PDA

View Full Version : Barefeats' G5 test!


Zenith
Aug 9, 2003, 07:45 PM
Barefeats (http://www.barefeats.com) has posted some detail-lacking results from testing a pre-production G5. Quote:

"8/9/03 -- Here's a tidbit for you weekend surfers: I tested a pre-production G5 Power Mac with dual 2GHz processors recently. I promised not to publish the results but I can say that it is faster than the 3GHz Pentium 4 in all but one of my famous "real world" tests. And that's without running Panther version of OS X or G5 optimized software.

Conclusion: If you ordered a G5, you won't be disappointed. If you are on the fence, I say jump... especially since I'm in the que ahead of you."


Looks promising enough to me! :D

bobindashadows
Aug 9, 2003, 08:32 PM
Originally posted by Zenith
Barefeats (http://www.barefeats.com) has posted some detail-lacking results from testing a pre-production G5. Quote:

"8/9/03 -- Here's a tidbit for you weekend surfers: I tested a pre-production G5 Power Mac with dual 2GHz processors recently. I promised not to publish the results but I can say that it is faster than the 3GHz Pentium 4 in all but one of my famous "real world" tests. And that's without running Panther version of OS X or G5 optimized software.

Conclusion: If you ordered a G5, you won't be disappointed. If you are on the fence, I say jump... especially since I'm in the que ahead of you."


Looks promising enough to me! :D

Well, I must say that I hoped it would beat the P4 in all of them. Oh well.

MrMacMan
Aug 9, 2003, 08:42 PM
Oh yeah This title is *in no way deceiving*

I thought they published results, nope, he is saying that it kicks, well wonderful, apple says it kicks also.


thhhannnkksss.




Not!

Fender2112
Aug 9, 2003, 09:12 PM
Originally posted by Zenith
Barefeats (http://www.barefeats.com) has posted some detail-lacking results from testing a pre-production G5. Quote:

"8/9/03 -- Here's a tidbit for you weekend surfers: I tested a pre-production G5 Power Mac with dual 2GHz processors recently. I promised not to publish the results but I can say that it is faster than the 3GHz Pentium 4 in all but one of my famous "real world" tests. And that's without running Panther version of OS X or G5 optimized software.

Conclusion: If you ordered a G5, you won't be disappointed. If you are on the fence, I say jump... especially since I'm in the que ahead of you."


Looks promising enough to me! :D

Promises are meant to be broken. :D
Gives us the details....pretty please...with sugar on top!

MrMacMan
Aug 9, 2003, 09:33 PM
Originally posted by Fender2112
Promises are meant to be broken. :D
Gives us the details....pretty please...with sugar on top!

'Its the fastest processor I've seen'
'Can you give an example, a test, a detail'
'No, I cannot, but I can say for certain that you will be pissed when I say its fast, and I can't say anything about the extensive testing i have done'


ARRG

idea_hamster
Aug 10, 2003, 01:52 AM
I'm glad to hear that a G5 is having what the testers say is "real-world" success -- technical benchmarks are great, but rarely what any of us does. I'm optimistic.



(Also, just so that none of us U.S.-based Macrumors re-make this mistake, the word for a "line of people" that's pronounced like the letter "Q" is spelled "queue", not "que" as in the article. And you U.K.-based folks have to help us Americans out with this stuff quicker before we think it's Spanish -- we'll be walking around London like Manuel from Fawlty Towers!)

macphoria
Aug 10, 2003, 02:28 AM
Sounds promising. But I would like to see the result of said test. Also in what "real world test" did G5 lose?

Zenith
Aug 10, 2003, 04:48 AM
Originally posted by MrMacman
Oh yeah This title is *in no way deceiving*

I thought they published results, nope, he is saying that it kicks, well wonderful, apple says it kicks also.

I just find this interesting because of Barefeats' critical, fair and reliable tests of computer hardware. I really don't care what Apple says, but hearing this from a site which has slaughtered Mac hardware for the last few years makes me more convinced that the G5 actually is the worlds fastest desktop computer.

Mac Kiwi
Aug 10, 2003, 05:59 AM
Excellent news :D



The Pentium is a single?.....how does it fair against a dual xeon though like Apple advertized.

Stike
Aug 10, 2003, 09:17 AM
Originally posted by macphoria
Also in what "real world test" did G5 lose?

Maybe bootup time :rolleyes:
compared to Win98...

GaelDesign
Aug 10, 2003, 11:56 AM
Originally posted by Zenith
-snip-
I can say that it is faster than the 3GHz Pentium 4 in all but one of my famous "real world" tests.
-snip-


OK, but hold on a minute. Is the Dual 2GHz machine faster than the P4, or is it faster with one processor turned off? They must mean that, otherwise, the Dual Xeon or Opteron would beat the livin' daylights out of the G5.

Having a Dual 2GHz machine beat a single processor P4 is completely and utterly terrible as far as any logical kind of speed comparison is concerned, so I really hope that's not what they meant. :confused:

Cheers, Jared

Powerbook G5
Aug 10, 2003, 12:43 PM
Perhaps the Pentium beat the G5 in the "How quickly can we get the OS to crash" test. :D

bennetsaysargh
Aug 10, 2003, 01:08 PM
i too would like to see benchmarks of the G5, but there are probably NDA up the wazoo.

shadowfax
Aug 10, 2003, 01:14 PM
Originally posted by GaelDesign
OK, but hold on a minute. Is the Dual 2GHz machine faster than the P4, or is it faster with one processor turned off? They must mean that, otherwise, the Dual Xeon or Opteron would beat the livin' daylights out of the G5.

Having a Dual 2GHz machine beat a single processor P4 is completely and utterly terrible as far as any logical kind of speed comparison is concerned, so I really hope that's not what they meant. :confused: i doubt it's with the other proc turned off. a question that arises in my mind, though, is whether Xeons support hyperthreading like P4s do, because a hyperthreading P4 is about as good as the non-existant dual P4 system would be.

Vlade
Aug 10, 2003, 01:18 PM
Good, Im disappointed in the G5 VS G4 tests though :(

Capt Underpants
Aug 10, 2003, 02:39 PM
Originally posted by bobindashadows
Well, I must say that I hoped it would beat the P4 in all of them. Oh well.

I guess we just can't have everything, now can we;)...... I sure am happy with the results. It just confirmed what we already knew. I am sure that intel has something up their sleeves, too. Here's hopin IBM will keep innovatin'.

Abstract
Aug 10, 2003, 03:01 PM
Originally posted by GaelDesign
OK, but hold on a minute. Is the Dual 2GHz machine faster than the P4, or is it faster with one processor turned off? They must mean that, otherwise, the Dual Xeon or Opteron would beat the livin' daylights out of the G5.

Having a Dual 2GHz machine beat a single processor P4 is completely and utterly terrible as far as any logical kind of speed comparison is concerned...

It doesn't matter. Who cares if a single G5 may (or may not) be beaten by a single 3.2GHz P4 processor? Apple doesn't offer a single 2.0GHz proc setup. Price equivalency should be the comparative factor. The fact is that a dual 2.0GHz G5 will probably beat a 3.2GHz P4, and they are also similarly priced. You get more from the Mac for the price you pay.

Fender2112: Promises are meant to be broken. :D
Gives us the details....pretty please...with sugar on top!

I like numbers. Numbers are our friend.

vniow
Aug 10, 2003, 03:11 PM
Originally posted by Shadowfax
i doubt it's with the other proc turned off. a question that arises in my mind, though, is whether Xeons support hyperthreading like P4s do, because a hyperthreading P4 is about as good as the non-existant dual P4 system would be.

http://intel.com/eBusiness/products/server/processor/xeon/wp020901_sum.htm?iid=ipp_srvr_proc+high2_120302&

bosskxx1
Aug 10, 2003, 09:31 PM
I am going out on a limb and guessing that the Bryce Test is where the P4 beat the G5. Bryce 5 doesn't support dual processors and had very little optimization for the Mac.

mim
Aug 10, 2003, 10:44 PM
Originally posted by GaelDesign
OK, but hold on a minute. Is the Dual 2GHz machine faster than the P4, or is it faster with one processor turned off? They must mean that, otherwise, the Dual Xeon or Opteron would beat the livin' daylights out of the G5.

Having a Dual 2GHz machine beat a single processor P4 is completely and utterly terrible as far as any logical kind of speed comparison is concerned, so I really hope that's not what they meant. :confused:

Cheers, Jared

Tough one...

Lots of software is not properly optimised for more than 1 proc to start with. And then wintel dual proc machines aren't terribly more efficient than single proc setups (mostly), while the macs/ppc really gets a big kick from 2 procs. And then the G5 dual setup is really really really good - full memory bandwidth to each processor, system bus so wide you could drive a..umm..bus down it.

But it's quite sensible to compare the dual G5 to even a single P4, when you consider that the Dual PMac is only going to be slightly more expensive than a similar spec single proc Dell. The Dual Xeon machines start to get way more expensive than the Mac.

And as for Opteron, it just doesn't perform in desktop applications - with 64bit server apps you get some performance benifits. The Mac will get those same benifits once 64bit app development begins under OSX (or yellowdog linux, etc)

MrMacMan
Aug 10, 2003, 10:44 PM
Originally posted by GaelDesign
OK, but hold on a minute. Is the Dual 2GHz machine faster than the P4, or is it faster with one processor turned off? They must mean that, otherwise, the Dual Xeon or Opteron would beat the livin' daylights out of the G5.

Having a Dual 2GHz machine beat a single processor P4 is completely and utterly terrible as far as any logical kind of speed comparison is concerned, so I really hope that's not what they meant. :confused:

Cheers, Jared

Intel Claims The P4 (hyperthreading version) is really acting like 2 processors.

Therefore because of their claim it is fair to test with 2 processors.


Look, if 1 Processor G5 wins against the P4 HT, I will *really* be impressed.

Cubeboy
Aug 10, 2003, 10:58 PM
Good to hear this from Barefeats, who are usually quite fair with there reviews, seeing that a single 2 GHz G5 will most likely be highly competitive with the P4, I wouldn't be very surprised if dual G5's does beats the P4 in all those tests. I'm especially interested in seeing how the G5 stacks up in 3d games which should prove very interesting.

Just to clear up some things up, a high-end P4 system will still be alot cheaper than the 2 GHz Dual G5 Powermac. For $3000, a midrange to highend Xeon workstation is very viable so the comparison is quite valid (especially considering both will be used for similar tasks).

In some applications like Aftereffects, hyperthreading does seem to offer the same performance increase as having dual processors. However, these are the exceptions, not the general case. HT on the Pentium 4/Xeon is typically seen as icing on the cake, especially considering that the Pentium 4 is destined towards desktop use.

Cubeboy
Aug 10, 2003, 11:13 PM
Originally posted by mim
Tough one...

Lots of software is not properly optimised for more than 1 proc to start with. And then wintel dual proc machines aren't terribly more efficient than single proc setups (mostly), while the macs/ppc really gets a big kick from 2 procs. And then the G5 dual setup is really really really good - full memory bandwidth to each processor, system bus so wide you could drive a..umm..bus down it.

But it's quite sensible to compare the dual G5 to even a single P4, when you consider that the Dual PMac is only going to be slightly more expensive than a similar spec single proc Dell. The Dual Xeon machines start to get way more expensive than the Mac.

And as for Opteron, it just doesn't perform in desktop applications - with 64bit server apps you get some performance benifits. The Mac will get those same benifits once 64bit app development begins under OSX (or yellowdog linux, etc)

We're talking about workstations here so nearly all your applications are going to be well threaded even for PCs. A fine example of this is Lightwave in which a dual 2.4 GHz Xeon beats the crap out of a 3 GHz P4.

High end P4 pcs (3 Ghz/3.2 GHz P4) can be very well configured from retailers (Dell, HP, Sony, etc) for $1600-1800 and built for significantly less so theres really no point in arguing about price. As for Xeons, I can configure a dual 2.8 GHz Xeon with 120 GB HD, 1 gig ram, and Quadro 500 graphics card from Dell for $2700. Personally, I think the best comparison would be to bench both the P4 and the Dual Xeons against single and dual G5 powermacs.

There have already been multiple sites that have benchmarked Opterons using typical desktop programs like games and office software.

Vlade
Aug 11, 2003, 10:18 AM
How much does HT really help, does it actually make the CPU alot faster, or does its multitasking make it feel like OS X, because everything gets CPU time ?

Cubeboy
Aug 11, 2003, 11:53 AM
Originally posted by Vlade
How much does HT really help, does it actually make the CPU alot faster, or does its multitasking make it feel like OS X, because everything gets CPU time ?

Depends on the applications. Since HT is essentially two (virtual) cores running on a single processor that share many common cpu components (ALU, FPU, Cache, Code-Decode units), applications that benefit from having dual processors will also benefit from HT. Exactly how much of a performance boost a cpu gets with HT turned on depends on the application. With multi-threaded applications, the performance gain will be anywhere between 5%-30%. Not nearly as good as dual processors or dual cores (with a few exceptions) but still quite significant. On the P4s and Xeons, HT is viewed as icing on the cake although it could become a major aspect of the cpu as the technology matures (more efficient, higher levels of HT). Oh, and yes it will help in multitasking, in the same way it boosts performance on multi-threaded applcations.

AngryAngel
Aug 11, 2003, 12:10 PM
How does performance of dual cores compare with dual processors?

Is there a cost advantage to dual porcessor systems (over dual cores)?

I am wondering why IBM/Apple didn't develop a version of the dual core POWER4 (with added VMX) instead of (or in addition) to the PPC 970, if they are going to sell many dual processor configured computers.

Any word on IBM's adoption of Hyperthreading (or a similar technology)?

Lanbrown
Aug 12, 2003, 08:25 AM
Originally posted by Stike
Maybe bootup time :rolleyes:
compared to Win98...

No, DOS.

Lanbrown
Aug 12, 2003, 08:33 AM
Apple published a SPECfp score of 15.7 for the Dual 2GHz G5. A Dell Precision WorkStation 350 (3.06 GHz P4) gets a 12.7. A dual processor Dell Precision Workstation 650 (3.06 GHz Xeon) gets an 18. And the single of that same system gets a 12.2. A dual processor IBM IntelliStation POWER 275 Workstation (1450 MHz) gets 19.6.

Lanbrown
Aug 12, 2003, 08:37 AM
Originally posted by AngryAngel
How does performance of dual cores compare with dual processors?

Is there a cost advantage to dual porcessor systems (over dual cores)?

I am wondering why IBM/Apple didn't develop a version of the dual core POWER4 (with added VMX) instead of (or in addition) to the PPC 970, if they are going to sell many dual processor configured computers.

Any word on IBM's adoption of Hyperthreading (or a similar technology)?

Dual cores are typically cheaper for software for servers. Some software you pay by the CPU, so if you have two cores, it is still one CPU though. Some software companies make the distinction though, others do not. It also only takes one slot in a server, compared to two.

It is cheaper to make a dual processor system then a dual core processor.