More Intel iMac Benchmarks (Native and Rosetta)

MacRumors

macrumors bot
Original poster
Apr 12, 2001
7,428
8,491


MacCentral reviews and benchmarks the new Intel iMacs released at Macworld Expo.

MacCentral's tested both Universal applications as well as PowerPC applications on the new 2.0GHz Intel iMac and compared them to the 2.1GHz iMac G5.

Of the Universal Applications tested, they focused on iMovie, iPhoto, iTunes, iDVD, iSquint, BBEdit and Zip Archiving. In these native applications, the Intel iMac was up to 1.82x as fast. Average increase however was closer to 1.2-1.3x, with one test coming in slightly slower (.91x) than the G5 iMac.

Meanwhile, PowerPC applications were tested using Rosetta emulation. They tested iTunes (PowerPC), Photoshop CS2 and Word. The tests showed the Intel iMac to be running .34-.48x as fast as the iMac G5. The performance penalty is expected due to the Rosetta emulation layer allowing PowerPC instructions to be run on the Intel processor.
 

Lacero

macrumors 604
Jan 20, 2005
6,639
2
And quieter, too. Remember folks, lack of fan noise is more important than pure speed. I wish these tests looked at the whole computing experience rather than how fast an app launches.

LifeIsCheap said:
Rosetta emulation scores look terrible!
Get out! Really?
Here's to the Crazy Ones
 

Dr. Dastardly

macrumors 65816
Jun 26, 2004
1,317
0
I live in a giant bucket!
The rosetta scores really are not all that bad considering that its all being emulated for an entirely different chip. I don't know if it shows the real speed of the intel chip or the skill of the Apple programers for creating a fairly decent emulation software.

All in all I'm not to upset, on the contrary I'm pretty impressed.
 

arn

macrumors god
Staff member
Apr 9, 2001
14,090
1,424
yeah, usually for emulation, I always picture an order of magnitude speed decrease. 30-50% isn't that bad.

They don't mention if they tried running applications again. The way the Transitive technologies works, I think it's supposed to get faster with each iteration. So the longer you work with it, it will perform better.

arn
 

sith33

macrumors newbie
Jul 16, 2002
18
0
While I certainly understand the tests they chose, but many of them (iPhoto exports to files and to web and whatnot, zipping files) are largely limited by other factors - the disk and whatnot. So, it's not terribly surprising to see limited speedup. I think it's all good - native apps won't lose speed in the short term, and we're much better off for the long term.
 

Thanatoast

macrumors 65816
Dec 3, 2002
1,005
134
Denver
And what kind of speed hit does vpc take on a G5? It ain't pretty, that's fer damn sure. I think that these scores are fine. It's just a temporary measure designed to shore up must-needed software in the short-term, unlike the alleged long-term solution that vpc was supposed to be.
 

skunkworks

macrumors member
Jan 11, 2006
49
0
The speed will improve over time and if you looking to buy an imac obviously this is the one, if not hold out a little while longer as the next revision of imac will be much better but think that the timing will be late summer for that.
 

EricNau

Moderator emeritus
Apr 27, 2005
10,589
31
San Francisco, CA
One test came in lower? I guess my iMac is still good for something...:p

Rosetta is a little slower than I would have expected, but I guess it is acceptable.

iTunes runs natively on Intel now, right? So why did they test that on Rosetta? I know it doesn't really matter, but it just seems weird to me.
 

noverflow

macrumors regular
Jul 4, 2002
188
0
"In these native applications, the Intel iMac was up to 1.82x faster"
Should say

"In these native applications, the Intel iMac was up to 1.82x as fast"
or
In these native applications, the Intel iMac was up to .82x faster

Other then that... Looks like the g5 can hold its own considering that you should expect at least this much of a gain just from the 2nd core alone
 

FoxyKaye

macrumors 68000
Nice to see that once everything gets Universal Binaries, that the apps will really (or at least, should) fly. I'm guessing that the pressure is really going to be on developers this year to re-compile. Rosetta is only (hopefully) temporary for a lot of major apps, and it might turn in to a PPC "classic" environment for apps that are never re-compiled in the future - presumably at which time the overall speed of the chips on the Intel roadmap will compensate for the emulation performance hit.

Though the early benchmarks on Rosetta would explain my experience with PhotoShop at MWSF - it felt like I was trying to run the CS2 suite on a Pismo. Thank goodness Adobe's promised universal updates by March.

Thanatoast said:
And what kind of speed hit does vpc take on a G5? It ain't pretty, that's fer damn sure. I think that these scores are fine. It's just a temporary measure designed to shore up must-needed software in the short-term, unlike the alleged long-term solution that vpc was supposed to be.
Anybody else get the feeling that M$ either has abandoned, or will be abandoning VPC in the near future? I have this eerie feeling that VPC 7.0.2 is the last version we'll see on PPC-based Macs, and perhaps may go the way of Internet Explorer and Windows Media Player...

It kinda makes sense, and it kinda doesn't - since folks are already trying to install Windows on the new Intel Macs, it might be that it would be a heck of a lot easier for M$ to put code directly into Vista that will install it on Apple Intel hardware.
 

kainjow

Moderator emeritus
Jun 15, 2000
7,745
3
I'd like to see some Xcode benchmarks. Does it compile much faster? I guess this would have to be compared with a dual processor Power Mac to be more realistic (the 2nd processor on the iMac will give it a huge advantage over the single processor iMac G5).

FoxyKaye said:
Anybody else get the feeling that M$ either has abandoned, or will be abandoning VPC in the near future? I have this eerie feeling that VPC 7.0.2 is the last version we'll see on PPC-based Macs, and perhaps may go the way of Internet Explorer and Windows Media Player...

It kinda makes sense, and it kinda doesn't - since folks are already trying to install Windows on the new Intel Macs, it might be that it would be a heck of a lot easier for M$ to put code directly into Vista that will install it on Apple Intel hardware.
No, and here's why (from MacMinute.com):
Microsoft has issued a statement on the development status of Virtual PC for Intel-based Macs."The Mac BU recognizes the need for the product and believes it is the best virtualization solution for PowerPC users, so it is committed to providing Virtual PC to new and existing PowerPC customers. However, Microsoft is still discussing with Apple the feasibility of bringing Virtual PC for Mac to Intel-based Macs in the future and has not made any announcements about if/how the product might work on the new machines."
 

Yvan256

macrumors 603
Jul 5, 2004
5,032
886
Canada
Rosetta, Intel Core Duo, G5... How are the new iMacs compared to my little Mac mini and its G4? :confused:

And do you think we'll see iBooks and Mac mini updated at the same time? Will they offer a "Mac mini Pro" with a Core Duo or will the iBook and Mac mini be stuck with single-core processors?
 

Rend It

macrumors 6502
Oct 27, 2003
265
5
United States
noverflow said:
...Looks like the g5 can hold its own considering that you should expect at least this much of a gain just from the 2nd core alone
Exactly! It would be nice to compare it with a fictitious iMac G5 dual-core. The 30-50% Rosetta performance seems like the most positive thing about this story. However, it's hard to determine if the speed differential is small because Rosetta is super-awesome, or if it's due to the performance increase of the hardware (or both?). I hope for Apple's sake the underwhelming performance of the native apps is due to un-optimized code.... I figure a single core Yonah is about on par with the single core G5, so while I don't expect performance of the apps to be exactly twice as fast with Intel Duo, I would have expected many more of those tests to show ~1.5x speed improvements. Certainly, there shoudn't be any tests that ran slower!?

Intel's chips may have been bored in PCs, but the G5 certainly wasn't bored in the Mac. :rolleyes:
 

Abstract

macrumors Penryn
Dec 27, 2002
24,378
110
Location Location Location
arn said:
yeah, usually for emulation, I always picture an order of magnitude speed decrease. 30-50% isn't that bad.
But it's not "34% to 48%" slower. It's 52% to 66% slower used under Rosetta!

May as well be using Virtual PC :eek:








Ok, that's a bit of an exaggeration
 

shawnce

macrumors 65816
Jun 1, 2004
1,442
0
Yvan256 said:
Rosetta, Intel Core Duo, G5... How are the new iMacs compared to my little Mac mini and its G4? :confused:
It is likely fully dusted by the new iMac Core Duo, possibly even when using Rosette (for some tasks at least).
 

ChrisA

macrumors G4
Jan 5, 2006
11,594
376
Redondo Beach, California
noverflow said:
Other then that... Looks like the g5 can hold its own considering that you should expect at least this much of a gain just from the 2nd core alone
No, I think the tests they used did not make much use of the second core. For example compressing a file is not multithreaed. In theat test the second core would only be running OS system level processes and be mostly idle,

A better test would be to run four compression tests at once and then compare how long it takes to compress all four folders on each system.

Notice the Intel web site talks about "multi core" and gives "dual core" as just one possable examle or "multi-core" In time there will be eight core Intel chips. In fact Sun is shipping an eight-core Sparc chips right now, Intel will follow. What does this mean? It means we'd better start today thinking about paralleliusm and multitasking as "normal". We should expect or demand that when we ask Photoshop to resize 50 image files that it will fire off 50 threads all at once rather then converting one image after the other. The Sun chip can run 32 threads at once using 8 cores each with 4-way "hyperthreading".

Why wrte about this? Educated consummers will create a demend. Demand will get the developers motivated.
 

dr_lha

macrumors 68000
Oct 8, 2003
1,587
0
LifeIsCheap said:
Rosetta emulation scores look terrible!
Are you serious? Emulating another CPU at 30-50% speed is amazing. Usually such technologies have a 10-20x Speed hit. Emulation generally has only be useful when the machine you're running on is orders or magnitude faster than what you're emulating.

I don't know what people were expecting with Rosetta, but I'm frankly amazed at these scores.
 

shawnce

macrumors 65816
Jun 1, 2004
1,442
0
kainjow said:
I'd like to see some Xcode benchmarks.
Xcode on a iMac Core Duo 1.83GHz out compiles Xcode on a PowerMac Dual G5 2.7GHz.

From http://www.gusmueller.com/blog/archives/2006/1/17.html#1415
Subversion compiled in 5 minutes, 16 seconds on my dual 2.7 g5 with 1.5 gigs of ram. It compiles in 4:32 on the 1.83ghz intel mac with 1 gig of ram. Which makes me happy.
..and from an Apple Xcode team member to Xcode users developer list...

How is the compile speed on the new machines?
On an iMac Core Duo, a tick or two slower than a Quad G5.
 

agreenster

macrumors 68000
Dec 6, 2001
1,892
2
Walt Disney Animation Studios
And this is why I have ZERO interest now in getting a MacBook. For real digital content creation (animation, video editing, etc) the MacBook is a waste of time until software is available (Maya in my case) I just dont want to wait 8 months for Alias to release Maya Universal Binaries

Im looking hard at an XPS M170...
 

dr_lha

macrumors 68000
Oct 8, 2003
1,587
0
agreenster said:
Im looking hard at an XPS M170...
Don't look too hard, your eyes will crack at the sheer ugliness of it.

Reminds me of my old Donkey Kong Game and Watch in the early 80s with that weird brushed metal effect.