penner | jpenne@bode.ee.ualberta.ca Edmonton, Alberta, Canada | OR alberta!bode!jpenne
jb10320@uxa.cso.uiuc.edu (Jawaid Bazyar) (06/21/91)
rhood@pro-gsplus.cts.com (Robert Hood) writes: >In-Reply-To: message from MQUINN@UTCVM.BITNET >Re: Theoretical memory limit on the GS - 8 MB? 16 MB? 17.125 MB? > >I'm shocked. In all of this thread I've seen, everyone has mentioned only >what is directly accessible. Has everyone forgotten the 128K 8-bits? The >address space on a 6502/65C02 is only 64K - surely someone could figure out >a way to do something similar with the GS, no? ICKY! NASTY! FOO! The whole point of having a processor that can directly access 16 megabytes is so you don't have to use bank switching. Adding bank switching would be a fruitless task on the GS. A much better project would be true demand paged virtual memory. According to the hardware reference, cards that follow Apple's design rules for memory cards can have to 4 megabytes, period. Add this 4 meg to whatever the machine you've got has on the motherboard (ROM 01- 256k, ROM 03- 1.125Meg), and that's the total in your machine. I doubt a card like the Octo RAM would work in a ROM 03, but I think AE's GSRAM+ does, so there you have 6Meg+1.125Meg for close to eight. It might be possible to build a 12Meg RAM card, but it would be useless because the Memory manager is only built for 8 megs. -- Jawaid Bazyar | "Twenty seven faces- with their eyes turned to Graduated!/Comp Engineering | the sky. I have got a camera, and an airtight bazyar@cs.uiuc.edu | alibi.." Apple II Forever! | I need a job... Be privileged to pay me! :-)
gray@ibis.cs.umass.edu (Lyle Gray) (06/21/91)
In article <43521@cup.portal.com> -Rich-@cup.portal.com (Richard Sherman Payne) writes: >>On Tue, 18 Jun 1991 20:20:04 GMT Lyle Gray said: >>>>The on-board RAM lies within the 8MB space. >>> >>>I seem to recall (from the original owner's manuals) that the RAM limit was >>>8.25 Mb, not 8 Mb. The remaining addressable space (up to 16 Mb) was reserve >d >>>for ROM expansions. >> >>It's actually 8.388608MB, but that's considered '8MB'. Just like, 1k is reall >y >>1024 bytes and 1MB is really 1 million plus 24kb. > >This is a little mixed up as I understand things. The addressable space >is 2^23= 8,338,608 bytes. Since 1K = 1024 bytes, we have 8,388,608/1024 = >8,192 KB. By common usage, this is called 8MB. Now 8.338608 MB would be >8.338608 * 1024 = 8,589.93 KB, which is about 400K over 2^23 bytes. > >The error seems to be dividing by 1000 to convert from 8,338,608 in >converting from bytes to MB. A common mistake to those not computer >literate. But I am surprised to see it here. > All right, now you've annoyed me. ;-) _I_ did not make that mistake (I'm quite familiar with it, from having to educate people about the real meaning of 1 megabyte). I was merely stating what I recall reading in the original owner's manuals for the IIgs. I can't find my manuals right now (dig, dig), but that _is_ what I remember: slightly more than 8 megs for RAM, slightly _less_ that 8 megs for ROM. It's possible that has changed since, but that's what I remember. -------------------------------------------------------------------------- Lyle H. Gray Internet (personal): gray@cs.umass.edu Quodata Corporation Phone: (203) 728-6777, FAX: (203) 247-0249 --------------------------------------------------------------------------
rhyde@koufax.ucr.edu (randy hyde) (06/22/91)
>> Bankswitching isn't necessary...
Gee, I wouldn't mind having several banks of bank 0 memory. That would
make a process switcher much easier to deal with.ericmcg@pnet91.cts.com (Eric Mcgillicuddy) (06/22/91)
The GS can address up to 16Megs of memory, however half of that is ROM and/or I/O spaces. The 8Megs can all be RAM, however there are only 2 bank selector lines for DMA, so only 4 banks in the memory expansion board can be accessed using DMA. How big these banks are depends on the address lines available, currently this is 1Meg per bank. The practical limit to GS memory is 5.125Megs of RAM. UUCP: bkj386!pnet91!ericmcg INET: ericmcg@pnet91.cts.com
gwyn@smoke.brl.mil (Doug Gwyn) (06/22/91)
In article <17284@darkstar.ucsc.edu> unknown@ucscb.UCSC.EDU (The Unknown User) writes: > If more memory were needed (I think 8 megs is a HELL of a lot, >and even if a MultiFinder GS ever existed, I doubt you'd ever fill that much), >go through the bother of building in VIRTUAL memory, which makes more memory >invisible to the application.. That is, the operating system handles >dealing with more memory, rather than the program doing bank switching >crap. I agree that explicitly dealing with bank switching is crap. However, I disagree on other points. No matter HOW much memory (also disk, also CPU power) one has, he can devise an application that genuinely needs more. More frequently, the programmer has assumed that a particular implementation technique will suffice for virtually all uses of the program; the most common example on the IIGS is requiring all data of a certain type to fit into an array contained in one 64KB data bank. This sort of program limitation is avoidable, but often avoiding such restrictions requires more work on the part of the programmer, and if it is not perceived to be a potentially frequent problem, many programmers would not bother. In fact, 8MB is not very much memory by the standards of a lot of applications that I've developed. Such applications are cramped when ported to the IIGS. However, CPU power also gets to be a factor when large amounts of data need to be processed, so 8MB may be a fairly good match to the 65816's computational abilities. One should not rely on virtual memory to extend the address space. While it technically does that, it does so at a severe cost, such that in many cases one would be much better off redesigning the application to take EXPLICIT control over disk I/O rather than allowing uncontrolled IMPLICIT disk I/O to occur via the virtual memory system.
q4kx@vax5.cit.cornell.edu (Joel Sumner) (06/22/91)
In article <32395@dime.cs.umass.edu>, gray@ibis.cs.umass.edu (Lyle Gray) writes: > In article <43521@cup.portal.com> -Rich-@cup.portal.com (Richard Sherman > Payne) writes: >>>On Tue, 18 Jun 1991 20:20:04 GMT Lyle Gray said: >>>>>The on-board RAM lies within the 8MB space. >>>> >>>>I seem to recall (from the original owner's manuals) that the RAM limit was >>>>8.25 Mb, not 8 Mb. The remaining addressable space (up to 16 Mb) was >>>>reserved >>>>for ROM expansions. [incorrect assumption Rich Payne didn't know how to multiply deleted] > All right, now you've annoyed me. ;-) _I_ did not make that mistake (I'm > quite familiar with it, from having to educate people about the real meaning > of 1 megabyte). I was merely stating what I recall reading in the original > owner's manuals for the IIgs. I can't find my manuals right now (dig, dig), > but that _is_ what I remember: slightly more than 8 megs for RAM, slightly > _less_ that 8 megs for ROM. It's possible that has changed since, but that's > what I remember. > You are correct. According to the GS/OS reference manual (page 32), the memory map is as follows [edited for briefness]. RAM - Banks $00-$7F (8Megs) + Banks $E0-E1 (128K Shadow Ram) Expansion ROM - Banks $F0-$FD Total ROM Space - Banks $F0-$FD + Banks $FE-$FF (Onboard ROM) If shadowing is on, of course, you only effectively have 8Megs of storage because Banks $E0-$E1 mirror banks $00-$01. -- Joel Sumner GENIE:JOEL.SUMNER This .sig may not be used q4kx@cornella.ccs.cornell.edu q4kx@cornella for public viewing or q4kx@vax5.cit.cornell.edu q4kx@crnlvax5 rebroadcast without the .................................................... express written consent The impedance of absolutely nothing is 377 ohms. of major league baseball.
q4kx@vax5.cit.cornell.edu (Joel Sumner) (06/22/91)
In article <864@generic.UUCP>, ericmcg@pnet91.cts.com (Eric Mcgillicuddy) writes: > The GS can address up to 16Megs of memory, however half of that is ROM and/or > I/O spaces. The 8Megs can all be RAM, however there are only 2 bank selector > lines for DMA, so only 4 banks in the memory expansion board can be accessed > using DMA. How big these banks are depends on the address lines available, > currently this is 1Meg per bank. Be carful when you use the term 'bank' in the above context. 1 Memory Bank as defined for the Apple IIgs memory map is 64K. In the above example, you mean 'rows'. The memory expansion port only allows 4 rows of chips (256K/row with 256 Kbit chips or 1Meg/row with 1MBit chips) to be addressed by DMA. Applied Engineering did try to make 6 addressable rows by reading some bus line in addition to the 2 bits (4 rows) used in the standard DMA hardware but I don't remember what degree of success they had. -- Joel Sumner GENIE:JOEL.SUMNER This .sig may not be used q4kx@cornella.ccs.cornell.edu q4kx@cornella for public viewing or q4kx@vax5.cit.cornell.edu q4kx@crnlvax5 rebroadcast without the .................................................... express written consent The impedance of absolutely nothing is 377 ohms. of major league baseball.
ericmcg@pnet91.cts.com (Eric Mcgillicuddy) (06/22/91)
>Re: Theoretical memory limit on the GS - 8 MB? 16 MB? 17.125 MB? > >I'm shocked. In all of this thread I've seen, everyone has mentioned only >what is directly accessible. Has everyone forgotten the 128K 8-bits? The >address space on a 6502/65C02 is only 64K - surely someone could figure out >a way to do something similar with the GS, no? >---- >ProLine: rhood@pro-gsplus | "Wherever you go...there you Ok, there is NO theoretical limit to the ammount of memory on the GS. TerraBytes and multi-mega-TerraBytes can be installed in a GS when multi-mega-TerraByte RAM chips become available. however anything about 16Meg is a RAM disk, 8MEg of that is ROM and 4Meg of the remaining is not DMA compatible. happy? UUCP: bkj386!pnet91!ericmcg INET: ericmcg@pnet91.cts.com
ericmcg@pnet91.cts.com (Eric Mcgillicuddy) (06/23/91)
>Be carful when you use the term 'bank' in the above context. 1 Memory Bank >as defined for the Apple IIgs memory map is 64K. In the above example, you >mean 'rows'. The memory expansion port only allows 4 rows of chips >(256K/row with 256 Kbit chips or 1Meg/row with 1MBit chips) to be addressed >by DMA. Applied Engineering did try to make 6 addressable rows by reading >-- >Joel Sumner GENIE:JOEL.SUMNER This .sig may not be used You are correct, Apple uses the term "row" for memory expansion, a bank is 64k and there may be several banks in a row. UUCP: bkj386!pnet91!ericmcg INET: ericmcg@pnet91.cts.com
tribby@hpindwa.cup.hp.com (David Tribby) (06/27/91)
I have an OctoRAM card in my ROM01 GS. Although I currently only have two
meg on it, I could load it up to eight meg. I also have an add-on board,
ESP, that uses battery-backed-up static RAM to form a ROM disk. This ROM
disk currently has a whopping 128K on it, but it is expandable out to
512K (1/2 meg).
One of the reasons I bought the OctoRAM back in '88 was that it was
designed to allow all of the expansion memory in the IIGS specification:
8 meg of RAM and 1/2 meg of ROM.
My memory map looks like this:
Bank (64K each)
$00 - $01 Built-in ROM01 RAM
$02 - $21 2-meg on OctoRAM
$22 - $81 Another 6-meg could be added to OctoRAM
<unused memory space>
$E0 - $E1 "Slow" built-in ROM01 RAM
<unused memory space>
$F0 - $F1 128K on ESP (ROM disk)
$F2 - $F7 Another 384K available for ROM disk
$F8 - $FD "Reserved ROM space" (some used by ROM03)
$FE - $FF ROM01 firmware
A few years back, I read about incompatibilities between the upper 4 meg
of expansion RAM and DMA, so if I fully load the OctoRAM, banks $42 - $81
would not be compatible with all cards that use DMA.
When the ROM03 GS was released, it came with 1 meg of built-in memory and
used some of the reserved ROM space ($FC - $FD).
If I were to fully load my OctoRAM and ESP cards and put them in a ROM03,
wouldn't the memory look like this:
Bank (64K each)
$00 - $FF Built-in ROM03 RAM
$01 - $4F 4-meg on OctoRAM compatible with DMA
$50 - $8F Another 4-meg on OctoRAM incompatible with DMA
<unused memory space>
$E0 - $E1 "Slow" built-in ROM03 RAM
<unused memory space>
$F0 - $F7 512K on ESP (ROM disk)
$F8 - $FB "Reserved ROM space"
$FC - $FF ROM03 firmware
That would give me a 9 meg GS (9.625 if you add the "slow" RAM and ESP).
Would this work in a ROM03? Have I missed any "reserved" memory?
-- Dave Tribby
- - - - -
tribby@hpda.cup.hp.com or tribby@cup.hp.com or hplabs!hpda!tribbygwyn@smoke.brl.mil (Doug Gwyn) (06/28/91)
In article <54240034@hpindwa.cup.hp.com> tribby@hpindwa.cup.hp.com (David Tribby) writes: > $22 - $81 Another 6-meg could be added to OctoRAM > $50 - $8F Another 4-meg on OctoRAM incompatible with DMA No, you cannot access "bank $8x" as RAM. The IIGS has a hard upper limit of 8MB total RAM space.
whitewolf@gnh-starport.cts.com (Tae Song) (06/29/91)
|It's actually 8.388608MB, but that's considered '8MB'. Just like, 1k is |really 1024 bytes and 1MB is really 1 million plus 24kb. People will have already pointed this out to you, but... 1024 bytes = 1K 1024 KB = 1M 1024x1024 = 1M = 1,048,576 256K = .25MB (which comes with the ROM 2 (01) GS) (1.125 MB on the ROM 3 GS) You can an additional 8MB. That gives you 8.25MB (which is correct) or 8,650,752 bytes total. INET: whitewolf@gnh-starport.cts.com UUCP: crash!gnh-starport!whitewolf ARPA: crash!gnh-starport!whitewolf@nosc.mil
whitewolf@gnh-starport.cts.com (Tae Song) (06/29/91)
|Re: Theoretical memory limit on the GS - 8 MB? 16 MB? 17.125 MB? | |I'm shocked. In all of this thread I've seen, everyone has mentioned only |what is directly accessible. Has everyone forgotten the 128K 8-bits? The |address space on a 6502/65C02 is only 64K - surely someone could figure out |a way to do something similar with the GS, no? |Re: Theoretical memory limit on the GS - 8 MB? 16 MB? 17.125 MB? | |I'm shocked. In all of this thread I've seen, everyone has mentioned only |what is directly accessible. Has everyone forgotten the 128K 8-bits? The |address space on a 6502/65C02 is only 64K - surely someone could figure out |a way to do something similar with the GS, no? We're talking about two different types of memory addressing here. The max amount of memory the 65816 can address DIRECTLY is 16MB. The max amount of memory the GS can address is 8.25M of RAM and 1MB of ROM (this includes ROM already built in) on the ROM 2 (01) GS. On the ROM 3 GS you can address 8.125MB and 1MB of ROM. (I'm sure someone will correct me if I'm wrong). All the stuff above deals with memory being addressed DIRECTLY. The other is bank-switching and I and have no idea how or why you'd want to. And one other type is virtual memory which hasn't been perfected for the GS. Question sticks out in my mind, though... if you have a ROM 3 GS and add 8M to it, you'd have 9.125MB... what would happen to the the top 1M? would it be available as a RAMdisk or would it end up been wasted... ??? INET: whitewolf@gnh-starport.cts.com UUCP: crash!gnh-starport!whitewolf ARPA: crash!gnh-starport!whitewolf@nosc.mil
curtis@achilles.ctd.anl.gov (Jeffrey Curtis ) (06/30/91)
In article <m0jtbiy-0000ozC@crash.cts.com> whitewolf@gnh-starport.cts.com (Tae Song) writes: >|Re: Theoretical memory limit on the GS - 8 MB? 16 MB? 17.125 MB? >| >|I'm shocked. In all of this thread I've seen, everyone has mentioned only >|what is directly accessible. Has everyone forgotten the 128K 8-bits? The >|address space on a 6502/65C02 is only 64K - surely someone could figure out >|a way to do something similar with the GS, no? >The other is bank-switching and I and have no idea how or why you'd want to. Why you'd want to? The same reasons anyone would want to on a //e: what you're given isn't enough. When you put 8 megs up against practical applications, it's a small amount. There are computers in Japan that have over 4 gigabytes of RAM (I'll get exact figures later). I'm sure they're complaining of not having enough, too. What if you want to digitize large amounts of sound and video? 8 megs runs out quick. So you add more and bank switch. It's easy to do if you're careful about the whole mess.. >Question sticks out in my mind, though... if you have a ROM 3 GS and add 8M to >it, you'd have 9.125MB... what would happen to the the top 1M? would it be >available as a RAMdisk or would it end up been wasted... ??? No, you just have to bank switch it. ;^) *+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+* + Jeffrey S. Curtis (708)972-8585 B41801 AT ANLVM curtis@achilles.ctd.anl.gov + * Computing and Telecommunications, Argonne National Laboratory * + Want a free copy of Heatseeker, the Apple's fastest disk utility? Mail me! + *+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*
jb10320@uxa.cso.uiuc.edu (Jawaid Bazyar) (06/30/91)
curtis@achilles.ctd.anl.gov (Jeffrey Curtis ) writes: > Why you'd want to? The same reasons anyone would want to on a //e: >what you're given isn't enough. When you put 8 megs up against practical >applications, it's a small amount. There are computers in Japan that have >over 4 gigabytes of RAM (I'll get exact figures later). I'm sure they're >complaining of not having enough, too. What if you want to digitize large >amounts of sound and video? 8 megs runs out quick. So you add more and >bank switch. It's easy to do if you're careful about the whole mess.. The IIgs isn't a //e- even though the hardware has been described as a '//e on steroids', because of the '816s extended instruction set and 16-bit architecture, the system software transforms the GS into a totally different beast. The GS architecture is strained as it is, adding any sort of Apple-supported bank-switching to the GS would destroy it completely. No, there are far better alternatives than bank switching (hard drives, anyone?) Besides, I disagree with your contention that 8 megs of RAM in a GS isn't enough. Rarely do I run out of memory in my 1.75meg GS. I'd like to point out that I'm a heavy-duty developer, and Orca/C isn't the most efficient of beasts, not to mention having things like GSbug, NiftyList, and half a dozen other programmer's DA's installed. I'd be totally satisfied with 4meg- any application I chose to develop for the GS would definitely fit in 4 meg. "But wait", you say, "there are Unix programs that require more than 4 megs to compile". I'm writing *GS* programs, not Unix programs. My programs have to work on a 1meg machine. In any case, digitizing large amounts of sound and video? I don't want to sample 15 minutes of something. What use would it be besides to pretend your GS was a poor-quality tape recorder? Same goes for video. Get a fast hard drive, and only having 8 megs of ram won't bother you. -- Jawaid Bazyar | "Twenty seven faces- with their eyes turned to Graduated!/Comp Engineering | the sky. I have got a camera, and an airtight bazyar@cs.uiuc.edu | alibi.." Apple II Forever! | I need a job... Be privileged to pay me! :-)