Motorola has released the latest update to its PowerPC 8500 – aka G5 – processor that ups AltiVec performance and delivers consistent 1GHz and up clock speeds, TheRegister rumours, based on their Apple sources. “Indeed, the source claims, two of the chips in the sample set of CPUs sent to the Mac maker, clocked at 2.4GHz. Most, however, ran at 1GHz, 1.2GHz or 1.4GHz, and some – a “considerable number”, says our Deep Throat – operate at 1.6GHz.” In related news, Advanced Micro Devices (AMD) has again upped the stakes in its processor performance race with rival Intel.
AMD launched the new Athlon XP 1900+, its highest performance desktop processor issued to date.
Preliminary PowerPC G5 Hits 2.4GHz
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
So… what is the G5 pipeline lenght? GHz without the lenght of the pipeline means nothing. Take the P4, for example… Yes, it runs at 2GHz, but it’s pipeline lenght is 22!!!!!!!!! By comparison, the PPC G4 runs a 800MHz, but with a 7 stages pipeline.
>So… what is the G5 pipeline lenght?
All the information that is known about the G5 today is linked from OSNews (the CPU is not officially released yet – little information is known).
Please type: G5
on the search box above and click the links of these returned stories to read more about the CPU itself.
I do not remember where I read it but the G5 is rumored to have a 10 stage pipelene. The longer your pipeline, the more you pay for branch misprediction, which is one of the major problems with the P4 with 20+. Sure, longer pipelines make your MHz go up, but peformance per Hz goes down as even with 95% accuracy, the 5% branch misprediction with a 20+ stage pipeline causes a major performance hit. Even wonder why many x86 server manufacturers still use the P3? Anyway, I can’t wait to get my hands on the G5.
Wow! It would be great if they could get the processors to run that fast!
Btw, Eugenia, have you heard anything about the VIA C3 processor? I’ve heard it mentioned a few times but I can’t seem to find any good articles on it…
Kevin, please http://www.theregister.co.uk/“>go and enter:
VIA C3
on their search box. There are at least 7-8 articles there about it.
It looks like processor leaps are becoming irrelevant. My dual Cel 400 gets by just fine. I have a 1Ghz Athlon, at work, that performs about the same. And, supposedly, the higher clock speeds aren’t gaining the same ratio of performance/clock that they used to. I’d have to agree with sean_r, above, if I upgrade I’ll be getting a couple of fast P3’s to replace my Cel’s. But, more likely, I’ll be getting a next gen machine before I ditch this one. This type of architecture can’t last for long, it’s getting boring. P4 is a snore.
[quote] Even wonder why many x86 server manufacturers still use the P3? [/quote]
Mainly because P4’s (apart from Xeon P4’s) don’t support SMP
From the November issue of <u>MaximumPC</u>:
The VIA C3 “Nehemiah” is expected to arrive late this/early next year.
(Guessed) Specs: 0.13-micron process, 900-1200 MHz, 256K on-die L2 Cache,
Socket 370 interface, SSE instruction set.
If the pricing is right it sounds like it could be a nice upgrade to a Celeron
or old P3 machine. So far the C3’s have been performance limited by low L2
cache sizes and failing to keep up in the MHz arms race.
I have never been totally clear on this: Is it possible to get a PPC chip
machine from someone other than Apple? If one could get such a machine at a
nonextortionate premium vs. an x86 machine and intall one of the PPC Linuxes,
the speed and lower power consumption may be worth it.
>> Even wonder why many x86 server manufacturers still use the P3?
> Mainly because P4’s (apart from Xeon P4’s) don’t support SMP
Sorry, I should have said PIII Xeon vs P4 Xeon, but that is my point
Intel would much rather have everyone on board with the P4, and dont think they have not pressured vendors to do so, but many major vendors are making the transition much more slowly than Intel wants and have not annouced any P4 Xeon product lines yet. Eventually most will switch, but Intel has lost its edge in recent years. I personally am really looking forward to the G5 and Hammer, from the reported SPEC benchmarks and what news has been releasod about the architectures so far, these should be excellent CPUs. I hope FreeBSD gets a 64 bit port to the G5 quickly, but I am sure NetBSD will have one as soon as the G5 avaliable or even before 
Unfortunately the register is about as trustworthy on this as my magic 8-ball.
The same source that claims these things has claimed several things that just can’t be true.