Power Mac G5s are fastest machines in world

Hi Folks,

I thought I'd write in regards to your recent article titled "Apple Slammed for misleading Power G5 Adverts"

First off, I'd like to point out there is no such thing as a "Power G5". It seems like you are confused between the Power5 IBM processor and the IBM PowerPC 970 (aka G5) The IBM Power5 is the latest architecture from IBM, high end server CPU's, the PowerPC 970 (aka G5) is a Desktop CPU (somewhat cut down Power4 CPU), less cache and dual cores, but plus altivec.

It does indeed allow more than 4GB main memory, infact the slots in the machine provide for more than that. I have one. You probably don't. On demand, means you put more in when you need it. Get it ? As for the fastest desktop claim, I don't know about this claim, All I can tell you is about my experience.

For example, I have a dual 2.4ghz Xeon machine, compiling source navigator 5.2b takes 222 seconds, on a 2ghz G5 it takes 180 seconds, so it's faster than a dual 2.4ghz Xeon at least. I think in my experiments it's the fastest machine we have in the office, sometimes over twice as fast as the Xeon, other times only marginally, but always faster, and for less CPU clock speed.

We are a demanding software development house, mostly windows users, a few linux power users, but our numbers speak for themselves. The G5 is fastest. Period.

I understand that some people in the UK believe that the G5 is not the fastest desktop in the world, so be it, that's what they believe. I also believe that Santa Claus, Superman and the loch ness monster don't exist but I don't have my views published without proof.

I'm ok with with some Windows point-and-drool person thinking they have the fastest machine on the planet, it's OK, really, I'm a software engineer and live in reality, not "weblog propaganda" and I can't help it if someone bought the wrong machine. We have many machines for different purposes, the G5 is a high-end machine and not for everyone.

What I would say to the enquirer is, before posting an article like this you should do the research, post the numbers. Get one, try it, really, not spec numbers... compile something, do something. Look past the pretty cheesegrater front... why don't you put your money where your mouth is ?

Get sourcenav from sf.net, compile it, do something real, The G5 is the fastest machine I've seen (other than Power5's - which ... errr...errr.... don't you think will make it to the Apple machines ?) - Intel is running out of steam, approaching 4ghz they are only marginally faster than the latest 2.5ghz PowerPC 970's (cut down Power4). Imagine what we have coming next year.

Best regards
Simon Spooner

A "Power G5" (no such thing) owner.
PS : My G5 does 2 seti jobs in 2 hours (and it's not even altivec or G5 optimized) , how many does your Athlon 64 do ?

More Mac Talk

Hi Mike,

Thanks for the reference to the ASA!

I read it and indeed, to my surprise, the ASA complained about the following:

Complaint: Objections to two magazine advertisements for the Power Mac G5 PC. One stated "The new Power Mac G5 is here. It's the world's fastest computer, and the first with a 64-bit processor ...". The other advertisement stated "... Introducing the revolutionary PowerPC G5 processor, the world's first 64-bit processor for personal computers ... the G5's 64-bit architecture addresses dramatically more memory - over 4 billion times more than 32-bit chips - so that the systems built around the G5 can shatter the 4-gigabyte memory ceiling that limits every other PC on earth ...". The complainants challenged the claims:

1. "the world's fastest personal computer";

2. "the first with a 64-bit processor" and

3. "the systems built around the G5 can shatter the 4-gigabyte memory ceiling that limits every other PC on earth".

It seems that Apple is publishing different things in different countries. The ASA complaint seems to be related to a publication done by a London-based agency. Do you know which magazine this was?

Point is, Apple has not been this blatant in other countries, at least not in the ones I'm looking at, which is the Netherlands, the US and Sweden. It seems that Apple UK is designing these ads, or approving the agency's texts, which are indeed false. And here's me getting my knickers in a twist thinking that you are unfairly and incorrectly quoting Apple.

On Apple's US and Dutch web site and in local magazines and publications, Apple never claimed to be "the world's fastest computer", and "the first with a 64-bit processor", which is clearly not true. What Apple claimed here and in the US was "World's fastest personal computer", and "world's first personal computer with a 64-bit processor". I remember looking for this distinction very closely at the time. Apple UK or their agency must have been using a little too much poetic freedom.

Also, I don't recall any publication that stated: "the G5's 64-bit architecture addresses dramatically more memory - over 4 billion times more than 32-bit chips". This is of course also not true. Apparently they were assuming that the G5 actually has a 64-bit addressing capability, which it does not. The G5 CPU does have 64-bit addressing internally, but only 42 address lines come out of the chip. Of these, only 36 are actually wired to the U3 memory controller chip, thus limiting the amount of actual memory to 68GB, assuming you could get 8GB DIMM's.

Maybe it makes sense to include a reference to a specific publication next time?

Thanks & regards,
Maarten

No sex please, we are nanobots

Nick,

Perhaps while we're on the subject of "fathers" of technologies, we should remind our readers that it was in fact Richard Feynman's famous 1959 lecture on nanotechnology at Caltech to the American Physical Society that really launched the field. The transcript of the lecture can be found here: here.

Cheers
Peter Dudek
Laboratory of Dr. Didier Picard

Valve and Theft

Charlie, as a respected journalist, I believe you should understand the importance of using the correct terminology when applied to software an information.

Your article of Friday, 11th of June titled "Valve thieves captured", contains a variety of factual errors, due in no small part to your terminology.

Theft, by definition, is the removal of property from the victim, and complete transfer of that property into the possession of the thief. Valve software was not the victim of the theft. Valve software was the victim of a security breach resulting in copyright infringement. Both are crimes, yes. Some may even say that morally, copyright infringement is as bad as theft. But they are not the same thing. To say that Valve was the victim of theft, even if that is what they (incorrectly) insist, is to imply that someone physically broke in, took every hard drive and every backup tape containing the source code and left. That would be theft. That is also not the case.

Of course many people have joked that since the code was "stolen", Valve no longer had the property, and so was unable to work on the HL2, until now, as the "thieves" have been captured and the code has been returned.

Please make sure that you are aware of the misinformation you are spreading around - news can be an educational resource for many people. And you're telling people the wrong things.

Best regards,
Glenn Birt

Linus invulnerable to Worms

In this piece Nick Farrell says:

"it does shake the complacency of Linux users. Some think they are almost invulnerable to the sorts of attacks suffered by Windows users"

Most of us know that we are not almost, but completely, "invulnerable to the sorts of attacks suffered by Windows users"

One of the things that really ticks us Linux users off, is when we hear so called "experts" opine that the only difference between Linux and Windows security is the number of attacks made against the respective platforms.

Linux is not invulnerable, but it is a straight forward, indisputable, fact that the kind of trivially exploitable vulnerabilities which make Windows a spam/virus heaven simply do no occur on any Unix-like system.

Best wishes,
Dave

Apples and Pears

Is there a place I can find some good, unbiased opinion and reviews of the Mac hardware compared to PC hardware. Because I have to say Apple's marketing department throws around some big numbers.

Like "Its frontside bus works at speeds up to 1.25GHz for an astounding 10GBps of total bandwidth. That makes it 450MHz faster than even the latest Intel 875 chipset, which sputters out at 800MHz.". -Apples marketing page for the G5 FSB.

Right now for graphical apps like Photoshop that's probably the most significant hindrance to performance on the PC platform. But to put out a number like 1.25GHz but ship it with 400 MHz RAM, doesn't seem to make sense. I mean does Apple's FSB even allow for a dual channel RAM configuration? Where does this 1.25GHz number come from? And what does it mean?

How do they arive at the 10GBps number when even their own website states. "A fast 128-bit DDR memory interface provides up to 6.4 GBps of memory throughput." -The configuration page for the 2.5GHz G5

What if someone bought RAM rated @ 533Mhz, and put it in there? Would the Mac take advantage of the extra headroom on its FSB and allow you to use a larger portion of the 10GBs of bandwidth it has available? How would that compare to an Opteron, 939, 754, XP, Northwood, Prescott with similar configurations?

Another point, why use an app like Photoshop to benchmark, when you have an OS like Linux that works for both architectures, and has tons of programs that will run on both architectures? Why not use the Gimp, and Blender to bench these systems? I really would like to see how the two architectures perform in a similar environment.

Honestly though getting info on Intel/AMD on an Apple website, or getting info on Apple from a site called AMDzone, doesn't seem to the best way to find unbiased reviews.

Matthew