Bigger performance leap: PPC-to-Intel or Intel-to-ARM?

[ThomasJL]

Which of Apple's processor transitions, at the time of the transition, offered a bigger leap in performance: the PPC-to-Intel transition, or the Intel-to-ARM transition?

In addition to voting, it is appreciated if you please explain your thoughts comparing how big of a difference there was regarding performance in two transitions.

 

[MK500]

Don’t forget the 68K to PPC transition. That one was a doozy!

History Repeating? (68K --> PPC --> Intel --> ARM)

Scenario: Intel (IBM in 2004) hasn't been able to bring the thermals down in order to ensure a powerful, thin and light laptop that sips battery. The Core series (G5 in 2004) is simply too hot even at 10nm (90nm in 2004). There have been promises of a cooler-powered CPU with the next 7nm Core...

forums.macrumors.com

Having been through them all as an IT consultant, I can assure you this one is the best. Apple has learned from each. They have it down at this point. So much just works silently without you even realizing there is translation happening. Also the performance difference is much bigger this time.

 

[Anonymous Freak]

The thing is, we don't know how *THIS* transition is going to play out yet. Like with the Intel transition, they're doing the lower-power systems first.

When the Intel transition happened, you had 64-bit CPUs at 2.7 GHz dual-socket or dual-core 2.5 GHz dual-socket setups supporting up to 16 GB of RAM with a *FAST* connection to the rest of the system in the Power Mac G5.

But the launch Intel Macs were 32-bit 2.16 GHz dual-core, topping out at 2 GB RAM with both the MacBook Pro and iMac.

The MacBook Pro may have been a *HUGE* boost compared to the PowerBook G4 it replaced; but the iMac was a sideways move at best compared to the iMac G5. It wasn't until the second generation that the Intel systems truly leaped ahead.

Same with the 68k to PowerPC transition. The launch lineup of PowerPC 601 CPUs topping out at 80 MHz were a bit anemic due to architectural issues, especially when you take in to account most software was still 68k code, and frequently slower than a 68040 at half the MHz speed. The later 603 and 604 systems are when its potential really took off.

We're seeing the same here. Yes, in some workloads, M1 is *WAY* faster than the CPU it replaced in the systems it replaced them in (like PowerBook G4 -> MacBook Pro.) But M1 isn't ready for the true Pro uses yet.

I'm sure it *WILL* be there. It just isn't yet.

 

[revs]

The PPC to Intel transition let me move from a PowerMac to a laptop and have about the same performance in a way smaller device.

So the device got smaller, but performance was about equal, if I remember correctly.

For the M1 we are getting faster (single code, and mostly multi-code) performance. The M1 doesnt beat the iMac Pro/Mac Pro multicore at the moment. Clearly it will.


This transition seems so much faster and easier than the intel transition. I am seeing complaints online that companies are taking *days* to release an ARM version... it took months/years for intel versions of some software.

Although I did the 68k to PPC transition, I had a 68k machine for a long time, and didnt jump to PPC until the G3 & G4, by then the software had updated, so I skipped that stage.

 

[matrix07]

ARM transition is like Intel transition on steroid.