iMac 1.8 Ghz G5 Benchmarks

[TheGimp]

Crappy sound card,eh? Take a look at this:

http://developer.apple.com/documentation/Hardware/Developer_Notes/Macintosh_CPUs-G5/iMacG5/03_Input-Output/chapter_4_section_15.html

Quote from Apple's Developer Connection:

"The sound circuitry and audio device drivers support audio data in multiple formats. Both digital and analog outputs support PCM audio at 16 and 24 bits with sample rates of 32.000 KHz, 44.100 KHz, 48.000 KHz, 64.000 KHz, 88.200 KHz and 96.000 KHz. In addition, the Digital Output also supports AC3 audio at 16 bits with sample rates of 32.000 KHz, 44.100 KHz, 48.000 KHz, 64.000 KHz, 88.200 KHz and 96.000 KHz."


No Macintosh has ever had 24bit 96Khz capability before. I'd say the iMac was made for audio apps.


You're absolutely right! It looks like the new iMac will be the first consumer machine to be able to take full advantage of virtual instrument libraries like the spectacular 24 bit Eastwest Bosendorfer 290 Grand Piano. I use this all the time on my G5 DP 2.5GB (soon to be 4.5GB), and my girlfriend (who owns a real baby grand) just *dies* everytime she comes over to my place. That app alone will get her to buy an iMac to use with her Yamaha digital piano so that she can go to heaven on a regular basis, if what you say about the 24/96 capabilities of the iMac is true.

Still, the iMacs could have used a *smidgen* more ram. Like I wrote, a little extra on the motherboard, and two empty slots.

[mrplow]

The results of these tests depends GREATLY on what your Energy Star settings are for the CPU. The following are my scores for my new 1.6Ghz iMac. (They were dramatically lower before I set the CPU to High instead of Auto)

Results 142.32
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7P35)
Physical RAM 256 MB
Model PowerMac8,1
Processor PowerPC G5 @ 1.60 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 1.15 GHz
Bus Frequency 534 MHz
Video Card GeForce FX 5200
Drive Type ST380013AS
CPU Test 151.66
GCD Loop 90.51 3.53 Mops/sec
Floating Point Basic 249.69 902.95 Mflop/sec
AltiVec Basic 109.54 3.18 Gflop/sec
vecLib FFT 176.42 2.74 Gflop/sec
Floating Point Library 320.75 12.84 Mops/sec
Thread Test 92.78
Computation 61.46 829.77 Kops/sec, 4 threads
Lock Contention 189.13 2.37 Mlocks/sec, 4 threads
Memory Test 207.98
System 228.16
Allocate 576.92 376.33 Kalloc/sec
Fill 178.42 1420.19 MB/sec
Copy 172.11 860.53 MB/sec
Stream 191.07
Copy 168.22 1229.69 MB/sec [G5]
Scale 165.93 1224.59 MB/sec [G5]
Add 216.90 1388.16 MB/sec [G5]
Triad 229.73 1403.68 MB/sec [G5]
Quartz Graphics Test 189.60
Line 177.92 4.53 Klines/sec [50% alpha]
Rectangle 165.02 11.61 Krects/sec [50% alpha]
Circle 185.24 4.27 Kcircles/sec [50% alpha]
Bezier 175.51 1.91 Kbeziers/sec [50% alpha]
Text 278.16 4.53 Kchars/sec
OpenGL Graphics Test 175.56
Spinning Squares 175.56 122.85 frames/sec
User Interface Test 225.89
Elements 225.89 72.66 refresh/sec
Disk Test 86.15
Sequential 77.71
Uncached Write 74.27 30.96 MB/sec [4K blocks]
Uncached Write 58.29 23.87 MB/sec [256K blocks]
Uncached Read 78.14 12.37 MB/sec [4K blocks]
Uncached Read 124.16 50.16 MB/sec [256K blocks]
Random 96.65
Uncached Write 89.66 1.35 MB/sec [4K blocks]
Uncached Write 97.69 22.03 MB/sec [256K blocks]
Uncached Read 93.25 0.62 MB/sec [4K blocks]
Uncached Read 107.87 22.20 MB/sec [256K blocks]

[Uragon]

I have indeed, thanks everyone.

So the question that remains to me is, would the matched pair of 512's be an advantage over the unmatched 256 + 1 gig?

would an extra quarter gig be more or less important than the matching issue :confused:

I checked with one of Apple's reseller, and he suggested that a pair 512 is better and will be cheaper than 1 gig + 512, though he reminded that this is his personal "opinion" only.

[H. Georgan]

Can someone with a new iMac check the CAS of Apple installed ram in the system profiler for me? Just go open the system profiler, select "memory" on the left pane, and select one of the Apple installed dimms on the right pane. It should say the speed in the lower pane. My iBook, for example, says "Speed: PC133-333" for the aftermarket dimm that I installed, which I think means it has a CAS of 3-3-3. Thanks in advance!

[kcmac]

Can someone with a new iMac check the CAS of Apple installed ram in the system profiler for me? Just go open the system profiler, select "memory" on the left pane, and select one of the Apple installed dimms on the right pane. It should say the speed in the lower pane. My iBook, for example, says "Speed: PC133-333" for the aftermarket dimm that I installed, which I think means it has a CAS of 3-3-3. Thanks in advance!

I just put in 2 sticks of 512. They are Princeton 512MB 400Mhz. CL3 184PIN (32X8) . This is what shows up on the System Profiler.

DIMM1/J4001:

Size: 512 MB
Type: DDR SDRAM
Speed: PC3200U-25440

Edit:

This is what is on the base 256 that I removed.

MTBVDDT3264AG-40BG4 200433 CBNBM2M012
COUNTRY OF ORIGIN SINGAPORE
PC3200U-30331-A1
256MB, DDR, 400, CL3

The following was added by another poster on macnn.
"Ah ha! Thanks. A quick Google reveals this link. So this memory module is made by Crucial:

Crucial Technology
PC3200U-30331-A1
MT8VDDT6464AG-40BC1
CCTAGA3013 200407
CT6464Z40B.8T
42545

Speed type: PC 3200, 200 MHz (DDR400)
Chip capacity: 8x 512 Mbit, 5 sn
Sides: single
SPD: 200 MHz 3.0-3-3-8, 166 MHz 2.5-3-3-7, 133 MHz 2.0-2-2-6"

[G4-power]

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.

I agree with you on that. Especially the Xbench CPU test sucks. It does well on RAM and HD testing, but the graphics test doesn't do very well either.
The idea is good, but lack of upgrades, probably because lack of money?
I like the graphics test idea, though. It has a separate window that has the same resolution, no matter what your screen resolution is.
Also, the results on the same machine with same specs may vary quite much, I have usually rebooted again to get better results.
Cinebech, then, isn't the best choice either, but is okay in testing CPU speed between different Macs, but it runs faster on PCs, because its been "optimized" more for them, and uses floating point operations, which I've heard P4's are good in. In testing graphics acceleration, it's the best benchmark apart from real world tests out there.
Of course, all this is based on my humble opinions, but this is what i think.

[chazmox]

Visited the Apple store today and ran Xbench on a 1.6 iMac G5 with 1 Gig of ram. I got an astondingly low 105-109 ( I ran twice ). I ran Xbench on a 12 inch PB ( 1.33 G4 and 256M of RAM ) and got 95.

I quit all apps but the Apple store usually have stuff running in the background, but I was still surprised at how low the scores were for the iMac G5.

I know that XBench isn't the greatest, but it is the most widely used tool and does split it's scores out so you can see which criteria has the most impact ( FP was very good on the G5 as expected ).

I tried to find any energy setting, but couldn't find one for the processor but the Energy Saver settings offered the same options as I have for my iMac G4...

[Elan0204]

Visited the Apple store today and ran Xbench on a 1.6 iMac G5 with 1 Gig of ram. I got an astondingly low 105-109 ( I ran twice ). I ran Xbench on a 12 inch PB ( 1.33 G4 and 256M of RAM ) and got 95.

I tried to find any energy setting, but couldn't find one for the processor but the Energy Saver settings offered the same options as I have for my iMac G4...

Based on others' reports, I would say that it safe to assume that if you had turned the processor setting to highest, the iMac would have done much better. The Processor Performance setting is under the Energy Saver preference pane in System Preferences. It is under the "Options" tab. The ability to change the setting is unavailable on the computers in the Apple Stores, however, unless you know the administrator password. You would need to convince an employee to change the setting for you.

[mizzman36]

I popped in to my local Apple Store yesterday ran xBench on a DP 2.0G PM G5 with 512 RAM and a 1.8G iMac with 256 RAM.

I shut all running programs and set processor to "Highest". Results were:

PM: 216
iMac: 153

[sjl]

Unfortunately top in OS X doesn't support the F keystroke :(

Only platform where I have gotten that to work is Linux...
Bugger. Bugger bugger bugger. I posted that when I didn't have access to my PowerBook, so I assumed that what worked for top on one platform would work for top on another. :mad: Excuse me whilst I quietly take my foot out of my mouth and check on my PowerBook... *trudge trudge trudge* (five or ten minutes later) Okay, you're right -- it doesn't support the F (or the equivalent O) keystroke. However, if you fire up top with the command "top -o -vsize", the end result is the same.

Nyah. :D

(man top is your friend. :D)

[IBSNOWEDIN]

Can't wait to get mine still havn't got the e-mail on shipping, but as soon as i get it i'm going to pop 2gb of ram in it that i just got and do a benchmark on it and post it up to see what the imac can do whne it is maxed up.

[stevesien]

So we have the iMac G5 scoring 134 with 256MB Ram and the Original single G5 1.8 w Radeon 9800 SE, 1.5GB ram scoring 124.

Based on that info, regardless of Xbench's qualities, we must conclude that the iMac is a very quick machine in comparison.

here is with full power, and auto below (same as previous post)
Results 118.14
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7M34)
Physical RAM 1536 MB
Model PowerMac7,2
Processor PowerPC 970 @ 1.80 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 1.80 GHz
Bus Frequency 900 MHz
Video Card ATY,R350
Drive Type ST3160023AS
CPU Test 164.59
GCD Loop 94.70 3.70 Mops/sec
Floating Point Basic 265.28 959.33 Mflop/sec
AltiVec Basic 123.24 3.58 Gflop/sec
vecLib FFT 202.74 3.15 Gflop/sec
Floating Point Library 332.96 13.33 Mops/sec
Thread Test 97.80
Computation 64.41 869.59 Kops/sec, 4 threads
Lock Contention 202.99 2.55 Mlocks/sec, 4 threads
Memory Test 249.06
System 250.27
Allocate 567.47 370.16 Kalloc/sec
Fill 186.42 1483.94 MB/sec
Copy 205.72 1028.61 MB/sec
Stream 247.87
Copy 192.60 1407.94 MB/sec [G5]
Scale 229.23 1691.70 MB/sec [G5]
Add 303.09 1939.79 MB/sec [G5]
Triad 304.55 1860.83 MB/sec [G5]
Quartz Graphics Test 184.89
Line 181.19 4.61 Klines/sec [50% alpha]
Rectangle 195.83 13.78 Krects/sec [50% alpha]
Circle 196.94 4.54 Kcircles/sec [50% alpha]
Bezier 165.79 1.80 Kbeziers/sec [50% alpha]
Text 188.37 3.07 Kchars/sec
OpenGL Graphics Test 128.14
Spinning Squares 128.14 89.67 frames/sec
User Interface Test 193.61
Elements 193.61 62.27 refresh/sec
Disk Test 48.64
Sequential 46.86
Uncached Write 22.63 9.43 MB/sec [4K blocks]
Uncached Write 68.13 27.90 MB/sec [256K blocks]
Uncached Read 78.17 12.37 MB/sec [4K blocks]
Uncached Read 72.95 29.47 MB/sec [256K blocks]
Random 50.56
Uncached Write 28.51 0.43 MB/sec [4K blocks]
Uncached Write 49.86 11.24 MB/sec [256K blocks]
Uncached Read 85.04 0.56 MB/sec [4K blocks]
Uncached Read 81.86 16.85 MB/sec [256K blocks]

here is with auto

Results 124.41
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7M34)
Physical RAM 1536 MB
Model PowerMac7,2
Processor PowerPC 970 @ 1.80 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 1.80 GHz
Bus Frequency 900 MHz
Video Card ATY,R350
Drive Type ST3160023AS
CPU Test 120.78
GCD Loop 76.95 3.00 Mops/sec
Floating Point Basic 173.62 627.89 Mflop/sec
AltiVec Basic 89.02 2.59 Gflop/sec
vecLib FFT 139.06 2.16 Gflop/sec
Floating Point Library 237.06 9.49 Mops/sec
Thread Test 81.74
Computation 56.76 766.22 Kops/sec, 4 threads
Lock Contention 146.00 1.83 Mlocks/sec, 4 threads
Memory Test 214.03
System 187.02
Allocate 343.28 223.92 Kalloc/sec
Fill 143.33 1140.93 MB/sec
Copy 162.57 812.84 MB/sec
Stream 250.17
Copy 200.43 1465.13 MB/sec [G5]
Scale 228.55 1686.70 MB/sec [G5]
Add 311.62 1994.34 MB/sec [G5]
Triad 292.79 1788.97 MB/sec [G5]
Quartz Graphics Test 156.14
Line 108.95 2.77 Klines/sec [50% alpha]
Rectangle 172.75 12.15 Krects/sec [50% alpha]
Circle 181.67 4.19 Kcircles/sec [50% alpha]
Bezier 161.15 1.75 Kbeziers/sec [50% alpha]
Text 187.07 3.05 Kchars/sec
OpenGL Graphics Test 131.86
Spinning Squares 131.86 92.27 frames/sec
User Interface Test 192.26
Elements 192.26 61.84 refresh/sec
Disk Test 84.10
Sequential 98.22
Uncached Write 123.57 51.51 MB/sec [4K blocks]
Uncached Write 114.14 46.74 MB/sec [256K blocks]
Uncached Read 65.57 10.38 MB/sec [4K blocks]
Uncached Read 116.01 46.87 MB/sec [256K blocks]
Random 73.54
Uncached Write 53.23 0.80 MB/sec [4K blocks]
Uncached Write 86.19 19.44 MB/sec [256K blocks]
Uncached Read 74.12 0.49 MB/sec [4K blocks]
Uncached Read 95.10 19.57 MB/sec [256K blocks]

[chazmox]

Based on others' reports, I would say that it safe to assume that if you had turned the processor setting to highest, the iMac would have done much better. The Processor Performance setting is under the Energy Saver preference pane in System Preferences. It is under the "Options" tab. The ability to change the setting is unavailable on the computers in the Apple Stores, however, unless you know the administrator password. You would need to convince an employee to change the setting for you.

I saw the lock on for admin privledges and knew I couldn't change anything - but I swear I did not see that processor selection pull down - I looked so I could note what the processor performance was set at. Maybe I missed it, but I did look in each tab of energy saver...

My 15" PB G4 at 1.25 gets around 118-120. I've compared a few scores and many places the G5 does better ( memory and drive access the G5 does worse? ) but it still gets a equal or lessor score...

[Telomar]

One of the big problems with Xbench is the way it weights things to get its final number. In a lot of respects those weightings don't match with what is most important to performance. I also have some serious questions over some of the subtests. Too much variation in general though for it to be useful.

[ki-goi]

wow, the xbench hard drive measurements are very unpredictable. they seem to be the thing that's turning all the numbers upside down. oops has that already been said?

anyway i expected any given imac to have lower relative hard drive performance compared to pro machines. these numbers though are very strange, what does it mean that a single computer's hard drive throughput varies as much as 2x from test to test? activity in the background?

what's good software for hard drive testing... somebody should do that.

[edit] i ran xbench on my 550MHz G4. the brains and graphics results were nearly constant, the random disk test results were nearly constant, and the sequential read tests were all over the place - but, this ATA-33 machine's disk scores beat some of the posted SATA imac scores - 64 overall, 55 sequential, 76 random. why is the imac slower.

[mrzippy]

Maybe I have missed this in this very long thread, but has the poor speed of the hard drive been discussed?

Considering it's supposed to be a Serial ATA drive with an 8mb buffer why are the benchmarks so poor.

I am looking to replace my PM Dual 1.25 (FW800) + 15" TFT with an iMac 20" but the HDD benchmark are really putting me off.

Just done some benchmarks below, they seem to be way lower than the last time I did them, maybe a bug in OS X 10.3.5 ? Otherwise the iMac HDD would seem to be better than mine :(

Neil

My benchmarks on PM DP 1.25 GHz (FW800)

Stock Power Mac 80 GB IBM IDE 8mb Cache
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7M34)
Physical RAM 1536 MB
Model PowerMac3,6
Processor PowerPC G4x2 @ 1.25 GHz
Version 7455 (Apollo) v3.3
L1 Cache 32K (instruction), 32K (data)
L2 Cache 256K @ 1.25 GHz
L3 Cache 1024K @ 250 MHz
Bus Frequency 167 MHz
Video Card ATY,RV250
Drive Type IBM-IC35L090AVV207-1

Disk Test 86.37
Sequential 94.18
Uncached Write 86.46 36.04 MB/sec [4K blocks]
Uncached Write 75.71 31.01 MB/sec [256K blocks]
Uncached Read 181.55 28.74 MB/sec [4K blocks]
Uncached Read 82.03 33.14 MB/sec [256K blocks]
Random 79.76
Uncached Write 74.14 1.11 MB/sec [4K blocks]
Uncached Write 100.16 22.59 MB/sec [256K blocks]
Uncached Read 70.14 0.46 MB/sec [4K blocks]
Uncached Read 80.51 16.57 MB/sec [256K blocks]

Added 2nd HDD - 120 GB IBM IDE 8mb Cache
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7M34)
Physical RAM 1536 MB
Model PowerMac3,6
Processor PowerPC G4x2 @ 1.25 GHz
Version 7455 (Apollo) v3.3
L1 Cache 32K (instruction), 32K (data)
L2 Cache 256K @ 1.25 GHz
L3 Cache 1024K @ 250 MHz
Bus Frequency 167 MHz
Video Card ATY,RV250
Drive Type IC35L120AVV207-1

Disk Test 83.75
Sequential 77.69
Uncached Write 76.36 31.83 MB/sec [4K blocks]
Uncached Write 58.05 23.77 MB/sec [256K blocks]
Uncached Read 165.03 26.13 MB/sec [4K blocks]
Uncached Read 66.20 26.75 MB/sec [256K blocks]
Random 90.84
Uncached Write 100.92 1.51 MB/sec [4K blocks]
Uncached Write 85.97 19.39 MB/sec [256K blocks]
Uncached Read 96.51 0.64 MB/sec [4K blocks]
Uncached Read 82.42 16.96 MB/sec [256K blocks]

[ki-goi]

i don't want to be accused of FUD - there was only one test quoted in this thread that was worse than my ATA/33 G4:

Disk Test 48.64 - Sequential 46.86 - Random 50.56

the rest of the imac results were around the same place, all scores beating my older computer:

Disk Test 79.19 - Sequential 67.00 - Random 96.82
Disk Test 93.76 - Sequential 86.24 - Random 102.70
Disk Test 92.47 - Sequential 87.38 - Random 98.20
Disk Test 86.15 - Sequential 77.71 - Random 96.65
Disk Test 84.10 - Sequential 98.22 - Random 73.54

i found on my computer that xbench's random test is more reliable, and this seems to be true here, too. for the imac, the range of sequential results (67.00-98.22, tossing the very low one) is more variable than the random results (73.54-102.7, again tossing that low one).

it looks like the new imac will regularly score above 96 on the random test, maybe that's how to use xbench to look at drive performance. it's certainly better than my computer's repeated 76 score.

[edit] i changed some stuff above... none of the numbers... just making it point blank clear.

[punkmac]

Has anyone thought of upgrading a 1.6 combo model with a dual layer Superdrive and third party 250 gig HD?

I feel Apple knows we are going to ditch the stock RAM anyway so why not keep it at the lowest possible. I would rather throw away 256 not 512.

Just wondering...


I.

[gooddog]

Why do i have the gut feeling that the iMac LCDs are not even close to the new Apple Cinema displays? I'm thinking they are probably either the old 20" or the same ones as the original G4 iMac line.



millions doesn't inspire much confidence for me -_-

**********************************

Well, the dot pitch works out the same for both.

{ {20cos[arctan (10/16)]}•(25.4) } ÷ (1680 ) = 2.58 mm

This is what is also posted for the Cinema 20".

But I could find nothing for iMac's response time to compare with Cinema's 16 ms.

---gooddog

[H. Georgan]

I just put in 2 sticks of 512. They are Princeton 512MB 400Mhz. CL3 184PIN (32X8) . This is what shows up on the System Profiler.

DIMM1/J4001:

Size: 512 MB
Type: DDR SDRAM
Speed: PC3200U-25440

Edit:

This is what is on the base 256 that I removed.

MTBVDDT3264AG-40BG4 200433 CBNBM2M012
COUNTRY OF ORIGIN SINGAPORE
PC3200U-30331-A1
256MB, DDR, 400, CL3

The following was added by another poster on macnn.
"Ah ha! Thanks. A quick Google reveals this link. So this memory module is made by Crucial:

Crucial Technology
PC3200U-30331-A1
MT8VDDT6464AG-40BC1
CCTAGA3013 200407
CT6464Z40B.8T
42545

Speed type: PC 3200, 200 MHz (DDR400)
Chip capacity: 8x 512 Mbit, 5 sn
Sides: single
SPD: 200 MHz 3.0-3-3-8, 166 MHz 2.5-3-3-7, 133 MHz 2.0-2-2-6"

Thank you! That confirms my suspicion that iMacs are shipping with CL3 memory. Now I can go ahead and order another 512mb CL3 and it should (fingers crossed) be considered a match with the one installed by Apple.

[markholloway@ma]

Tested a G5 iMac/1.8/512MB at the Apple Store today!

1) FYI - Cinebench for G5 is beta. Some optimizations missing.
2) I didn't have an admin password to change the Energy Saver from Automatic to Highest Performance on the iMac G5.
3) Aluminum Powerbook is my personal machine tested at home (Energy Save set to Highest Performance)





CINEBENCH 2003 v1
****************************************************

Tester : Mark Holloway

Processor : Apple Aluminum Powerbook
MHz : 1500
Number of CPUs : 1
Operating System : 10.3.5

Graphics Card : ATI Mobility Radeon 9700
Resolution : 1280x854
Color Depth : millions

****************************************************

Rendering (Single CPU): 133 CB-CPU
Rendering (Multiple CPU): --- CB-CPU


Shading (CINEMA 4D) : 163 CB-GFX
Shading (OpenGL Software Lighting) : 435 CB-GFX
Shading (OpenGL Hardware Lighting) : 819 CB-GFX

OpenGL Speedup: 5.02

****************************************************

CINEBENCH 2003 v1 (G5 Optimized Beta)
****************************************************

Tester : Mark Holloway

Processor : imac g5
MHz : 1800
Number of CPUs : 1
Operating System : 10.3.5

Graphics Card : nvidia Geforce FX 5200 Ultra
Resolution : 1400x900
Color Depth : millions

****************************************************

Rendering (Single CPU): 241 CB-CPU
Rendering (Multiple CPU): --- CB-CPU


Shading (CINEMA 4D) : 231 CB-GFX
Shading (OpenGL Software Lighting) : 546 CB-GFX
Shading (OpenGL Hardware Lighting) : 674 CB-GFX

OpenGL Speedup: 2.92

****************************************************

CINEBENCH 2003 v1
****************************************************

Tester : Mark Holloway

Processor : iMac G5
MHz : 1800
Number of CPUs : 1
Operating System : 10.3.5

Graphics Card : nvidia Geforce FX 5200 Ultra
Resolution : 1400x900
Color Depth : millions

****************************************************

Rendering (Single CPU): 183 CB-CPU
Rendering (Multiple CPU): --- CB-CPU


Shading (CINEMA 4D) : 226 CB-GFX
Shading (OpenGL Software Lighting) : 515 CB-GFX
Shading (OpenGL Hardware Lighting) : 699 CB-GFX

OpenGL Speedup: 3.09

****************************************************

[MegaSignal]

Just curious if anyone's been able to do a non-scientific test with regard to the iMac G5's performance of iDVD encoding: if so,

How long does it take in reference to the length of the movie?

Do the fans spool up and get noisy?

With regard to the displays: Any major problems with burnt or stuck pixels?

[ssamani]

Why should someone use such Energy Saver settings in a *desktop* computer anyway?

To do our bit to help save the planet.

[Stridder44]

Now that i've expressed my dissapointed state, i'd also like to say that other than the GPU, i believe this to be a well priced and zippy consumer level mac. I just can't get over the fact that this can't even run Halo.

Halo blows anyway :cool:

[toti]

Halo blows anyway :cool:

It may either suck or blow... dunno, havn't played it :)

But I'd say the requirements for Halo are WELL BELOW that the iMac G5 so I guess that dude is just trolling....

[nospleen]

Just curious if anyone's been able to do a non-scientific test with regard to the iMac G5's performance of iDVD encoding: if so,

How long does it take in reference to the length of the movie?

Do the fans spool up and get noisy?

With regard to the displays: Any major problems with burnt or stuck pixels?

1. I just did a little 8 minute movie of my sons birth. (about 80 pictures with the Ken Burns effect). I burned it in iDVD and it took about 25 minutes from start to finish. I know it is not a full length one, but I wanted to give you an idea.)
2. Yes, the fans spool up and get noisy. Not too bad, but I think it is louder than the old imac.
3. I have a 20" and I did not have any dead or stuck pixels.

[chazmox]

When you think about it - that seems like a real world comparison, but it's really not that consistent of comparison...

iDVD encodes in the background while you are assembling the movie/DVD, so depending on how long you take, it may be well on the way to finishing encoding before you select "burn." Now if there was a way to force iDVD to only encode after you've selected "burn" then it could be a good comparison.

[andyblac]

Now if there was a way to force iDVD to only encode after you've selected "burn" then it could be a good comparison.


there is, if you open your project in iDVD, then go into preferences in general disable "enable background encoding" :D

[MegaSignal]

When you think about it - that seems like a real world comparison, but it's really not that consistent of comparison...

iDVD encodes in the background while you are assembling the movie/DVD, so depending on how long you take, it may be well on the way to finishing encoding before you select "burn." Now if there was a way to force iDVD to only encode after you've selected "burn" then it could be a good comparison.

Very true - but in doing this test, I always monitor the "Status" menu after having selected "Best Performance" along with "Background Encoding Enabled"; in this way, differences in actual media burn times (2x, 4x, etc.) will not be included. I time the process from the moment the progress bar indicates progress, to the time when "Done" is displayed. My apologies for not mentioning this earlier. Nonetheless, I appreciate nospleen's input with regard to all of my inquiries with regard to speed, noise, and display. Sounds like a great machine so far!

[MegaSignal]

there is, if you open your project in iDVD, then go into preferences in general disable "enable background encoding" :D

In addition, "Best Quality" uses Compressor's two-pass system for better quality and up to 2 hours of movie and menu time. I used the "Best Performance" option merely as a comparison test (see above).

[mcdawson]

Just curious if anyone's been able to do a non-scientific test with regard to the iMac G5's performance of iDVD encoding: if so,

How long does it take in reference to the length of the movie?

For best comparison, I would burn to a disk image--that takes out the burn time as a factor (I personally always burn to a disk image to better figure out if any iDVD jitters are part of the burning or encoding process). The hack to allow burning to disk images (or external DVD burners) can be found (with instructions) at http://homepage.mac.com/geerlingguy/personal_site/mac_guide/pages/15-
burn_idvd-faq/idvd_hack_faq.html

[squeak]

nospleen says:
2. Yes, the fans spool up and get noisy. Not too bad, but I think it is louder than the old imac.


This is where having Airport Express shines. Simply crank up the music
over there on the stereo.

I have 0db fan noise in my iMac - a Ruby DV. Apple says they have quiet fans but compared to silence nothing beats nothing (when my PC is on-whirr whirr whirr).
I even hate what my dvd player goes through to play a dvd.
I can't wait for solid state.

squeak

[nospleen]

nospleen says:
2. Yes, the fans spool up and get noisy. Not too bad, but I think it is louder than the old imac.


This is where having Airport Express shines. Simply crank up the music
over there on the stereo.

I have 0db fan noise in my iMac - a Ruby DV. Apple says they have quiet fans but compared to silence nothing beats nothing (when my PC is on-whirr whirr whirr).
I even hate what my dvd player goes through to play a dvd.
I can't wait for solid state.

squeak

Now I have an excuse to buy the Airport Express. :p The noise is pretty annoying though, but it does not bother me that much. But, I am surprised they advertise it as 'whisper quiet'.

[sniggip]

I'm a newbie, but I've been wondering what kind of a speed bump I'd get from my 1st edition G4 450 to a dual G5 2ghz?

[harperb]

My iMac g5 was set to "automatic"; when i turned it to highest GUI performance almost quadrupled. iTunes went from using 14-16% of my CPU to 2-4%.

[ki-goi]

My iMac g5 was set to "automatic"; when i turned it to highest GUI performance almost quadrupled. iTunes went from using 14-16% of my CPU to 2-4%.
that's what's cool about "automatic" mode. when the processor only has a few things to do, it slows itself down. so you're seeing a single app take more of the processor's time, because there's less time to go around.

to see why "automatic" is cool, turn it on, then turn on activity monitor, and start more apps. have the apps do things. play a movie, import songs in itunes, things like that. as you load it up, when it looks like the processor's going to max out, watch the magic, suddenly everybody's using less processor time, the computer's not going to max out anymore, the CPU's put itself back to full speed to do lots of work...

[Invizzible]

When I upgraded my 400 Mhz G4 Powermac to a single processor 1.4 Ghz, I got a 2.5 X real world speed increase. When I then compared my upgraded Powermac to my sister's 1.5 Ghz G4 Powerbook, hers was approx. 20% faster than mine. If you look on barefeats.com, you'll see benchmarks that compare the dual 2.0 Ghz and dual 2.5 Ghz G5's to, among other things, the 1.5 Ghz Powerbook. I'll leave the math up to you.

[gooddog]

A few points...

By now, I'm sure, Apple knows you're unbelievably, extraordinarily, upset by this graphics card. They got it, ok?

This card allows Apple to incrementally 'upgrade-to-death' the new iMac so that for the next 3-5 years they can introduce a 'new' model with better specs. Apple is known for this, right?

If you don't like the card, buy a PM and customize to your hearts content.

Now get OVER IT!!!!!!!!!!!

******************

BTW, about when do you all think we might see the first upgrade to this iMac ?

I am budgeting for right after the X-mas holidays....

Man , would I love a 23" iMac with a faster GPU !!!!


--------

I saw my first 20" G5 iMac
today and played a DVD ( Frida ) ---- ohmygosh :) :) :) !!!!
It was a dead perfect playback with jet black blacks, awesome reds, etc. and not a single frame skip , even though it had only the 256 RAM.

I calculate the dot pitch as equal to the new 20" cinema. The store dude seemed very knowledgeable and said he was very sure the response time for the 20" G5 iMac LCD is the same ( 16ms ) as for the stand alone new cinema 20" display: it is the same LCD.

Does anyone with experience expect that
the 20" / 1.86 GHz / 2GB RAM G5 iMac will be able to generate and render/playback
complex MAYA UNLIMITED for OS X files when this app' streets ?

This is my only GPU concern.

I am not a gamer at all, but would like to try MAYA soon and I don't like the huge clunky PowerMac --- too much : I love the look/feel of the all-in-one iMac G5.

---gooddog

[aswitcher]

******************

BTW, about when do you all think we might see the first upgrade to this iMac ?

I am budgeting for right after the X-mas holidays....

Man , would I love a 23" iMac with a faster GPU !!!!



---gooddog

I think Apple would be concerned about the cost limiting sales and/or it bitting into the PowerMac line if they produced a 23". It would be an interesting product to see but can't say it would necessarily increase over all Apple sales and might be a big risk....

[Gee4orce]

But, I am surprised they advertise it as 'whisper quiet'.

I dunno. I can whisper pretty loud...

[Little Endian]

Barefeats has posted some realworld Benchmarks. http://barefeats.com/imacg5.html

I question the accuracy of the Benchmarking as in the Motion Playback without render seems to show the PowerMac Dual G5 in very bad light while the Dual 2Ghz sprints ahead at more than three times faster. I would have also liked to see some Photoshop Benchmarks as well as Quake III scores, old but still effective.

[nospleen]

Not bad at all, I knew this thing felt fast. :) I was kind of surprised to see it beat the dual 1.42 in quite a few things.

[corbin_a2]

Results 206.02
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7P35)
Physical RAM 1024 MB
Model PowerMac8,1
Processor PowerPC G5 @ 1.80 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 1.15 GHz
Bus Frequency 600 MHz
Video Card GeForce FX 5200
Drive Type ST380013AS
CPU Test 171.21
GCD Loop 102.72 4.01 Mops/sec
Floating Point Basic 284.00 1.03 Gflop/sec
AltiVec Basic 122.58 3.56 Gflop/sec
vecLib FFT 199.07 3.09 Gflop/sec
Floating Point Library 361.50 14.47 Mops/sec
Memory Test 258.61
System 281.25
Allocate 666.97 435.06 Kalloc/sec
Fill 220.96 1758.84 MB/sec
Copy 215.45 1077.24 MB/sec
Stream 239.33
Copy 211.89 1548.89 MB/sec [G5]
Scale 210.54 1553.77 MB/sec [G5]
Add 268.74 1719.93 MB/sec [G5]
Triad 283.88 1734.48 MB/sec [G5]

[Telomar]

It's been said that 80% of Mac buyers never upgrade their machine beyond adding memory. If that's true, the iMac G5 will be ideal for most Mac desktop buyers. I wonder if that's because around 80% of Apple's sales consist of either portables or desktops that resemble portables and as a result aren't that upgradeable or if it isn't just because that number is made up :p

[Little Endian]

I wonder if that's because around 80% of Apple's sales consist of either portables or desktops that resemble portables and as a result aren't that upgradeable or if it isn't just because that number is made up :p

No I think 80% not upgrading beyond RAM is relatively accurate however I do believe it is that high due to Portables, imac, and emac sales probably make up nearly 80% of Apple Computer sales. The imac tempts me big time especially with it's nice price point to performance ratio. I will skip an imac this time around and my next machine will be a PowerMac.

[ki-goi]

I will skip an imac this time around and my next machine will be a PowerMac.
yep for those who wait, maybe a dual-dual-core machine is waiting around the corner... quad-PPC... if you wait... a... little... longer...

[H. Georgan]

Results 206.02
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7P35)
Physical RAM 1024 MB
Model PowerMac8,1
Processor PowerPC G5 @ 1.80 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 1.15 GHz
Bus Frequency 600 MHz
Video Card GeForce FX 5200
Drive Type ST380013AS
CPU Test 171.21
GCD Loop 102.72 4.01 Mops/sec
Floating Point Basic 284.00 1.03 Gflop/sec
AltiVec Basic 122.58 3.56 Gflop/sec
vecLib FFT 199.07 3.09 Gflop/sec
Floating Point Library 361.50 14.47 Mops/sec
Memory Test 258.61
System 281.25
Allocate 666.97 435.06 Kalloc/sec
Fill 220.96 1758.84 MB/sec
Copy 215.45 1077.24 MB/sec
Stream 239.33
Copy 211.89 1548.89 MB/sec [G5]
Scale 210.54 1553.77 MB/sec [G5]
Add 268.74 1719.93 MB/sec [G5]
Triad 283.88 1734.48 MB/sec [G5]

This is possible if you don't run the full suite of xbench tests, as this guy did. If there's a "*" next to the score on the xbench site that means it's only a partial test.

If, for example, you only run the cpu test on the iMac you would quite likely get that kind of score.

Too bad -- it would have been nice to see such a high score! I think the upper-threshold for the whole suite of tests is around 165 for G5 iMacs, judging from what I've seen.

[andyblac]

i am wondering if they had the processor set highest or automatic ?

[RealDeal]

17inch G5 iMac 1.6GHz 1024MB RAM (pair 512/s) set a maximum performance, no HD sleep (cook the chips!)...

CINEBENCH 2003-
rendering 1 cpu- 170
graphics C4D shading 204
OpenGL SW-L 529
OpenGL HW-l 928
Speedup 4.59

XBENCH
Results 146.68
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7P35)
Physical RAM 1024 MB
Model PowerMac8,1
Processor PowerPC G5 @ 1.60 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 1.15 GHz
Bus Frequency 534 MHz
Video Card GeForce FX 5200
Drive Type ST3160023AS
CPU Test 152.33
GCD Loop 91.67 3.58 Mops/sec
Floating Point Basic 253.68 917.39 Mflop/sec
AltiVec Basic 109.14 3.17 Gflop/sec
vecLib FFT 175.35 2.72 Gflop/sec
Floating Point Library 321.72 12.88 Mops/sec
Thread Test 93.57
Computation 61.58 831.34 Kops/sec, 4 threads
Lock Contention 194.69 2.44 Mlocks/sec, 4 threads
Memory Test 241.68
System 256.77
Allocate 603.26 393.50 Kalloc/sec
Fill 202.57 1612.48 MB/sec
Copy 196.49 982.46 MB/sec
Stream 228.27
Copy 203.78 1489.62 MB/sec [G5]
Scale 203.62 1502.71 MB/sec [G5]
Add 252.85 1618.24 MB/sec [G5]
Triad 266.67 1629.38 MB/sec [G5]
Quartz Graphics Test 192.20
Line 180.74 4.60 Klines/sec [50% alpha]
Rectangle 167.97 11.82 Krects/sec [50% alpha]
Circle 187.06 4.31 Kcircles/sec [50% alpha]
Bezier 177.26 1.93 Kbeziers/sec [50% alpha]
Text 282.44 4.60 Kchars/sec
OpenGL Graphics Test 175.58
Spinning Squares 175.58 122.87 frames/sec
User Interface Test 233.97
Elements 233.97 75.26 refresh/sec
Disk Test 89.61
Sequential 80.76
Uncached Write 74.68 31.13 MB/sec [4K blocks]
Uncached Write 58.73 24.05 MB/sec [256K blocks]
Uncached Read 88.36 13.99 MB/sec [4K blocks]
Uncached Read 128.34 51.85 MB/sec [256K blocks]
Random 100.65
Uncached Write 95.51 1.43 MB/sec [4K blocks]
Uncached Write 96.80 21.83 MB/sec [256K blocks]
Uncached Read 96.92 0.64 MB/sec [4K blocks]
Uncached Read 115.93 23.86 MB/sec [256K blocks]

[Santaduck]

Latest results of my collaboration with Rob-ART of barefeats.com here:
http://www.macologist.org/viewtopic.php?p=2009#2009

We test at CPU highest, network disabled, all apps closed, all finder windows closed, etc.

We are using my Santaduck benchmark for UT2004 (http://www.santaduck.com/ut2k4/Santaduck_Toolpak_pub.sit), the same that Apple used (http://www.apple.com/imac/graphics.html).

We have tests from 256M through 2G, as well as different UT2k4 patches including versions 3204, 3236, and 3323.

Apple is also helping Rob-ART, with some information on what map, resolution, and detail settings they used.

These results are STILL IN PROGRESS. Nothing is definite yet. Wer're trying to see if we can reproduce Apple's 212% figure for the G5 vs G4 17" iMac.

[stingerman]

17inch G5 iMac 1.6GHz 1024MB RAM (pair 512/s) set a maximum performance, no HD sleep (cook the chips!)...

CINEBENCH 2003-
rendering 1 cpu- 170
graphics C4D shading 204
OpenGL SW-L 529
OpenGL HW-l 928
Speedup 4.59

Make sure you use the Cinebench G5 beta, your single render score is too low, it must be an error...

[J.Allen]

is that good or bad?

i need a comparison to maybe the powermac g5
or an Athlon 64

[IBSNOWEDIN]

these are my cinebenchand xbench score on my 20"imac 1.8ghz 1024mb ram. Anyone tell me if these results are anygood? thanks

Cinebench
Rendering ( 1 CPU ) 250 CB-CPU
C4D Shading 237 CB-GFX
OpenGL SW-L 659 CB-GFX
OpenGL HW-L 965 CB-GFX
OpenGL Speedup 4.08x

xbench
Results 160.54
System Info
Xbench Version 1.1.3
System Version 10.3.5 (7P35)
Physical RAM 1024 MB
Model PowerMac8,1
Processor PowerPC G5 @ 1.80 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 1.15 GHz
Bus Frequency 600 MHz
Video Card GeForce FX 5200
Drive Type ST3160023AS
CPU Test 170.53
GCD Loop 101.98 3.98 Mops/sec
Floating Point Basic 281.77 1.02 Gflop/sec
AltiVec Basic 122.96 3.57 Gflop/sec
vecLib FFT 197.43 3.07 Gflop/sec
Floating Point Library 361.33 14.46 Mops/sec
Thread Test 104.75
Computation 69.19 934.09 Kops/sec, 4 threads
Lock Contention 215.49 2.70 Mlocks/sec, 4 threads
Memory Test 252.98
System 270.75
Allocate 669.44 436.68 Kalloc/sec
Fill 198.64 1581.14 MB/sec
Copy 219.68 1098.40 MB/sec
Stream 237.39
Copy 208.43 1523.59 MB/sec [G5]
Scale 210.90 1556.42 MB/sec [G5]
Add 266.65 1706.55 MB/sec [G5]
Triad 280.90 1716.31 MB/sec [G5]
Quartz Graphics Test 214.80
Line 202.78 5.16 Klines/sec [50% alpha]
Rectangle 187.56 13.19 Krects/sec [50% alpha]
Circle 208.31 4.80 Kcircles/sec [50% alpha]
Bezier 197.42 2.15 Kbeziers/sec [50% alpha]
Text 317.57 5.18 Kchars/sec
OpenGL Graphics Test 211.32
Spinning Squares 211.32 147.88 frames/sec
User Interface Test 257.61
Elements 257.61 82.86 refresh/sec
Disk Test 91.16
Sequential 82.46
Uncached Write 81.64 34.03 MB/sec [4K blocks]
Uncached Write 62.48 25.59 MB/sec [256K blocks]
Uncached Read 81.51 12.90 MB/sec [4K blocks]
Uncached Read 125.17 50.57 MB/sec [256K blocks]
Random 101.93
Uncached Write 97.66 1.47 MB/sec [4K blocks]
Uncached Write 102.19 23.05 MB/sec [256K blocks]
Uncached Read 95.64 0.63 MB/sec [4K blocks]
Uncached Read 114.14 23.49 MB/sec [256K blocks]

[Mac_Niarb]

Nope. There is no "Automatic" setting for Energy Saver on desktop machines--what would it be used for? No battery, thus, no energy conservation system. You control system sleep and display sleep, that's it.

So this score is a little odd.

Actually Bluefusion, the new iMac G5 does have an energy conservation system. Yep thats right. And yes you can change the processor speed from Automatic to Highest. You should knows your facts before making blanket statements like this one. The reason for this is that at a lower speed the computer runs quieter when the processor is at full speed it makes some noticeable noise but is about 28% faster. Read the 100th issue of MacAddict magazine if you want the details.

[MeanD3feat]

Sorry I couldn't wait. If a mod wants to edit this post to make it scroll feel free. :D

iMac 20"
2GB RAM CRUCIAL
250 GB HDD BTO
AIRPORT BTO
BLUETOOTH KEYBOARD AND MOUSE BTO
OS 10.3.7

Energy saving in Highest mode, all windows closed nothing running except the benchmarking tool which was started after a restart (for every test) to ensure there was nothing running in the background. Airport was disabled. Ram is a matched pair 128 bit bus size. (just adding info for sake of completeness, yes I have read the barefeats article)

XBENCH:

Results 159.15
System Info
Xbench Version 1.1.3
System Version 10.3.7 (7S215)
Physical RAM 2048 MB
Model PowerMac8,1
Processor PowerPC G5 @ 1.80 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 1.80 GHz
Bus Frequency 600 MHz
Video Card GeForce FX 5200
Drive Type WDC WD2500JD-40GBB2
CPU Test 161.51
GCD Loop 101.63 3.97 Mops/sec
Floating Point Basic 285.59 1.03 Gflop/sec
AltiVec Basic 123.03 3.57 Gflop/sec
vecLib FFT 150.14 2.33 Gflop/sec
Floating Point Library 353.42 14.15 Mops/sec
Thread Test 103.71
Computation 68.43 923.80 Kops/sec, 4 threads
Lock Contention 214.11 2.69 Mlocks/sec, 4 threads
Memory Test 274.45
System 297.86
Allocate 634.39 413.81 Kalloc/sec
Fill 271.53 2161.36 MB/sec
Copy 207.79 1038.95 MB/sec
Stream 254.45
Copy 226.67 1656.93 MB/sec [G5]
Scale 228.24 1684.37 MB/sec [G5]
Add 281.37 1800.79 MB/sec [G5]
Triad 296.46 1811.39 MB/sec [G5]
Quartz Graphics Test 202.04
Line 198.75 5.06 Klines/sec [50% alpha]
Rectangle 186.62 13.13 Krects/sec [50% alpha]
Circle 204.88 4.72 Kcircles/sec [50% alpha]
Bezier 195.01 2.12 Kbeziers/sec [50% alpha]
Text 229.93 3.75 Kchars/sec
OpenGL Graphics Test 178.96
Spinning Squares 178.96 125.23 frames/sec
User Interface Test 251.97
Elements 251.97 81.04 refresh/sec
Disk Test 99.99
Sequential 96.90
Uncached Write 85.22 35.52 MB/sec [4K blocks]
Uncached Write 66.17 27.09 MB/sec [256K blocks]
Uncached Read 152.07 24.07 MB/sec [4K blocks]
Uncached Read 127.31 51.44 MB/sec [256K blocks]
Random 103.27
Uncached Write 101.97 1.53 MB/sec [4K blocks]
Uncached Write 106.69 24.06 MB/sec [256K blocks]
Uncached Read 93.04 0.61 MB/sec [4K blocks]
Uncached Read 113.58 23.37 MB/sec [256K blocks]


CINEBENCH:

CINEBENCH 2003 v1
****************************************************

Tester :

Processor : iMac 1.8 G5 20" 2GB RAM 250 GB HDD
MHz : 1.8
Number of CPUs : 1
Operating System : OS X 10.3.7

Graphics Card : 5200 64
Resolution : 1680 1050
Color Depth : millions?

****************************************************

Rendering (Single CPU): 188 CB-CPU
Rendering (Multiple CPU): --- CB-CPU


Shading (CINEMA 4D) : 228 CB-GFX
Shading (OpenGL Software Lighting) : 567 CB-GFX
Shading (OpenGL Hardware Lighting) : 959 CB-GFX

OpenGL Speedup: 4.21

****************************************************


CINEBENCH BETA OPTIMISED FOR G5:

CINEBENCH 2003 v1 g5
****************************************************

Tester :

Processor : iMac 1.8 G5 20" 2GB RAM 250 GB HDD
MHz : 1.8
Number of CPUs : 1
Operating System : OS X 10.3.7

Graphics Card : 5200 64
Resolution : 1680 1050
Color Depth : millions?

****************************************************

Rendering (Single CPU): 247 CB-CPU
Rendering (Multiple CPU): --- CB-CPU


Shading (CINEMA 4D) : 233 CB-GFX
Shading (OpenGL Software Lighting) : 645 CB-GFX
Shading (OpenGL Hardware Lighting) : 949 CB-GFX

OpenGL Speedup: 4.06

****************************************************

COCOABENCH:

CocoaBench 1.2: 30.79
Number Crunching: 34.32
Image Processing: 43.14
File I/O: 11.38

[matticus008]

Those OpenGL scores are extremely low. You probably could not find any model PC made in the last four years that scored that low.

Try all the Dells made...ever...that don't have dedicated graphics cards (i.e. most of them). They absolutely suck at everything. You have to get their high end systems in order to get a graphics card. A homebuilt PC of course with a non -MX line card (or the ATi equivalent) certainly makes your statement true. I think even a Radeon 9000 can pull better OpenGL scores, but much of that is driver-dependent.

[mac-daddy]

It seems to me that the main downfall in the iMac G5 lies in the non-upgradable 64mb video card. Considering one of the main functions I want is video editing, I think I will wait until the next round of upgrades, hoping for a video card upgrade or at least an option to upgrade like every other model mac makes. It's frustrating that I just can't buy what I want. Does anyone out there have some experience with editing with the 64mb card. Am I overreacting or is it enough for this task?

[Daveway]

My iMac has the same Cinebench score as my P4 2.4ghz laptop. It was a bit disappointing but its the Mac experience that counts.

[Gregory]

It's quite clear that the iMac G5 is on a par with a DP G4 1.42, both the cinebench and xbench results confirm this. This is a fantastic result. A 20" iMac with a 250GB HDD and maxed out ram will be a formidable machine and is likely to outrun any G4 system.

Yeah! The iMac G5 1.8 is fast, I tested some Final Cut Pro Edits and Renders with it and is the same speed as a Dual G4 1.42 , with the iMac you even have some advantages with other apps that only use single procs, its even faster when using it for DVD Studio Pro 3 interface and creating the DVD Layout. . .
Cool Stuff. . . Not a bad choice for a iMac G5, I'm thinking about buying one for a extra DVD creations station and maybe replacing my Dual G4 1.25 with one.

[Gregory]

My iMac has the same Cinebench score as my P4 2.4ghz laptop. It was a bit disappointing but its the Mac experience that counts.
Just remember that Cinebench is still a Beta Version for the G5, and is not as optimized as the windows version. . . Even with the same GHZ and Graphic card the Mac comes in slower, but when I benchmarked a GAME the mac was the same speed as the PC. How Strange it was. . . Even when rendering in Lightwave 3D 7.5 it was a little faster, but only 7 seconds, but the interface was faster on the PC, about a third faster in the Open GL. I called Newtek, and the PC Version is more optimized than the Mac Version. Most Programs have better optimiztion than the Mac Version, It takes some time for these changes to take place on the MAC side, but when it does it is worth the wait. . . Just look at what FCP can do with over 7 Layers of Uncompress SD Video in real time with no accel. video card, just a scsi raid, FCP & a G5 computer is all that is needed. To have fun editing in record time !!!

[Gregory]

It seems to me that the main downfall in the iMac G5 lies in the non-upgradable 64mb video card. Considering one of the main functions I want is video editing, I think I will wait until the next round of upgrades, hoping for a video card upgrade or at least an option to upgrade like every other model mac makes. It's frustrating that I just can't buy what I want. Does anyone out there have some experience with editing with the 64mb card. Am I overreacting or is it enough for this task?
You are over reacting big time. FCP doesn't use the graphic card that much, the only apps that use the graphic card are games, and 3D apps with Open GL thats it, you will be fine with editing video. I am a video editor professional and have used FCP since Version 1.0. Even if you use Dual Monitors you only need 32mbs per screen, a total of 64mbs for Dual screens in Mac OSX. If you use OS 9 you only need 16mbs per screen. The iMac G5 1.8 is just as fast as a Dual G4 1.42 in FCP, the G5 does very well in performing editing tasks even a single proc. is faster in editing, you will notice that the interface will perform much faster than a G4 and even a Dual G4, the interface only uses one proc. rendering use two proc. but the 1.8 G5 does better than one might think against a Dual G4.

[Santaduck]

Yikes is this thread still going?

There are already lots of results both here and at barefeats (http://www.barefeats.com)... but as before, remember on the iMac G5 to:

* set energy saving's CPU to "highest"
* quit all other apps
* don't touch keyboard or move mouse during the benchmark
* you can even hide the dock so dock computation (bouncing, magnification) doesn't affect results.

That said, the darn thing is too noisy at "highest", so many owners should be doing 'real world' benching at CPU "automatic". I myself will put it to 'highest' only when I'm doing something computationally intensive, or gaming, but have it on automatic almost all of the time. . . unfortunately bench results on 'automatic' are a bit inconsistent, at least for some benchmarks.

[yanges]

here are the comparisons @ barefeats:

http://www.barefeats.com/imacg5.html

http://www.barefeats.com/imacg5b.html

http://www.barefeats.com/imacg5c.html

http://www.barefeats.com/imacg5d.html

overall i would say the iMac G5 stacks up pretty well!! :cool:

[ i apologize if any of these were already posted ]