IBM PowerPC (Apple)

A CPU corner wouldn’t be complete with some CPUs from competing camps. In this case, these are some Apple Computer CPUs from the IBM PowerPC days. Unfortunately, I don’t think there are proper socket designations for these, so hence they are lumped under the CPU architecture instead.

IBM PowerPC IBM2506K5319 933Mhz G4

This one was salvaged from a G4 Tower.

G4 933mhz CPU Module G4 933Mhz Side G4 933Mhz Heatsink Off G4 933Mhz CPU Underside G4 933Mhz CPU Chip

IBM PowerPC PPC750L-FB0B500 500Mhz G3

This one was a pull from a PowerBook of some sort, and so it seems to be a CPU, northbridge, VRM and RAM on a single module.

DSC_7733 DSC_7734

And here’s the same module without the SDRAM SODIMMs.

DSC_7735 DSC_7736

6 Responses to IBM PowerPC (Apple)

  1. sparcie says:

    I believe here Apple were following the lead of others by placing their CPU’s on replaceable modules rather than sockets. This had many advantages, for instance they could wedge more powerful chips into older systems, and reduce reliability issues as people handled the CPU itself much less.

    I’m not sure how they handled these modules whilst they were water cooled.

    Other architectures even had modules with multiple processors on them, both Sun and SGI did this with their earlier systems. For example I have a dual 50Mhz SuperSparc in my Sparcstation 20. Later SPARC and MIPS don’t have multi-CPU modules, but in principle could have.


    • lui_gough says:

      Some of the advantages also involve the constant progress of silicon chips – newer lithography formats often demand and operate at lower voltages than previously. By integrating a voltage regulator module on the module itself, this prevents an upgrade issue when the motherboard’s VRM isn’t capable of the current/voltage requirements of a new CPU. It also allows the integration of separate cache chips with the module itself, to reduce compatibility issue from cache chips mounted on the board (or COAST modules). In some sense, Intel’s Slot 1/Slot 2 took these ideas to some degree, but then the inflexibility caught up to them (cooling/mounting/form factor wise) and so we’re back to the socketed approach.

      It’s a bit of a shame that I don’t have more older examples in my collection at this tine, although I can’t say that I’ve ever made any concerted effort to collect these …

      – Gough

  2. sparcie says:

    I didn’t think of that, a very good point!

    I love this new section, it was interesting πŸ™‚


  3. Henrik says:

    It’s quite a find to get the very very rare IBM sourced G4 processor (or PowerPC 7400 which would be its more common name). Congratulations. Almost all were made by Motorola, who failed to get the yield good enough for the promised 500 MHz PowerMac G4. Motorola struggled, and IBM had to be contracted by Apple to help pick up the slack. Since there’s no available documentation for the IBM variant of the G4 I can’t discern what speed it’s rated for, but it’s certainly not a 933 MHz part. It’s in the 400-500 MHz range. The daughter card is from year 2000, and Apple didn’t produce parts with higher speeds until 2001, and with PowerPC 7450 which was solely sourced by Motorola, and those processors had another geometry of the chip (more elongated) and a different arrangement of capacitors on the substrate (3 on each side).

    Thanks for the high resolution photo though! It’s the second photo of this processor I’ve ever seen outside my own collection. Very nice find!

    • lui_gough says:

      Thanks for the comment, and quite right you are – it’s definitely not a 933Mhz part. Oops. I can’t remember where I got that number from … but you’re right – it’s likely to be a <500Mhz part. Unfortunately, it was salvaged from a disposed machine which was partially gutted, so I didn't have the chance to get more details from it before I made the pull and "ran away" with the loot :). The details are much appreciated!

Error: Comment is Missing!