[Salty]

Is it just me... or wasn't the G5 supposed to be like almost twice as fast as a similarly clocked G4. Did the G4 suddenly just get way more awesome, or did the G5 suddenly start sucking or what...?

[Al G]

Originally Posted by Todd Madson

I'd say of the last twenty laptops I've set up for work, the vast majority are
in the 1.6-1.8 ghz range. Not one has been over 2 ghz yet except one and
that was someone with deep pockets.

It's still competitive. Especially if it's dual core.

But a G4 or G5 at the same clock rate as a Pentium M is really not competitive. I don't think XP feels all that fast but in terms of raw number-crunching computing power, a Pentium M Dothan at 1.6-1.8GHz is startlingly fast. And that's while pulling either 21 or 27 watts, depending on the model.

I have a 1.67GHz G4 Powerbook and it feels quite Snappy™ running OS X. But for pure CPU power (something like Folding@Home for example), a 1.7GHz Dothan spanks the G4 in quite a spectacular fashion.

[Al G]

The G5 is twice as fast as the G4 in floating point, but for everything else, not so much.

Memory bandwidth is better too but we're talking about very different memory-related architectures.

[G5man]

does the G5 run at full speed even thoguh you have a loaded hard drive?

[Salty]

This doesn't make any sense. The G5 when it wasn't out yet was all anyone could talk about because it was some God like chip... now apparently we've decided it has faults or... huh? Blah you're all soo bloody fickle.

[Catfish_Man]

Salty: The G5 was a nice improvement for the high end over the aging G4. Unfortunately, it then failed to scale as expected, and used a lot of power, so by comparison to other high end modern chips it's not spectacular. *If* it had done as promised (i.e. 3GHz over a year ago), it'd still be kicking ass.

[mousehouse]

IMHO the whole issue doesn't seem to be _if_ IBM can deliver but _when_ it can deliver. Steve wants to show nice incremental speed increases and have a steady release schedule of nice new "i want" machines and gadgets. IBM was unable to implemenent the complete spectrum of requirements put forward by Apple, such as the low-power version of the G5, faster and dual-core G5's etc. I expect the Apple Labs to be filled with very cool prototypes of PowerBook G5 logic boards and Quad G5 Macs. But for Apple it's all about volume and steady improvements which IBM has shown are quite hard to achieve.

Apple likes to have 6-month release schedules, which is something that IBM obviously doesn't match with, and Intel does. The 970MP will be here, same for the low-power G5. They will all be here, but by the time there is sufficient quantity you'll be happily typing along on your Intel-powered 64-bit Leopard machine. Going the Intel route allows Apple to focus more on the rough edges (custom board design, the OS, iApps, etc) because it gains the complete solutions arsenal that Intel has. And it comes with the nice bonus of having new stuff every few months.

But back on topic,
Any chance of seeing Quad G5's in the next 6 months?

[Simon]

Originally Posted by mousehouse

Apple likes to have 6-month release schedules, which is something that IBM obviously doesn't match with, and Intel does.

I agree with you that IBM probably can't match Intel's steady update pace and that Apple therefor also felt encouraged to switch.

However, IMHO in this six month update cycle also lies maybe one of the most dangerous marketing aspects of the Intel Macintosh. Up to now, Apple updated their lines roughly every six months or twice a year. Sometimes it took longer, but sometimes also less. No longer with the Intel Mac; now there's Intel that basically dictates when CPU updates come out and thus also when Apple can update their boxes with new CPU versions.

When Intel starts shipping a new version of a CPU companies like Dell are normally very quick to offer this new CPU on one of their (often already existing) PCs. Apple doesn't work that way. They update when a new system is ready; but often that involves a new case, new ports, new features, etc. How is this going to work out in the future? Will Apple innovate, get a new system ready and then just wait till Intel's got the newest CPU version ready? Or will Apple release new systems whenever they feel like it and put in whatever Intel offers them at that time?

I'm worried about the latter case. Apple's hardware would get a reputation of being behind other manufacturers like Dell. When Intel drops a new CPU version, Dell will offer it a couple of days later and Apple might take another four months because their 6 month cycle says it has to be that way. I doubt that will do Apple's hardware rep any good, after all Apple hardware and PC hardware will become very comparable once they both use the same CPU.

I'm wondering how Apple plans to deal with that.

[mousehouse]

usually during a year there are several speed improvements in the intel arena. the big speedrace being over, and assuming apple will use intel chipsets as well, apple has the possibility to implement quiet speed bumps all year long. they will be slower then Dell, but small speedups usually don't involve a chipset change and are therefore easier to implement (they've become a supply chain issue).

and before intel it was IBM and before that it was Motorola (Freescale) that dictated product development. Mot. caused the whole performance problem (that IBM solved) anyway! next to Intel Apple can still go to AMD if they prove to be the winner in the long run now that x68-64 has basically killed IA64 in the volume market.

the biggest deal will indeed be how Apple can hold up the image of being different when it's obvious that the box can run Windows and applications can be benchmarced on the same machine using a different OS. I sure hope they come out as winner...