PowerMac G5: World's Fastest Desktop Computer

[plumbercrackboy]

Originally posted by ddtlm
plumbercrackboy:

despite the artificial classes you are trying to push them into.

I guess only your "classes" are the proper ones. Intel and AMD's classes must be the "artificial" ones. The Xeon is the server version of p4. It is also in high end workstations. The Opteron is the server version of Athlon 64. It is (acually will be) also in high end workstations. The p4 and Athlon 64 are the consumer versions of them. These are the classes the manufacture puts them in. Look at intel.com and amd.com.

What I am saying is that the G5 is a workstation/consumer chip. It is not designed for server performance. The Xeon and Opteron is. Likewise the Power4 is a server chip. Granted the difference between the Xeon and P4 is a lot less than the Power4 and the 970. But it is still the same thing.

Case in point. Who makes an Opteron workstation that you can buy today? The only people I know of is boxxtech. Oh yeah, a comperable system to the dual g5 2.0's in an opteron runs over $4,200 from them.

They were not included in the benchmarks by Apple because they are not available to the general public. No large manufacture puts them in workstations. No one uses them for anything except servers. I am sorry for putting it in the light for you, but it is the truth. They will in the future I am sure. But, why would Apple show benchmarks for something that doesn't exist yet that they did not create.

I did not say that you can't benchmark them. I didn't say that individual reviews cannot benchmark and compare them. I gave a sound reason why Apple did not and should not have compared them.

 

[MacBandit]

Originally posted by ddtlm
mastermix:


I think the author's points are valid, but I'm nonetheless happy to see Apple talk about SPEC. People need to get their hands on these machines before we really know how fast they are.

The author is a moron because he isn't comparing equal compilers. On one hand the benchmarks he is using are from a highly intel optimized compiler which shows in my oppinion falsely high scores on the other hand Apple used a compiler that isn't specifically compiled for any one cpu and therfore shows a better more equal comparrison. In any case screw the benchmarks. I hate comparing computers with benchmarks because I don't run benchmarks all day on my computer I do work on my computer. So therfore I want to see how the computer does in real world performance with Apps that I or anyone else would use. In those cases the G5 thorougly stomped the competition just watch the keynote. Also those calling for an Opteron comparisson please quite the Opteron doesn't perform as well as the Xeon system and the G5 trounced it in real world tests.

Also the author says using a $99 pricing scheme is underhanded and not upfront with the consumer. Give me a break who doesn't know that $99 is $1 different then $100. The whol world uses this pricing scheme so get use to it. I think that argument just underminds the whole report and shows that this guy is just an internet Troll. I highly doubt he is even a real Mac user.

 

[iLilana]

Clustering

I don't know if this was metioned before but I wonder how well these machines would perform in a cluster? 10-50 dual 2 (or 3) gig with 8 gig ram each all clustered for extreme workouts.

 

[ddtlm]

MacBandit:

Look at his article! He accuses Apple of disabling SSE2 and disabling hyperthreading in the SPEC rate test. He says they used a special malloc library for the G5.

Go to the bottom of that article and read the angry Mac people making fools of themselves. Don't be like them. Cool down and read/think about what the guy said.

 

[MacBandit]

Originally posted by ddtlm
MacBandit:

Look at his article! He accuses Apple of disabling SSE2 and disabling hyperthreading in the SPEC rate test. He says they used a special malloc library for the G5.

Go to the bottom of that article and read the angry Mac people making fools of themselves. Don't be like them. Cool down and read/think about what the guy said.

I read the article yesterday. The guy is blowing hot air and starting rumors about things he doesn't actually know about. I'm not saying Apple isn't masking the downfalls of the new computer it would be unlike any company to not do that. I do think that benchmarks are useless though. The only way to really know real world performance is to test real world apps.

 

[frozenstar]

Originally posted by MacBandit
I read the article yesterday. The guy is blowing hot air and starting rumors about things he doesn't actually know about. I'm not saying Apple isn't masking the downfalls of the new computer it would be unlike any company to not do that. I do think that benchmarks are useless though. The only way to really know real world performance is to test real world apps.

It's convenient to ignore benchmarks that are not in favor of your preferred platform.
He is using facts and publically available data to support his claims. So far, everyone that has challenged his claims has done nothing but bitch and moan about him being a "Mac hater".

 

[Wonder Boy]

I've made decision...

I DON'T LIKE THE NEW COMPUTERS.

There, I said it, and I feel great.

1. One optical drive = garbage
2. 2 HD bays = trash
3. rediculous price = classic apple
4. cheese grate front = smelly
5. handles = they just look bad
6. the all around design looks cheap

repramand me, hunt me down and string me from my nickers. i don't care. after thinking long and hard on the subject, ive decided i don't like them.

PS. plus paying fiull price for panther is (explitive) terrible on apples part. greedy bastards.

PPS. Ill still get one next year.

PPPS. I am not a good speller, sorry.

 

[Cubeboy]

It's easy to lower a chips score on SPEC simply with compile flags, code scheduling, instruction selection, etc theirs plenty of things you can do to make a Pentium 4 score poorly. However it's very difficult or downright impossible to artificially increase the performance of a cpu by a significant amount mostly due to the nature of SPEC CPU2000. Perhaps I'll go more into this later.

 

[Frohickey]

Re: I've made decision...

Originally posted by Wonder Boy
I DON'T LIKE THE NEW COMPUTERS.

There, I said it, and I feel great.

repramand me, hunt me down and string me from my nickers. i don't care. after thinking long and hard on the subject, ive decided i don't like them.

No one here is forcing you to like the new PowerMac.
No one here is forcing you to buy the new PowerMac.

Thats why freedom is so great. You can be a putz, and I can call you one, and there ain't nothing the government can do about it.

Just kidding.

 

[frozenstar]

Re: I've made decision...

Originally posted by Wonder Boy
the all around design looks cheap

If aircraft-grade anodized aluminum alloy looks cheap to you, then I don't even want to know what you thought of the transparent polycarbonate used in the previous generation of Power Macs.

 

[ddtlm]

plumbercrackboy:

Now you're just being annoying. The Xeon is a Pentium 4 with a higher price, some non-performance modifications, and SMP support, nothing more. Xeons are often used in servers but that does not mean that they are "server chips" and more than the when the G5 will be a "server chip" when shows up in Apple and IBM servers. You are essentially claiming that a purpose-built server chip (Power4) loses its cache and suddenly is less designed for server performance than a Pentium 4 with a price markup. I have no idea why you are even trying to debate this, the case is and always was closed. People who want a powerful workstation are not going to ignore either G5s or Xeons/Opterons because of the wall you are trying to errect between them.

Oh yeah, a comperable system to the dual g5 2.0's in an opteron runs over $4,200 from them.

How about all those $1500 P4s that is supposedly compeditors to the $3000 dual G5? Does that argument only work when it favors Apple?

They were not included in the benchmarks by Apple because they are not available to the general public.

So Boxx would refuse my check if I sent it to them? I bet they'd have the Opteron to me before Apple could get me a dual G5.

I gave a sound reason why Apple did not and should not have compared them.

You stated the excuse that Apple would state, but the fact is that the Opterons are going to be available to compete against the G5's in the 64-bit arena probably from the time the first G5 ships. Of course Apple wants to ignore them because it allows them to display top SPEC scores and it allows them to drum up their 64-bit achievement as much as they can.

 

[jcormac]

If my standard for buying a computer was that I'll only purchase one with a processor made by someone who doesn't "screw with benchmarks" in someone's view as part of marketing then I'm probably going to end up keeping my 5 year old box.

Just a thought....there's an interesting note on arstechnica today along the same lines.

 

[Sadu]

That article cracks me up. I would take everything apple says with a grain of salt until we get some non-biased benchmarking done (not that such a thing exists of course). I mean WTF disabling hyperthreading on the intels and tweaking the registers on the mac?

If this was such a good performance enhancing idea, then maybe they should do that tweak as standard, rather than just applying the tweak for the benchmark then turning it off for distribution.

They should perform the tests "out of the box" or the results do not mean anything.

and lol at all the mac zealots at the bottom ignoring a perfectly valid argument and flaming the author with personal criticism.

 

[jcormac]

Agreed. Real world testing is best.

My only point was that a poster indicated that he was going to buy a G5 but now was not because of what he viewed as excessive puffery on Apple's part. "Messing with benchmarks? In the PC industry?" I'm shocked, positively shocked.


That everyone may do it does not make it right. It simply means that you have to do your homework and figure out which system best fits your needs. Given the platform differences, comparing benchmarks becomes a bit tricky.

 

[Maxkraft]

the opetron is'nt that great

The opetron has been soundly beaten by the AMD barton chip on several bechmarcks.

 

[gezzas525]

the opterton really shows its stuff in mp configurations, check out the quad opteron SPEC scores there phenominal.

the integer bench mark is faster than the quad CPU ITANIUM2, remember the ITANIUMS are $3000 per CPU !!! OUCH

KLEOS

 

[Maxkraft]

here

This is in case people don't believe me that the opteron can be decked by an athlon.

http://www.tomshardware.com/cpu/20030422/opteron-23.html#3drendering

 

[gezzas525]

Re: here

Originally posted by Maxkraft
This is in case people don't believe me that the opteron can be decked by an athlon.

http://www.tomshardware.com/cpu/20030422/opteron-23.html#3drendering

didnt you read what i said, OPTERONS PERFORM BETTER in a MP environoment. You have to consider also that they are still not mature and will eventually pull ahead with time as AMD optimise the design including BIOS's/MEMORY TIMING ETC...

Have you used any 3d APPS in all cases they recommend a DUAL PROCESSER MACHINE so that it is possible to work and render at the same time. If your doing any kind of CAD work you should be using a DP machine anyway, so I realy cant see where your coming from.


KLEOS

 

[jbomber]

Re: case design

Originally posted by andrewh
Initially, I was disappointed with the case design. I thought it could have been cooler looking. But, I think it will grow on us. I like the way it looks very industrial -- clean front, no logos, etc. Mesh is awesome. The more I look at it the more I like it. I wonder if it was Jonathan Ives that designed it again.

Anyway I suppose it's what's inside that counts...

Sorry if I'm raining on anyone's parade, but I've been holding my breath for this machine to come out for months. When it did, I felt like I'd completely wasted my time. Don't get me wrong, the specs are AMAZING and the processor will CRUSH anything out there, but goddamn, this is an ugly little beast.

It's not too terrible from the side, but from the front? ugh.... I feel like it's an ugly bird that fell out of an ugly tree, and hit every ugly branch on the way down before breaking it's ugly neck. I can't believe that Jonathan Ives presented this thing to Steve Jobs with a straight face.

Whew! just had to get that off my chest...

 

[Zaid]

Originally posted by gezzas525
OK i agree zaid, the larger cache on the xeon does make quite alot of difference but again it depends on the application. But xeons are ment for the workstation market and have a price tag to match, ok i see what your saying as apple computers are expensive they can be compared to xeon workstations??

kleos

In terms of comparison on performance, i think the Xeon is a fair comparison. The reason that the dual Xeons are being compared on price is simply because the pentium can't be used in dual proc configurations. Sure Xeons are more expensive, but thats the only Intel machine that is comprable on specs, hence a price comparison is perfectly valid.

 

[mjones4th]

Kids, kids, kids..... KIDS?!?!?!?!

Calm down now. Okay? Everybody's got their panties in a bind. Okay? Relax, relate, release. Then repeat.

Who gives a flyin fizzle about a benchmark.

I know a 3GHz P4 can smoke my G4 733 w/no L3. Who gives a ****?

I run a music studio on an antiquated (not really, but I have the upgrade bug! Gee Five, Here I come, but don't tell my pregnant wife!) Quicksilver, two years old and still going.

Every time a new machine comes out, all the speed geeks tout the fact that "MY NEW SPACELY SPROCKET V6 CAN RUN 1.5% FASTER THAN THE OLD V5 _AND_ IT SMOKES THE PANTS OFF YOUR PIECE OF **** ONE YEAR OLD GADGET V4! THAT WAS SOOO LAST YEAR! HA HA!!!!"

I must admit, I am a speed geek too, but reality check folks. I've done hundreds (maybe close to 300) of songs on my 733 in the past two years and I have pushed the processor to the limit less than 20 times.

OKay?

(When I feel the need, I'll upgrade the damn processor. But Hip Hop is not a needy music, my compositions are normally pretty thin.)

Now, I did used to overload the drives a lot, until I upgraded to 2 120 GB drives on a ATA controller card. Now it still happens, but much less frequently.

I sometimes run out of RAM, but my 640 usually does me right most of the time.

I want a G5 too, I've been following the rumors for the past 10 months too.

But do I need it? no

Do I care if its really faster than an Opteron or a Xeon? no
(well do opterons and Xeons run OS X? And what's the real reason we all have macs kids?)

I'd be willing to bet that for the most part, all of your current computers meet your needs adequately right now.

So why, oh why, dear children, are we arguing about specs?

Get over it, I already am.





Oh and btw, on the PCI slot and HD expansion topic:

I have 3 drives
60
+120
+120

=300

Imagine this: One 100 story building vs. four 25 story buildings. All other things being equal (same # of staircases, etc), which gets the people out faster in the event of fires?

the trick is not storage size, but storage speed and bandwidth.
If my 120s are pumping 40MB/s each, on separate controllers, that is better for me, than one 250GB drive pumping maybe 50MB/s

_AND_ SCSI costs too damn much for me and doesn't provide the real world gains that the costs imply. I mean TOTL (top of the line, since we all love acronyms) Cheetahs may double my track count, but for 7 times the cost. That math actually works on mars. ($300 for 2 120s and a controller, or $2000 for two 73s and a controller)

Got it?

Okay, in audio, this is important. I have my audio streaming load spread out over the two drives (they're not RAIDed) and it does the job for me.

I'm sure it matters in video also, B/C video is audio on steroids.

And the 60 gig drive hosts the OS and Logic 6. (I'd really like to have separate drives for those two)

One PCI for an Audio card, maybe even two, depending on your I/O needs. (I have two, M-Audio Delta66 and Korg OASYS PCI)

One PCI slot for my ATA controller.

One PCI slot for a FW800 card (coming soon) so I can set up a CD burning farm w/4 FW burners and my internal burner. Two burners on one FW channel slow each other down from about 48X each to about 32X (yes I measured), so it is a realistic need.

So that's the 4 I have in the QS right? But...

What about that UAD-1 card from Mackie I want? What about the TC Powercore? What about adding a third monitor? www.emagic.de had a pic of a logic system running 3 monitors for the longest time.

It may be irrational to want these things, and I realize that not having them is not a show stopper, but why the **** are you guys beating people over the head for wanting one or two more PCI slots, or space for one more HDD?

Grow up.

Let the flames begin

I made it Ma! Top o' The World!

 

[hacurio1]

Originally posted by ddtlm
MacBandit:

Look at his article! He accuses Apple of disabling SSE2 and disabling hyperthreading in the SPEC rate test. He says they used a special malloc library for the G5.

Go to the bottom of that article and read the angry Mac people making fools of themselves. Don't be like them. Cool down and read/think about what the guy said.

Yes, but he also neglects to mention that the PIV and Xeon machines had 2GB of RAM and that the 970 had only 1.5. I know this wouldn't have made a big difference, but I'm going trough the 30 page Veritest PDF and there are a lot of things this guy neglects. Also, he assumes the malloc library is improving the performance of the 970. As far as I know, malloc librarys are used only for memory allocation. I'm not sure though. I don't believe this guy and I'm going through the Veritest PDF point by point. I'll come back with results tonight or tomorrow, it will take a while.

The author at http://www.haxial.com/spls-soapbox/apple-powermac-G5/ states: “Using publicly available information, I am going to show you how Apple is attempting to deliberately mislead its loyal customers and fans when it claims that "The Power Mac G5 is the world's fastest personal computer.""

But he also mentions that: “Apple/Veritest used some special hardware tweaks that end-users may or may not be able to use. I don't know ENOUGH about these hardware tweaks to know if they are fair, so I will just mention that it *might* be a cheat.”

Clearly we have to spend some time analyzing and doing some research by our selves. What he wrote is his opinion, but he is no expert. If somebody is up to the task, I propose: LETS GET TO THE BOTTOM OF THIS!

 

[Cubeboy]

Originally posted by hacurio1
Yes, but he also neglects to mention that the PIV and Xeon machines had 2GB of RAM and that the 970 had only 1.5. I know this wouldn't have made a big difference, but I'm going trough the 30 page Veritest PDF and there are a lot of things this guy neglects. Also, he assumes the malloc library is improving the performance of the 970. As far as I know, malloc librarys are used only for memory allocation. I'm not sure though. I don't believe this guy and I'm going through the Veritest PDF point by point. I'll come back with results tonight or tomorrow, it will take a while.

The author at http://www.haxial.com/spls-soapbox/apple-powermac-G5/ states: “Using publicly available information, I am going to show you how Apple is attempting to deliberately mislead its loyal customers and fans when it claims that "The Power Mac G5 is the world's fastest personal computer.""

But he also mentions that: “Apple/Veritest used some special hardware tweaks that end-users may or may not be able to use. I don't know ENOUGH about these hardware tweaks to know if they are fair, so I will just mention that it *might* be a cheat.”

Clearly we have to spend some time analyzing and doing some research by our selves. What he wrote is his opinion, but he is no expert. If somebody is up to the task, I propose: LETS GET TO THE BOTTOM OF THIS!

Read my analysis on page 14 of this thread, the SPECmarks (except for the rates) seem to match up with IBM's so I would say they are accurate and fair, the problem is, the comparison was with a older 3.06 GHz Pentium 4b, the newer Pentium 4c's do significantly better.

 

[dethl]

I've been waiting for the G5 since I've heard of it. I've been wanting to purchase a new computer since last year, but I have held off. I love this machine. I cannot wait to have it at my house. This computer along with Panther will be a programmers dream...I will enjoy college very much so.

The case design is totally different than what we have all have been familiar with. I personally like it, and I saw a live shot of it...and it looked much better than the Apple Store photos.

Everyone ranting about how ugly the case looks, think about this: Apple has finally come out with the G5. I could really care less about the case as long as whats inside will run with no problems.

On another note: I went to the link a few posts above (Opteron vs Xeon link) and yes, the Opteron got smoked in the Workstation category.

Apple has given us the first CONSUMER (for regular use, not workstation/server, etc) 64-bit machine. I'm very happy at what Apple has done.

Now...I do have one problem: Us G5 buyers will not be getting Panther, but rather 10.2.7 (Smeagol?)..and there is (for now) no option for getting Panther at a reduced or free rate. I'm kinda irked...but at least I'm going to get my machine.

Ok, I've typed enough....have fun
*Dons flameproof suit*

 

[Brother Mugga]

Hmmmmmm...

From Jbomber:

It's not too terrible from the side, but from the front? ugh.... I feel like it's an ugly bird that fell out of an ugly tree, and hit every ugly branch on the way down before breaking it's ugly neck. I can't believe that Jonathan Ives presented this thing to Steve Jobs with a straight face.

Whew! just had to get that off my chest...

I know what you mean. If you managed to catch the conference feed, there was a definite "......" from the audience when it was unveiled.

HOWever...

I actually quite like it; it's kind of minimalist (whatever that means, and judging from the state of my study, I clearly have not a bleedin' clue). Ives described it as "essential" and, design-technobabble aside, I *think* I can see what he means. It's really stripped down and functionally linear.

Er...not quite sure what that means either, now I come to think of it. This isn't going too well.

And my internal monologue seems to have broken down as well.

Aaaanyway...

As I've said previously, I'd prefer it to have had maybe an Apple logo picked out in denser mesh on the front, but generally it does the business for me (especially when I realised the grill allows you to see the machines innards in a *slightly* gay mesh-tee-shirty-pecks-bulging kind of way) - see:

http://www.spymac.com/gallery/showphoto.php?photo=36511&papass=&sort=1&size=big&thecat=

for a photo.

I think it might grow on us all.

And, to be honest, if the design genuinely lets it run at around 35db then they could paint it s**te-brown and make it out of kleenex and spit for all I care...

Brother Mugga