NASA Benchmarks the New G5 Powermac

#1 aStRaLgOd on 04 Jul 2003 - 21:51

hum those benchmarks are also tricked!

#2 WeeJames on 04 Jul 2003 - 21:57

Yes... its always a trick.. its always a conspiracy.. watch out Apple are going to suck out your soul with their evil benchmarks!

(1 reply) #3 Wickedkitten on 04 Jul 2003 - 22:00

omg if nasa will lie about computer benchmarks, what the hell else are they lying about? space aliens in the white house?????

#3.1 macrosslover on 04 Jul 2003 - 22:01

how'd you know...it was suppose to be a secret lol

#4 macrosslover on 04 Jul 2003 - 22:00

oh give me a break lol. i've glanced over it...and it seems fair. they give indictaion they will perform another test, when full g5 optimizations are out and when they have the latest p4 chip etc to get a final idea of performance....i wonder why they didn't bench against AMD......but oh well.

(1 reply) #5 TheDeputy on 04 Jul 2003 - 22:01

Big Deal, who cares.. blah blah blah

#5.1 nacs on 04 Jul 2003 - 22:38

I care. Stop trolling if you're not interested.

#6 kairon on 04 Jul 2003 - 22:09

NASA is fair. Stop thinking every benchmark that shows Apple performing equal or almost-equal to today's chips is rigged. Good read!

(1 reply) #7 Eth3l on 04 Jul 2003 - 22:51

The conclusion of the benchmark:

QUOTE

The primary purpose of this test was to determine how G5 scalar floating point performance compares to G4 performance in CFD applications. As a secondary part of this test, G4 and G5 benchmark results were compared to similar results obtained on Pentium 4 systems. Overall, the scalar floating point performance of G5 systems is much improved over G4 systems due to better per clock cycle efficiency combined with higher clock speeds. Based on preliminary testing with an existing version of Jet3D (not recompiled or optimized for the G5), it appears that the G5 has about 22% better scalar floating point performance per clock cycle than the G4 systems tested and 32% better floating point performance per clock cycle than the P4 systems tested. Based on raw scalar floating point performance in Jet3D, a 2GHz G5 system can match a 2.66GHz P4 system, and this is a dramatic improvement from earlier tests where G4 systems lagged behind higher clock speed P4 systems. Based on an extrapolation of current P4 results, the 2GHz G5 would lag newly announced 3.2GHz P4 systems in Jet3D scalar floating point performance by about 20%, but this kind of comparison is best deferred until G5-aware compiler tools become available (since a 20% performance gain is well within the potential of compiler optimization).

Vector performance of the G5 remains excellent, and is inline with current G4 systems on a per clock cycle basis. As a result, raw vector performance of the G5 will be boosted simply by its higher clock speeds relative to current G4 systems.

Finally, it is important to note that the current test does not factor machine cost or intended use into the picture, and that can have a large impact, especially in clustering applications.

I would make the assumption that the dual proc would destory the P4, sadly it didn't. Although it did do very very well!

Yeah, good question Macrosslover . . . where was AMD?!

#7.1 timdorr on 05 Jul 2003 - 04:19

Well, this *was* a Single CPU test, so it should beat it handily, because there was only one proc being used at the time

#8 divertom15 on 04 Jul 2003 - 23:17

opteron would wipe the floor with that g5 in single or dual processors and opteron doesnt cost very much

GO DOWN ON YOUR PRICES JOBS!!!!

#9 jwkuipers on 04 Jul 2003 - 23:23

The problem with this test is that the best Apple offers was not tested against the best that Intel can offer. Both of the Intel processors had been out for over a year. It should have been tested against a Pentium 4 3.0CGhz, or enable the second processor on the Mac and test it against a dual Xeon 3.06Ghz.

(12 replies) #10 Animaniac on 04 Jul 2003 - 23:41

I <3 AMD fanboys.

AMD is not, unfortunately, not the best thing since sliced bread. The cult following behind AMD, and Apple for that matter, disgusts me.

These benchmarks are what they are. For this purpose, it seems the P4 is blown out of the water by the G5 (~5000 MFLOPS v. ~300 MFLOPS on optimized code). Why is it so hard for people to just accept that the G5 is a great chip? Intel was never good at making CPUs.

Last edited by 5584 on 04 Jul 2003 - 23:48

#10.1 NXTwoThou on 05 Jul 2003 - 00:11

Huh?

<i>...and within 1 MFLOP of the 2.66GHz P4...

Combined with higher clock speeds, this results in significantly better floating point performance than G4 systems and performance on par with a 2.66GHz P4</i>

It gave me the impression my notebook I bought in october last year is roughly equal to a single brand spankin new G5.

And the new machine I built at work back in march, which makes my notebook feel pathetic, was a good buy since it was over $600 less than what the new apples will be sold for.

#10.2 Animaniac on 05 Jul 2003 - 00:21

"As before, it is worth noting that the G5 system benchmarked at 5177 MFLOPS and 1.29 MFLOPS/MHz for vector Jet3D performance when two processors were used."

It's irrelevent what a single 2 GHz G5 can do, since you can't buy a single processor configuratioin. The Intel compile couldn't even handle vector benchmarks.

#10.3 ike on 05 Jul 2003 - 00:43

QUOTE (#10.0)

I <3 AMD fanboys.

AMD is not, unfortunately, not the best thing since sliced bread. The cult following behind AMD, and Apple for that matter, disgusts me.

agreed. there are different processors in this world, they excel at different things, big deal.

i think a lot of people like AMD just to be "different" than others. they may not be able to tell the difference between their computer and their friend's computer (except oh the wonderful benchmarks) but they just like being different.

apple on the other hand, actually IS significantly different than a PC platform. but that's not to say the apple crowd doesn't have its share of idiots...

#10.4 NXTwoThou on 05 Jul 2003 - 00:47

"Again, note that the vector benchmark does not include P4 systems because the AltiVec instruction set is only available on G4 and G5 systems"

The last test was to show just how much faster the G5 is over the G4, it didn't involve a P4. I assume you got the ~300MFLOPs from the previous test.. The single test showed P4 at 255MFLOPS and G5 at 254MFLOPS with a dual G5 at 498MFLOPS(with no dual P4/G4 tests to compare against).

#10.5 macrosslover on 05 Jul 2003 - 00:58

no, it has nothing to be do with people wanting to be different because they hate Intel or anything like that, too many benchmarks have shown that are SOME applications and many tasks, a lower clocked amd chip can match and sometimes, although increasingly rare, beat a higher clocked p4. generally AMD charges less for their chips at a given performance than Intel does. AMD makes good chips, that's it. and the opteron is a damn good chip, that people seem scared to test against. i have nothing against Intel personally, they make good chips, if their chip was the cheapeast and offered the highest performance, i'd have one in my system. my next one might be an Intel, so it has nothing to do with fanboys or anything, just if you are going to compare chips, compare them all, because who knows the Opteron/Athlon could possibly eat the p4 up in these tests NASA did, which would imply it beats the g5 in several tests.

#10.6 malebolgia on 05 Jul 2003 - 01:12

QUOTE (#10.5)

no, it has nothing to be do with people wanting to be different because they hate Intel or anything like that, too many benchmarks have shown that are SOME applications and many tasks, a lower clocked amd chip can match and sometimes, although increasingly rare, beat a higher clocked p4. generally AMD charges less for their chips at a given performance than Intel does. AMD makes good chips, that's it. and the opteron is a damn good chip, that people seem scared to test against. i have nothing against Intel personally, they make good chips, if their chip was the cheapeast and offered the highest performance, i'd have one in my system. my next one might be an Intel, so it has nothing to do with fanboys or anything, just if you are going to compare chips, compare them all, because who knows the Opteron/Athlon could possibly eat the p4 up in these tests NASA did, which would imply it beats the g5 in several tests.

I second that macrosslover.

#10.7 ripgut on 05 Jul 2003 - 02:51

QUOTE (#10.0)

Intel was never good at making CPUs.

and apple has made as much money as intel with there so much better chips? cmon bro apple is barely at the ghz range with there chips intel has been at that range lmao

#10.8 hjimmy on 05 Jul 2003 - 03:11

Err, intel is one of the biggest CPU maker...
what are ya smoking?

#10.9 Animaniac on 05 Jul 2003 - 05:03

Study their architecture and processor designs, then we'll talk. Anyone edcuated enough knows they're at the bottom of the list. IBM processors are designed the most elegantly and efficiently.

#10.10 Animaniac on 05 Jul 2003 - 05:04

No, it shows how a vector optimized version of Jet3D perfomed. They could not get it be vector optimized for x86.

#10.11 threedaysdwn on 05 Jul 2003 - 17:10

You people are absurd!

Intel currently has the lead in desktop chip design. No question there. They already have their 90nm strained silicon process ready to debut, they're just waiting for a reason to. Right now their 3.2Ghz P4 is the fastest desktop CPU around.

Where an IA-64 environment is available, Itanium 2 dominates. It gets beaten into the ground in 32-bit mode though. But the architecture itself is very, very impressive. Even moreso than IBM's much ballyhoo'd Power4.

If you think there's ANYTHING elegant about Power4 (cramming several chips into one), you've got a lot to learn about CPU design.

Intel and AMD both currently make better chips than IBM. Hell, IBM uses Opterons now.

The Jet3D benchmarks have *NOTHING* to do with an inability "to get it be vector optimized for x86," as Animaniac so eloquently put it.

So what the hell does that mean anyway? Do some freakin' research before you start posting in defensive of a company that cares nothing about you, using arguments you made up 'cause you thought they sounded good.

Intel's Prescott cores are by far the most technologically advanced chips designed to date. Opteron is an impressive architecture that, with higher clock frequencies, could translate to a desktop chip worthy of competing with the Pentium 5 (prescott).

The G5 is a single, crippled Power4 core with Altivec strapped on. There's nothing elegant about it. Motorola killed their semiconductor division and Apple desperately needed something new to market. So the PPC 970/G5 was thrown together.

And finally, there is no one reason why people by AMD processors. There are several:

1) Better price/performance ratio than comparable Intel chips, especially at the high end.

2) Nvidia nForce. While Intel's own P4 chipsets continue to improve, they still don't offer the same allure to enthusiasts that nForce 2 and 3 do.

3) AMD's 64-bit chips are unmatched in their ability to run 32-bit software AND easily port software to 64-bit mode.

I'm sure there are some people that buy them in spite of Intel's shoddy marketing practices (especially when it comes to dealing with AMD-friendly OEMs). But most people simply don't care.

#10.12 mealbundy on 07 Jul 2003 - 02:51

Please please please, STOP COMPARING 32bit chips to 64bit CHIPS!!!!!!!!!!!!!!!

I cant stand the stupidity of some reviewers. DONT EXPECT A 32BIT CPU TO OUTPERFORM A 64BIT CPU!!! Thats what happened with the 486 vs. Pentium era.

(1 reply) #11 Mr. Black on 05 Jul 2003 - 00:04

It's nice to see Apple has made a machine that actually gives prestige to it's name - PowerMac.

#11.1 threedaysdwn on 05 Jul 2003 - 17:13

If you consider performance that is less than or equal to that of last year's PCs to be the banner that follows the PowerMac name, then you're right.

Apple has certainly held true to their name.

#12 Ciderx on 05 Jul 2003 - 00:14

That's better. These are the scores that were expected. The PPC970 at 1.8MHz was always meant to hit the levels of a mid 2-GHz P4, and these results prove it. It proves that the dual 2 GHz machine is probably about the spec of a 2.8 GHz P4, not 95% faster than the 3.06GHz P4...

#13 kuroneko on 05 Jul 2003 - 00:54

Good for Apple that IBM came out with a better processor for them to tout about. Still as a loyal PC fan I'm going to wait till the G5 ships along with Intel's and AMD's latest processor till I draw a conclusion.

#14 NIMASoft on 05 Jul 2003 - 02:41

thats pretty tight!

#15 psw on 05 Jul 2003 - 03:17

it seems as if Apple is getting more praise here than IBM w/ this new G5 and these test.

shouldnt it be IBM vs. Intel vs. AMD .... ?

#16 The_Rain_King on 05 Jul 2003 - 03:18

this is absolutely rediculous in practicality. Raw bewnchmarks don't mean as much as actual performance in applicable ways. The comparison between the PCs and Mac G5 should be head to head, same software on both platforms - OS X and WinXP - NOT RH Linux. Disclaimer: I am not here to debate the practicallity and growth or lack thereof the linux fanbase or anything like that. If you like linux or any other alternate os fine by me, please encourage programmers to write more programs and help it's growth even more. I hope it serves your needs. Anyways, the point of all this is to say that Apple's sales pitch for these machines is highend graphics/multimedia/video production and post-production/rendering. I also know numerous meteorolgist that run high end weather models from unix apps and the likes. The same can be said for the PC side. So how about somthing practical - a benchmark to comparable systems running software thaty might actually matter. like ANYTHING by Adobe - like Premier 6.5, After Effects 5.5, Photoshop 7 or 8 beta (talk about great advertising). Simply put, Linux just can't meet the typical graphic design / video production needs that the Windows and Mac platform possess at this time. And if it's someone who really jjust want's to get on the internet and send photos and email and stuff like that, none of this shold matter anyways.

I also feel like adding that most people interested in jet noise prediction tools will more than likely be large corprate companies who perhaps build engines, planes, and the lkike. Perhaps others, but i imagine that would be an accetable use for such things. Uh, these companines have moeny, lots of it - enough to potentially purchase any pimped out rig they want, which too would greatly differ the result of the benchmarks for this app.

I personally can't stand benchmarks that are nothing but pure pissing contests. It's even worse when the results are distorted/crippled just to show for a brief moment in time whose is "bigger."

This now concludes my commentary for the night. Happy 4th.

Last edited by 3482 on 05 Jul 2003 - 03:27

#17 EnIgMa-PenGuIn on 05 Jul 2003 - 07:50

more wasting of taxpayer funds....yay

(1 reply) #18 David3k on 05 Jul 2003 - 13:37

look at this

P4 = least calulations per mhz = most mhz
G5 = balanced calculations per mhz = balanced
Athlon XP = most calculations per mhz = less mhz

AMD64 = BOOM, BOOM, BOOM....ROAR!!!!!!

lol.

no flame plz. If I spot flaming, I report it to the mods.

#18.1 threedaysdwn on 05 Jul 2003 - 17:16

How can you call the G5's IPC "balanced?"

IPC means just as little as clock speed.

It's the resulting processing power from both that matters. Whichever is "balanced" most is the CPU that is fastest. In this case, it's the P4. Since there are no AMD benchmarks, we can't make a comparison there. A shame, really. They've certainly earned inclusion in such a test.

(3 replies) #19 David3k on 05 Jul 2003 - 13:42

I also note that the G4 cpu's are 128-bit, and the G5 is 64-bit?

someone explain the step down?

#19.1 aaron901 on 05 Jul 2003 - 17:07

G4 is 32-bit man.

#19.2 threedaysdwn on 05 Jul 2003 - 17:19

It depends on what you mean. Technically a Pentium III has 128-bit components... namely, its SIMD registers.

I don't know if the Altivec unit in the G4 was 128-bit or not, but this could be what you're thinking of.

64-bit in the case of the G5 is simply a way of saying it can address very large amounts of memory. It can also perform calculations on reeeeaaaaallly big numbers in a more efficient way. Basically, it means absolutely nothing to desktop users. That's why Intel hasn't bothered with a 64-bit desktop chip yet. There's no point. Except for marketing, of course.

#19.3 David3k on 07 Jul 2003 - 06:28

you mean they lied about their CPU being 128-bit?!?

NOTE: I also notice that all references about the G4 being 128-bit are gone from their website.

(2 replies) #20 Da22in on 05 Jul 2003 - 14:52

Perhaps NASA should spend more time making sure the space shuttle doesn't blow up again instead of playing in the computer lab! The G5 is once again tested against older, inferior products. Where was the Opteron? They put it back in the cabinet once they realized that it will smoke the G5 up one side and down the other.

What you have is convergence between a CPU and the OS that controls it, this is the most ideal condition and it's one of the biggest reasons Apple's performance numbers are always up. The G5 itself is NOTHING special, neither is Apple for that matter.

#20.1 threedaysdwn on 05 Jul 2003 - 17:20

There's only one test that includes (old) P4s. And the P4 is tied with the 2Ghz G5.

I don't think anyone "cheated."

I just think the G5 got spanked.

#20.2 roadwarrior on 07 Jul 2003 - 02:06

QUOTE (#20.0)

Perhaps NASA should spend more time making sure the space shuttle doesn't blow up again instead of playing in the computer lab!

Did it occur to you that maybe one of the reasons for this test was to select which processors would be most efficient for running simulations used to improve the current shuttle, or to design the next generation one?

#21 David3k on 05 Jul 2003 - 16:35

[braces himself for the galaxy to tear itself apart]

#22 David3k on 05 Jul 2003 - 17:04

can somebody PLEASE explain why they use 64-bit now?!? they had an advantage before. why did they drop it?

#23 thetman on 05 Jul 2003 - 19:49

haha my dad works at nasa langley. maybe i can get some pictures when he goes to work monday ......

(2 replies) #24 Colonel Sanders on 05 Jul 2003 - 20:50

for those of you mentioning the fact that the g5 is equal to the laptop you bought last year, you have to tkae the test in context. It's only one of the processors, and it's un optimized. These systems are designed for dual processor configurations.

#24.1 macrosslover on 05 Jul 2003 - 22:48

so wait a minute, are you telling my a system is only best when it's under it's optimized condition and operating under the conditions and stuff it was intended to use for?? so obviously if you have a g5, you would expect g5 optimized performance with the best apple compilers etc.....is that what you're saying???

#24.2 Colonel Sanders on 06 Jul 2003 - 02:15

what I was implying was that a G5 runs best when in a dual processor configuration. They are designed to dual, and that's where they perform the best. I'm not talking about optimized compliers, but rather real world systems.

(1 reply) #25 thetman on 05 Jul 2003 - 23:11

oh and the guy who commented on the space shuttle, Nasa Langley does not do anything in terms of the shuttle

#25.1 Da22in on 06 Jul 2003 - 18:22

Who cares!? It's taxpayer money being spent on BS (what a surprise) It matters not whether it's in Cape Canaveral, FL or Langley, VA. Take the couple of million they spent in benchmarking CPU's and send it to the shuttle program.

#26 Gary_Player on 06 Jul 2003 - 01:12

...That was the stupidest set of benchmarks I've ever seen...honestly, you'd think freakin NASA could come up with something a little better than a group of 7th graders could have done.

Admitedly the report was better but asside from that

#27 ripgut on 07 Jul 2003 - 01:47

lol if only apple could run sisoft sandra(can they?) then well see who comes out on top till then thats jut an app optimized for g5 processors so stfu apple fans intel still owns you

#28 David3k on 07 Jul 2003 - 06:40

4MD PWNS!!!

Neowin.net