Hacker News Reader

Comment by cmiles74 | Hacker News Reader

I purchased one of the first Macs that ran on a PowerPC processor and I was profoundly disappointed. Performance was no better than a 68040-based Mac and it was pretty crashy. As time went on and software was updated that machine became less crashy but performace was always poor.

I don't have any insider information, but I wouldn't be surprised if Apple felt like they needed a backup plan in case the PowerPC alliance was unsuccessful at matching the speed growth of Intel's products. I'd say the G3 was the first PowerPC CPU that impressed me, I think Apple only stuck with PowerPC for two more generations (the G5).

View on HN

I got one of the first PowerPC computers in a lab full of high end spare no expense PCs and it cranked. If your experience was otherwise, I suspect you were running software under emulation so the fact it was “no better” was a borderline miracle. The second gen PowerPC macs were “the fastest 680x0 boxes ever shipped” (an Apple engineer boasted to me at WWDC and he wasn’t wrong)

The PowerPC allowed Apple to overtake the Pentium and gave it boasting rights until the failure of the G5.

And the G5 really only failed because of cooling issues - obviously, IBM has been able to continue to push the POWER platform that undergirded PowerPC far beyond the G5

Not just cooling -- the G5 drew too much power to be viable in one of Apple's growth markets at the time: laptops. The Intel Core chips were based on the Pentium M microarchitecture and so already had power characteristics that were suited for laptop use.

Of course, POWER would go on to be a great arch for desktop and server use, but that wasn't enough for Apple after the G4.

Excessive power draw ... which, while seen in abysmal battery life, was most directly seen in an inability to cool

Yes, I believe the majority of software and even parts of the OS ran in emulation. Still, even when they moved a lot of that code native, performance wasn't amazing. Someone else mentioned that the rest of the hardware may have been holding these machines back and I could easily believe that.

My opinion: when released, performance didn't meet my expectations and failed to meet my expectations for another five years (when the G3 was released). IMHO, I don't believe the Mac was competitive against the PC until the release of the G4. I could understand why Apple would want to hedge against the PowerPC.

That's funny, I too remember things differently. As Mac advocates often, and loudly pointed out, the Mac consistently smoked the PC -- especially at the only benchmark that mattered, Photoshop filters. It was the 90s and RISC architecture really was gonna change everything. At the time, Intel was playing the "megahertz myth" heavily in its marketing, the false belief that more MHz made for a faster CPU, which just wasn't true when comparing across different microarchitectures like Pentium and PowerPC.

And of course, as those from the Lost Amiga Civilization recall, it took about a 500 MHz Windows 9x PC to feel as responsive as a 25 MHz Amiga.

It's been a long time, my memory is less a collection of data and more a hazy soup of anecdotes.

I do suspect that Photoshop filters might get some real performance gains out of AltiVec instructions, maybe my boring developer style workloads weren't reaping similar benefits from the availability of those instructions.

What the G4 giveth, Objective-C/Carbon/Cocoa taketh away.

The computer lab manager at the Uni I went to was obsessed with Apple, and he had one of the 601 based PowerMac's as well. And yes, I can confirm how underwhelming it was. Granted much of the OS was 68000 based, which certainly didn't help things in the slightest, but even Wolfenstein 3D on the thing really didn't run all that much better than our existing 486DX/2 66's.

Running on Intel certainly saved NeXT in that by version 3.3 with the ability to support IDE CD-ROM's the cost of a machine that could run NeXTSTEP 3.3 had dropped dramatically. It's just a shame that the OS cost nearly as much as the machine.

Being more open like the later Darwin sure may have helped, although I don't think NeXT much like later on Apple really wanted to either take in code from outsiders, or made the open software rediably usable.

The PPC processors were not the problem. The performance issues were caused by both a crippled bus on the first gen PPC computers and MacOS wasn’t fully native.

Right, BeOS ran like a bat out of hell even on low-end 603 PowerPC's. The 68K emulation in MacOS worked surprisingly well in terms of compatibility, but at a significant performance cost.

Most of the software was emulated early on, so not completely unexpected.

Reply on news.ycombinator.com