Results 1 to 6 of 6

Thread: real maximum memory

  1. #1
    Join Date
    Dec 2000
    houston, tx, USA


    What is the real maximum memory for a power pc 6116 Mac?
    This has probably been discussed to death already but I just stumbled in here and want to know.
    The Tech i use sticks with the apple guide - 72 meg - is the most supported by the Mac but what about the jumbo memory chips 64/128/256? do they work corretly? does the computer recognize and utilize the memory space?
    Also The question of FPM chips vs. EDO chips for memory? Do the EDO chips work? or for my 6116 just stick with FPM chips?

  2. #2
    Join Date
    Aug 2000
    Concord, CA



    Though unsupported, go here and it looks like 264MB including the 8MB on the motherboard. It also goes into the EDO/FPM question. Of course the disclaimer, YMMV. k

  3. #3
    Join Date
    Sep 2000


    on this page:

    The gurus state "MacGurus does not recommend the use of 64MB SIMMs in any Mac besides the 6100 at this time"

    With dual 64 MB, and the built in 8 Mb, will result in 136 MB ram in the 6100. Not bad if you ask me for the pizza box computer.

    I had no idea you could go past 72 Mb (two 32Mb simms) in the 6100. Nice to know that little guy has alot of life left. My original 6100 is still running today. Have a Sonnet Tech upgrade with a Seagate Barracuda installed and 40 Mb ram.


  4. #4
    Join Date
    May 2000
    wherever I hang my hat


    you can use 64MB SIMMs, and Apple claims it doesn't matter whether you use EDO or FPM. Then again, Apple claims 32MB SIMMs are the largest.

    there is something going on with the memory controllers in these original Power Macs, though, as the 6100 will handle a pair of 64MB SIMMs, but the others in the series, the 7100 and the 8100, are much more finicky. At one point, we thought they would also work in these other machines, but we encountered problems in some cases, and since they are obsolete machines, never spent much time trying to figure it out with build houses and different memory revisions.

    we do know that our 64MB SIMMs work fine in 6100's. We're running a pair in our FTP server right now, and have been, for over a year.

    the other important factoid is that you must use a matched pair of SIMMs, regardless.

    [This message has been edited by magician (edited 05 December 2000).]

  5. #5
    Join Date
    Feb 2001
    Austin, TX


    Quote Originally Posted by magician View Post
    there is something going on with the memory controllers in these original Power Macs, though, as the 6100 will handle a pair of 64MB SIMMs, but the others in the series, the 7100 and the 8100, are much more finicky.

    Way late, but for completeness sake, and because someone might come looking for this old information in this excellent resource.

    The 6100, 7100 and 8100 all have exactly the same memory controller chips and ought to support the same SIMMs, except...

    The 7100 and 8100 run the upper address wire and the RAS signals through a CPLD chip on the board. The 6100 lacks that CPLD.

    The end result is that the 6100 can address up to 128 MB SIMMs (if they're wired properly) and the 7100 and 8100 are limited to 32MB SIMMs. The CPLD makes the additional (more than the two in the 6100) SIMM sockets work.

    If you try to install a 64 or 128 MB SIMM in the 7100 or 8100 it will be seen as only 1/4 its capacity. There's a good reason for this.

    The 72 pin SIMM has 12 address wires. Memory addresses are multiplexed (sent in two consecutive parts) so 12 address wires X 2 yields 24 memory bits on a SIMM. Twenty-four bits of address translates to 2^24 = 16M addresses on a 72 pin SIMM. Remember that the 72 pin SIMM is 4 bytes wide (it delivers 4 bytes of data per access) so 16M addresses X 4 bytes = 64 MBytes of storage.

    So how can there be 128 MB SIMMs if only 64MB can be addressed? Remember how the addresses are multiplexed? To make that work there are two more signals, called RAS and CAS. When RAS is active, the memory knows that the address is the first part (the Row address). When CAS is active, the memory knows that the address on the wires is the second part (the Column address).

    72 pin SIMMs have four RAS lines. These can be used to create banks on the SIMM. In the case of a 128 MB SIMM, there are the equivalent of two 64 MB SIMMs on the module. One set is connected to two of the RAS lines. The other set is connected to the other set of RAS lines.

    Now more memory can be accessed by translating CPU addresses into not just memory address lines, but also memory RAS lines.

    This is also how multiple SIMM sockets are typically handled. All the SIMMs see the same address signals all the time, but only the desired SIMM socket gets an active RAS signal.

    Sooo, the memory controller in the X100 family apparently has 12 address lines and 2 RAS signals. In the 6100 all 12 address lines go to the SIMMs, giving a bank size of 64MB (as discussed above) and the two RAS lines are connected to two each of the RAS lines in the SIMM sockets, giving the possibility of two banks.

    Two banks of 64MB is 128 MB per SIMM.

    In the 7100 and 8100 those two RAS lines, plus the upper address line are massaged into providing four RAS lines in the 7100 and eight RAS lines in the 8100. So each socket has two independent RAS lines, but each pair of sockets share RAS lines (remember, memory must be installed in pairs).

    The net result is that the 7100 and 8100 only support 11 address lines but still support 2 banks. Eleven addresses X 2 = 22 => 4M addresses X 4 bytes => 16 MB per bank X 2 banks = 32MB per SIMM.

    Yes, 32 MB SIMMs are almost universally two banked.

    So that is why, even though they use exactly the same chipset, the 6100 supports SIMMs 4X larger than the 7100/8100.

    One last thought, the 6100 ties the four RAS lines in the 72 pin SIMM into pairs. Some 128 MB SIMMs are not wired properly to support this. So even though 128 MB SIMMs are supported, not all 128 MB SIMMs will work. It depends on how they routed the RAS lines to which memory chips.

  6. #6
    Join Date
    Aug 2001
    Grangeville, ID USA


    That goes a lot deeper than I ever attempted. It makes sense, but it also makes you wonder why? Why would Apple leave out the CPLD chip? It can't have had much bearing on cost. These motherboards were being manufactured at the same time. Love to be able to read design notes.
    molṑn labe'
    "I am a mortal enemy to arbitrary government and unlimited power. I am naturally very jealous for the rights and liberties of my country, and the least encroachment of those invaluable privileges is apt to make my blood boil."
--Ben Franklin

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts