problems with the G5 benchmarks?

Re: problems with the G5 benchmarks?

Originally posted by f-matic
apparently some mac watchdog somewhere has compiled a list of reasons why apple's new g5 benchmarks are skewed. it's interesting reading, and some of the points certainly seem valid (hyperthreading disabled on pentium 4???).

here's the slashdot link:
http://apple.slashdot.org/article.pl?sid=03/06/24/1232237&mode=thread&tid=107&tid=187

here's the argument itself:
http://www.haxial.com/spls-soapbox/apple-powermac-G5/

Click to expand...

After releasing these (faked) benchmarks and making those fraudulent claims about "The world's fastest personal computer", I think Steve ought to fire Apple marketing. They will never gain any credibility again.

Why couldn't they just say: This computer is twice as fast as the current PowerMacs? Or, why couldn't they just say: We're almost faster than the fastest P4s, and by the time we hit 3ghz., we will be faster?

Shame on you Apple. The problem is, the average Joe at CompUSA or the Apple store will simply believe their claims when they say "The fastest desktop computer in the world."

I did also notice that every single app used during the bake-off yesterday was optimized for Altivec. Did you notice that as well?

You people slay me. Why can't anyone just ever be happy that we have a new kick ass system. Why does EVERYONE have to find SOMETHING to b*tch about!

The new G5 Rocks! Does anyone really belive that Apple wouldn't steer things their way? Intel, MS, Apple we all do that.

OF COURSE Apple would use AtliVec enabled apps. That's one of things that make Macs faster.

ALSO...there is NO WAY the guy who wrote that on slashdot is a Mac fan... he's a PC User... the whole "...I am a Mac user. I have been using Macs for years. I am writing this article on a PowerMac G4. I enjoy using Macs..." is a dead give away.

Shame on you Apple. The problem is, the average Joe at CompUSA or the Apple store will simply believe their claims when they say "The fastest desktop computer in the world."

Click to expand...

...uhhhh this is bad? Ummm I don't think so! Good I hope they do buy them, increase market share baby!

MacBoyX

It's funny you mentioned this, because when I saw Apple's claims I immediately went off to read the veritest document just to see for myself where they fudged the numbers. Even BEFORE I had read any of this hoopla, I automatically assumed Apple was lying (as always) and went to find out where. The fact that I immediately thought of debunking these claims for my own amusement should suggest how much credibility Apple has lost by using these kind of 'contests' in the past.

The fact of the matter is that Macs have been slower than comparable PCs for several years, and yet, Apple insists on periodically making marketing claims in the area where they are the WEAKEST. It's like someone selling hybrid electric cars trying to market them for their SPEED and POWER instead of their fuel efficiency.

In Apple's defense, ATI and nVidia do similar things, even going so far as to 'cheat' on benchmarks with special optimizations of their hardware to make things look better, however, even THEIR claims aren't as ludicrous as Apple's.

When all's said and done though, a G5 will look pretty damn slick sitting on your desk....

Originally posted by MacBoyX

The new G5 Rocks! Does anyone really belive that Apple wouldn't steer things their way? Intel, MS, Apple we all do that.

Click to expand...

This post actually reminds me of why Apple still uses those particular marketing tactics. I mistakenly thought they were designed to entice PC users to a Mac ('wow, I want the fastest computer'), but let's be realistic, there's not going to be a sizeable exodus from Windows PCs to Macs any time soon.

Apple is preaching to the converted, preaching to them to buy a new system. The 'Mac faithful' (those of you who aren't as religious users must forgive me for this generalization) will swallow anything Apple throws their way, and even after seeing such a claim debunked, they don't really care. 'WOW! This computer beats the pants off a Dual Xeon 3.06, my Winblows friends will be so jealous. I HAVE to have one!'.....

and later on: 'So, Apple was telling a little white lie? So what? The G5 still AWESOME!'

Never mind that when the users buys the 1.6GHz SINGLE processor system that he or she can likely afford, it will have a hard time keeping up with even a modern budget PC.....

When it comes right down to it, performance isn't really a particularly relevant factor in the purchase decisions of the average consumer.

Nobody really knows that the G5/970 does in real life. Staged cook-offs are always going to play to the side of those doing the presentation (duh). Default install of everything, without any special optimizations, on similar hardware (if you can really compare 32bit to 64bit fairly), then we may know something.

I think it is kinda funny that a 64bit workstation (G5 isn't a desktop by current standards) barely eeks out over 32bit desktop (P4).

Quake 3 on a 3Ghz P4 with Radeon 9700pro gets over 400fps, while the 2Ghz dual G5 gets just over 300fps. According to http://www.tomshardware.com the AMD 2700XP is actually 10fps faster than the P4. It looks like we are still behind Intel/AMD in terms of gaming speed.

I know that we do not need anything more than 60fps but Doom3 requires a lot of processing power and it looks like the best PC is 25% faster than the best Mac. At least this has gone down from the 60% it was before, but it is still not very good.

Although this may change soon since UT2003 is being optimised for 64bit machines, so I will hold off on my final judgement until I see 64bit UT2003 running on a Mac vs the PC.

Edit: The above is actually wrong. I have said this more than once in this thread, but here it goes again. I was wrong when I said the benchmarks for quake 3 are wrong and PCs are faster. I was comparing the wrong benchmarks and the dual 2Ghz G5 Mac with Radeon 9800pro is in FACT faster than the Dell 3Ghz P4 with HT, 800Mhz bus, Radeon 9800pro and Rambus 1066.

IMO, there is no such thing as an objective test. Everyone has an agenda, whether it is to pump up one platform, or just to make liars out of others.

I don't know enough about these benchmark tests to tell the truth from the lies -- either from apple or from this debunker. Keep in mind that the specs he sites about the PCs are from internal tests by the manufacturers of those computers! Sure, apple has reason to tell you that a Dell is not as fast as it really may be, but do we trust Dell as the alternative source??

As far as I'm concerned, the G5 is a *huge* improvement over the G4, and regardless of what is technically the fastest personal computer in the world, the speed gap is effectively closed. We can sit here and argue about the specifics of this test or that test or how this guy or that guy manipulated the truth, but by any account, these computers (2 Ghz G5, 3 Ghz Xeon or Pentium) are in the same ballpark.

[QU
Apple is preaching to the converted, preaching to them to buy a new system. The 'Mac faithful' (those of you who aren't as religious users must forgive me for this generalization) will swallow anything Apple throws their way, and even after seeing such a claim debunked, they don't really care.

yes u r right , but to tell you the thruth, apple can throw a rock at me and i will be happy.

My dream death would be to catch a g5 tower that someone dropped from a 3rd or 5th floor in the head!!!

i will die happy with the apple g5 tattoed in my forehead.

Stop complaining. After years and years of waiting, we finally have G5!

well it was interesting. but i know nothin gabout benchmarks so i have no idea how much this means, doesnt look too good though. but as for the dual 2ghz on the gaming part, apple has always lacked in games, and those fps are simply a result on s****y drivers and lack of upkeep from nvidia and ati. if ati, nvidia and the developers gave a damn about apple we would have better results. although i do believe nvidia drivers will get better because they just hired 5 developers strictly for mac drivers.

iJon

Re: more relevant...

Originally posted by f-matic

personally, i think the most relevant comparison at the moment (and one that apple isn't providing, i assume, to avoid discouraging sales of current G4 systems) is the G5 vs. G4 comparison,

Click to expand...

you know, it just might be the other way around....yikes!

Originally posted by hvfsl
Quake 3 on a 3Ghz P4 with Radeon 9700pro gets over 400fps, while the 2Ghz dual G5 gets just over 300fps. According to http://www.tomshardware.com the AMD 2700XP is actually 10fps faster than the P4. It looks like we are still behind Intel/AMD in terms of gaming speed.

I know that we do not need anything more than 60fps but Doom3 requires a lot of processing power and it looks like the best PC is 25% faster than the best Mac. At least this has gone down from the 60% it was before, but it is still not very good.

Although this may change soon since UT2003 is being optimised for 64bit machines, so I will hold off on my final judgement until I see 64bit UT2003 running on a Mac vs the PC.

Click to expand...

Well, that demo was using the 9600 + not the 9700 that you mentioned in the PC.

Originally posted by Falleron
Well, that demo was using the 9600 + not the 9700 that you mentioned in the PC.

Click to expand...

and as has been shown on a number of occasions, the 9600 is actually slower than its predecessory the 9500, and the 9500 is slower still than its big brother the 9700, so whichever computer was using the 9600 is at a significant disadvantage to the computer using the 9700.

first-person shooter FPS performance is more closely related to the speed of the graphics subsystem than that of the CPU

Originally posted by klozowski
and as has been shown on a number of occasions, the 9600 is actually slower than its predecessory the 9500, and the 9500 is slower still than its big brother the 9700, so whichever computer was using the 9600 is at a significant disadvantage to the computer using the 9700.

first-person shooter FPS performance is more closely related to the speed of the graphics subsystem than that of the CPU

Click to expand...

Exactly, if you put the same graphics card in the Mac + the PC + I think it will be very close. Especially if the game uses the second processor.

The demo used by Apple is 02, I think, while all the others use 01. Not sure how much a difference that makes, but it could be significant.

I almost want to bring a quake 3 cd to an apple store when the machines are available, install, and play -- then do the same thing with a 3.2ghz dell at compusa

then I can post honest, no BS benchmarks -- unless tom's hardware beats me to it first, but I don't think they cover macs.

Originally posted by Falleron
Well, that demo was using the 9600 + not the 9700 that you mentioned in the PC.

Click to expand...

the quake demo actually used the 9800 card...

Originally posted by etoiles
the quake demo actually used the 9800 card...

Click to expand...

Ok then.

what a load of whiny crap

hahahahaha...

How many of the people slammning Apple here were actually at WWDC?

I'm typing this from the Moscone center.

I can't believe the fud and crap i'm seeing all over this site.

* No L3 cache??? No Crap! IBM SAID THAT THE 970 WOULDN'T USE L3 LAST YEAR!!!! The G5 doesn't need it. The bus has as much bandwidth to the main memory as the old L3 cache had.

* Apple is lying about the specs... Um, YEA. Apple went out of their way to demonstrate that the specs were generated by an objective testing firm. Jobs ADMITTED that the single 2GHz 970 was SLOWER than the 3 GHz (800MHz bus) P4 in integer math (in SPEC)... and that it was slightly faster in Floating point. He then went on to demonstrate that the dual 2GHz was faster than a dual Xeon in all regards due to the enhanced, point to point architecture.
The specs that Apple demonstrated were completely in line with what IBM posted last year at the microprocessor forum.
If anything, both the 970 AND the P4 scores were under what I expected but that's likely due to the development gcc 3.3 compiler.

* Apple cheats because they use Altivec enhanced apps to demo.....
Well, I don't know how Altivec enhanced the Mathematica demo was... but the G5 certainly did wipe the floor with the dual Xeon (which cost $1000 more with less HD space).
The other demos were extremely valid. The Photoshop run off was done with a real production design piece. It was over 2x faster than the dual Xeon.
In the audio demo... they did use different applications, but the results and stats were AMAZING... 1000 simultaneous voices.... a dozen stereo channels with 100 digital affects per channel running at under 25% cpu utilization. For the live demo... the PC couldn't even continue playback while the Mac cruised through at 50% CPU utilization.
The thing I found interesting was that the spread of performance advantage over the dual Xeon ranged from a factor of 2.2x to 2.4x on all the tested applications. Coincidence?

* Quake runs faster on a PC... Who the crap cares if an old software engine runs at 300 of 400 frames per second? How fast does your monitor refresh? And APPLE DIDN'T ANNOUNCE THE 9700PRO ON THE G5 YET... JUST THE NVIDIA 5200 AND ATI 9600 PRO SO THE PC MENTIONED HAD A BETTER VIDEO CARD.
I just noticed that the 9800Pro is a BTO in the store.. which I figured would be the case, but SJ didn't mention it during the keynote. The only place that _anyone_ would have benched the G5s at would be the performance lab downstairs... I'll see what Apple's got loaded in them

* The entry level model will have trouble keeping up with a budget level PC....
Um... perhaps you didn't look at the design. The single 1.6 has an open cpu socket. Why don't you try putting a second P4 in your budget PC 6 months down the road?
And anyway.. the 1.6 is only 20% slower than a 2GHz cpu (which keeps up with a 3GHz P4 on a 800MHz bus). If all things scaled linearly.. you'd expect it to be as fast out of the box as a 2.4GHz P4 box or better...
That's not a shabby box (especially the 'C' version with the 800MHz bus). And your cheap PC won't have gigE, PCI-X,.. it probably won't have dual channel memory or 8x AGP if it's really a low end box...

Face it... the G5 rocks. The architecture rocks. Apple matched or overtook Intel today and while the P4s roadmap is slowing, Apple and IBM promised a 50% increase in clock in the next 12months... Intel just pushed out the 3.2GHz P4, but the will only get to 3.4GHz by the end of 2003. They are shooting for 3.6, maybe 3.8 GHz in 2004 (according to ArsTechnica). By the time the next significant revision of the p4 core is ready early next year, Apple will be prepping the 980s.

Don't forget that all of these test were performed on an early version of gcc 3.3 that is just be optimised for the 970 processor. They were also conducted on a developmental OS.
The performance you see on the G5 will only get better as time goes on.

It's a good time to be a mac user... well done Apple.

Re: what a load of whiny crap

Originally posted by ffakr
How fast does your monitor refresh?

Click to expand...

Good point. I wish PC gamers had half a brain sometimes

wow, *very* good point. How many people could really tell the difference? 300 fps is so fast anyway...

He says he's a Mac user, right? And that he's writing the article from a Mac? I'd like to point out something that maybe most of you didn't notice. I could be way arong here, but I've noticed that there are certain fonts that are rendered way better on macs than on PCs. This first picture that I'm posting is from the article:

And here's a screen shot of the same page, different spec showing, but that doesn't matter, taken from a *real* PowerMac G4:

Originally posted by LimeLite
He says he's a Mac user, right? And that he's writing the article from a Mac? I'd like to point out something that maybe most of you didn't notice. I could be way arong here, but I've noticed that there are certain fonts that are rendered way better on macs than on PCs. This first picture that I'm posting is from the article:

Click to expand...

Ahaha, good point, Maybe he downloaded it on his pc and then uploaded it to his server. Then wrote the article at home on his favorite mac? LOL. I knew the G5 would get the hornets nest stirring.