iMac 1.8 Ghz G5 Benchmarks

[Kelvin]

Yes there is an even 120 $ card in it now! -20$ for only 64 MB ;-), so a 100 $ card. But I would LOVE to have the Radeon 9600XT finally. I hope there will be midplane upgrades or something like that... (I ordered the 17" with 1.8 Ghz and SD)

With Apple only having 2 GPU vendors in their fold, they have to split the line evenly in their product lineup... so, as a result the GPU in the iMac will be an nVidia, not an ATi.

The powermac goes (good) nVidia, (better) nVidia, (best) ATi
The powerbook goes nVidia, ATi, ATi
The iBook goes ATi, ATi, ATi
The eMac goes ATi
The iMac then goes... nVidia.

So, even with the iMac going nVidia, it's still a little ATi heavy (especially at the upper end; though not if you don't count the eMac). Ah... politics.

 

[gekko513]

XBench is very unreliable in my experience, so I wouldn't read much into it before we get more samples to compare.

 

[eric67]

it is so true

XBench is very unreliable in my experience, so I wouldn't read much into it before we get more samples to compare.

c
guys stop looking at xbench values except for the HD test end RAM test, for the rest, only cinebench gives a real and reliable idea of the iMacG5 performance, so it seems almsot as fast as a PMG5 1.8Ghz SP...not bad at all!! I will even say rather really good.

 

[rdowns]

If upgrades are to be an option for these iMacs in the future, why would Apple marketing keep that bit of info from us now? It seems to me that that is a selling point and therefore would be a fact worth mentioning at the release of the new iMac.

Because you sell what you have today and not some future product. If Apple announced that they would be producing upgrades for it, many would delay purchasing for a faster processor, better video card etc.

Apple used to offer logic board upgrades for many of its models back in the early to mid 90s. They were very expensive and not very popular. Of course costs are lower today and if user installable????

I don't think Apple will offer upgrades.

 

[gerrycurl]

imac sux

I have a dual g4 1.25ghz, i get about 158 using xbench, imacs suck, get a powermac.

 

[Neuro]

Madness

I find it bizarre that people are defending Apple's choice of graphics chip in the G5 iMac. The point is they could have included something much better for a tiny increment in cost, thus including gamers in their product offering.

I'd love to buy an iMac for my lounge, but as I like games as well as my music production tools (which it is fast enough for), the graphics chip is just too slow to justify the purchase.

The G5 towers are too big and expensive for my requirements, so where do I fit into their product offering? I don't, and they will really miss out. I could buy a 'nice' fast lifestyle pc with gamers card, but I really don't want yet another pc in my house.

The iMac specs are ok at present, apart from the graphics card. They will be depressingly poor in a year's time...

 

[iMeowbot]

Apple will never (nor will a 3rd party) offer a replaceable board that contains the CPU, Video Card, etc. It's replaceable incase it breaks, but its just not profitable.

Apple offered exactly that type of upgrade the last time they sold systems with easily-replaced boards, in the LC series.

 

[iMeowbot]

Apple used to offer logic board upgrades for many of its models back in the early to mid 90s. They were very expensive and not very popular. Of course costs are lower today and if user installable????

None of those upgrades from any vendor sold well. They were, however, easy to offer if a newer model used the same chassis, and it allowed manufacturers to eliminate "you can't upgrade it" as a reason not to buy. It's strange that people would buy stuff based on features they will never use, but people are funny like that.

 

[jouster]

All of you consumers are falling into the trap of forking out a ton of money for a crappy investment.

Do you know what an investment is?

It is something you buy with the intention of subsequently selling and making a profit.

All computers are crappy investments.

The powermac goes (good) nVidia, (better) nVidia, (best) ATi

The top of the line GPU for the PowerMac is the nVidia GeForce 6800.

 

[JasonL]

I find it bizarre that people are defending Apple's choice of graphics chip in the G5 iMac. The point is they could have included something much better for a tiny increment in cost, thus including gamers in their product offering.

I'd love to buy an iMac for my lounge, but as I like games as well as my music production tools (which it is fast enough for), the graphics chip is just too slow to justify the purchase.

The G5 towers are too big and expensive for my requirements, so where do I fit into their product offering? I don't, and they will really miss out. I could buy a 'nice' fast lifestyle pc with gamers card, but I really don't want yet another pc in my house.

The iMac specs are ok at present, apart from the graphics card. They will be depressingly poor in a year's time...

I wholeheartedly agree. Some people are fanatical in their defense of Apple, for reasons that make little or no sense to me. Some of their arguments, like how the PM starts with the 5200 so that iMac shouldn't have a better card, make some sense (although the PM having a 5200 is an absolute joke); and other comments are just ridiculous.

The fact is that it is pretty clear to many of us that there is a large demand for a higher end GPU in the iMac. I am personally waiting for the next revision of the iMac in hopes that it will come with (or have at least as an option) a better GPU. If it doesn't then I'll have to get a PM I guess, despite the fact that I'd really like the iMac to help cut down on my workspace clutter.

 

[jouster]

I wholeheartedly agree. Some people are fanatical in their defense of Apple...

Some people just don't care about the GPU. Of course, not many of them are the type that will post here. But a lot of them are the type that will buy this machine.

 

[Neuro]

Some people just don't care about the GPU. Of course, not many of them are the type that will post here. But a lot of them are the type that will buy this machine.

If they don't care about the GPU, then they won't mind if it's better! Gamers would be happy and Apple would sell lots more iMacs.

Let's not forget that in Tiger, numerous processing tasks will be moved from the CPU to the GPU to speed things up. That's a reason for everyone to want a better chip...

 

[G4-power]

Definitely agree with the posts regarding more RAM. I think that Apple really should have made it 512 standard on at least the top two models. At any rate it is good to have some ammo for when I go to the wife to say I need more memory.

I agree with you on that. Most G5 (PM or iMac) probably will increase the memory, so why don't Apple do it already? The 1.8 GHz iMac's could get 512 MB as well as the low-end PM and the Dual 2 and 2.5 GHz PowerMacs should get a gigabyte.

 

[IBSNOWEDIN]

My 733mhz bench

Sorry for the long thing don't know how to put it in a frame.

Results 91.88
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7M34)
Physical RAM 768 MB
Model PowerMac3,4
Processor PowerPC G4 @ 734 MHz
Version 7450 (V'ger) v2.0
L1 Cache 32K (instruction), 32K (data)
L2 Cache 256K @ 367 MHz
L3 Cache 1024K @ 184 MHz
Bus Frequency 134 MHz
Video Card GeForce2 MX
Drive Type Maxtor 5T060H6
CPU Test 87.51
GCD Loop 86.84 3.39 Mops/sec
Floating Point Basic 84.55 305.78 Mflop/sec
AltiVec Basic 89.82 2.61 Gflop/sec
vecLib FFT 87.64 1.36 Gflop/sec
Floating Point Library 88.89 3.56 Mops/sec
Thread Test 64.25
Computation 46.57 628.68 Kops/sec, 4 threads
Lock Contention 103.60 1.30 Mlocks/sec, 4 threads
Memory Test 105.88
System 117.53
Allocate 468.80 305.80 Kalloc/sec
Fill 129.96 1034.50 MB/sec
Copy 63.70 318.51 MB/sec
Stream 96.34
Copy 95.66 699.27 MB/sec [altivec]
Scale 96.62 713.07 MB/sec [altivec]
Add 96.40 616.96 MB/sec [altivec]
Triad 96.68 590.74 MB/sec [altivec]
Quartz Graphics Test 102.75
Line 83.06 2.11 Klines/sec [50% alpha]
Rectangle 88.16 6.20 Krects/sec [50% alpha]
Circle 100.41 2.31 Kcircles/sec [50% alpha]
Bezier 99.49 1.08 Kbeziers/sec [50% alpha]
Text 189.82 3.09 Kchars/sec
OpenGL Graphics Test 82.43
Spinning Squares 82.43 57.68 frames/sec
User Interface Test 150.28
Elements 150.28 48.34 refresh/sec
Disk Test 89.04
Sequential 93.83
Uncached Write 87.80 36.60 MB/sec [4K blocks]
Uncached Write 86.30 35.34 MB/sec [256K blocks]
Uncached Read 127.75 20.22 MB/sec [4K blocks]
Uncached Read 84.58 34.17 MB/sec [256K blocks]
Random 84.70
Uncached Write 74.05 1.11 MB/sec [4K blocks]
Uncached Write 92.48 20.86 MB/sec [256K blocks]
Uncached Read 85.36 0.56 MB/sec [4K blocks]
Uncached Read 89.35 18.39 MB/sec [256K blocks]

 

[segundo]

Apple will never (nor will a 3rd party) offer a replaceable board that contains the CPU, Video Card, etc. It's replaceable incase it breaks, but its just not profitable.

Everything is profitable at a certain price. I'm not saying you are not correct that Apple will not offer an upgrade to the iMac 'mainboard' but I certainly wouldn't rule out upgrades being available from 3rd parties.

That said, I'm pleased with the early benchmark reports we are getting in for the new iMac's. I wish the gpu was a bit faster but yeesh, that horse is dead so let's quit flogging it.

 

[stockscalper]

Not impressive at all. The 1.8 GHZ G5 imac posts a benchmark of 1.34, while the old 17" 1.25 G4 posts a 1.21. That's not that big an improvement.

 

[stuepfnick]

Hello!

Cinebench results of a friend:

Rendering 1 CPU: 189

C4D shading: 203
OpenGL SW-L: 923 CB-GFX
OpenGL HW-L: 1376 CB-GFX

OpenGL Speedup: 6.78

It is a 3 year old Pentium 4 1.8 Ghz, 512 MB RAM with a GeForce3 Titanium with 64MB

I really wonder why the iMac is so weak in OpenGL?? The FX5200Ultra should be about 10-20 % faster than a GeForce3 Ti500 and is in certain performance tests. So what? Why is this? Because of RAM?

 

[JasonL]

What's up with the XBench score of 134? My 15" 1.5GHz Powerbook gets 150. Shouldn't a 1.8G5 w/ full size SATA hard drives do significantly better? (True, I bet this is with the stock 256 RAM).

My PB gets closer to 118. Of course it's a 1.25GHz example; but I do have 1GB of RAM.

Just did Cinebench testing, too, with a CPU render of 229.9. I'd really like to see the iMacs tested with more memory.

An iMac with comparable memory should smoke our machines. Of course, one can get a better GPU in a PB...

 

[rosalindavenue]

I find it bizarre that people are defending Apple's choice of graphics chip in the G5 iMac. The point is they could have included something much better for a tiny increment in cost, thus including gamers in their product offering.

I really think that the graphics card that was chosen for the G5 iMac is all about heat. Graphics cards are hot, so are G5s. It was more important to get the G5 in the iMac than it was to get a newer, hotter graphics card. The card is not terrible (I have the mobile version of the same card in my 9 month old Toshiba laptop), but it is by no means cutting edge either.

 

[H. Georgan]

Here's another guy who got 156 on his xbench test of the new iMac. That's the second time I've seen that score, and it makes sense to me.

http://www.learningloop.org/imacg5/imacg5.html

 

[Chip NoVaMac]

Here's another guy who got 156 on his xbench test of the new iMac. That's the second time I've seen that score, and it makes sense to me.

http://www.learningloop.org/imacg5/imacg5.html

Still based on 256mb memory. My PB 12" rev. B with 1.2gb memory scored 97 on XBench. I seem to remember a score of 88 when it had just 256mb memory.

 

[stuepfnick]

My PB gets closer to 118. Of course it's a 1.25GHz example; but I do have 1GB of RAM.

Just did Cinebench testing, too, with a CPU render of 229.9. I'd really like to see the iMacs tested with more memory.

229.9 can't be true, that's what the Dual 1.25 G4 gets in Cinebench, the 1.25 Ghz single should be around 100 CB points.

BTW: 100 Cinebench points are equal to 1 Ghz of a Pentium 4 system. So the new iMac performs like a 2.54 Ghz Pentium 4 system.

I wonder why OpenGL values are so extremely weak, even software, any ideas?? (see my previous post for more details)

PS: This is for floating point operations only!! (mainly 3D) in other things the G4 is A LOT faster than a equally clocked Pentium 4.

 

[AmigoMac]

My 1GHz iBook G4, 768 RAM, 60 GB, got 100.7 yesterday after the iMac G5 benchmarks were published... we need results with 512 MB and 1 GB in RAM ... that should give us a better idea for those who want to click "Buy now" soon

 

[dudeami]

Plan and simple xBench CAN NOT be used as a benchmarking tool. It is an extremely outdated program that in my honest opinion can not accomplish what it is intended for, Benchmarking. It is a toy, not a reliable tool. It clearly does not benchmark G5's correctly. If you further compare the results of the G4 towers, like the DP 1.42 GHz and the G5 Towers you will see very unexpected results in the CPU test. The program clearly does not accuratly benchmark the difference in CPU performance between G5's and G4's and actually favors the G4 processors on the CPU tests. It has been admitted by the programmer in the discission forum that xBench will not test Dual Processor performance, so it can not be used to benchmark performance differences between single and dual processor systems. You will also see in the results comparison site, that it also does nopt benchmark graphics card performance correctly. This opinion is re-iterated on several other sites that commonly post performance results. You will actually see higher scores using a Gforce 4MX than you will with an ATI 9800 Pro in the same persons machine. No other benchmerking tool would give you those results. I've also read people getting lower scores with the ATI 9200 then with an ATI Rage 128. There are many more examples of how poorly xBench actually benchmarks performance on the results comparison site and in the xBench discussion forum, however he usually does not respond to threads regarding the lack of functionality any more. Xbench has not been updated in almost a year, and in my opinion never ran correctly then either. I truely think the programmer has lost interest, however in the thread where he admitted that xBench does not benchmark dual processor performance, he did mention that you could donate money and mark it for upgrading the program. Again xBench CAN NOT be used to benchmark systems as it is. Maybe it would become a useful tool if it ever gets updated. However, I would hold your breath. I really think that xBench is a dead product, and probably will not see an upgrade.

 

[NicoMan]

If they don't care about the GPU, then they won't mind if it's better! Gamers would be happy and Apple would sell lots more iMacs.

OK but there *might* also be a heat reason for Apple's choice of GPU in the iMac: high-end GPU's need their own fan, huge amount of power, not to mention space: the best GPUs available to the PMac line take 2 card slots!!!

Let's not forget that in Tiger, numerous processing tasks will be moved from the CPU to the GPU to speed things up. That's a reason for everyone to want a better chip...

Agreed but my bet is that's still very small work for *even* a GeForce 5200 Ultra.

My $0.02.