Canadian Mac Forums at ehMac banner

Disappointed with my G5 Quad (poor Xbench)???

7K views 18 replies 10 participants last post by  uselessobserver 
#1 ·
So I got my Quad A couple 2 weeks ago and I have been trying to get it to perform to what my mac loving friends believe it should perform to...

I am getting only 138% And the new Duo core imac scores 150% (what is going on)

I have the OS and all the apps running off a raptor (WD) at 10,000rpm...
I have the geforce 7800,
4.5 Gigabyte of ram (kingston) not the Value Ram
quad 2.5
it is set to highest performance in the energy setup
I have done all the disk utility permissions and repairs
What else can I do...

I was told that it should be running at 200% if not more (this may be wrong but a mac specialist at my local store told me this)

So If anyone has any ideas of what is wrong let me know... I want it to be a monster that it should be.

I am doing a apple hardware test right now to see if the ram or anything else is wrong so I will post that when it finishes...

thanks for the help.....

haven2020
 
See less See more
#2 ·
xBench is not a very good benchmarking tool. It is not consistent or MutliCPU friendly. Most people I have seen use Cinebench http://www.benchmarkhq.ru/english.html?/be_video.html or a suite of photoshop tests for getting accurate meaningful benchmarks.

I am downloading cenebench now and will post my benchmark for my duocore 1.83 Ghz imac 17" 1.5 GB Ram system.
 
#3 · (Edited)
xBench doesn't offer much in the way of support for multi-core or multi-CPU machines; the 138 score is a representation of the result for 1 of your 4 cores (and the 150 score for newer Intel dualcores is for 1 of the 2 cores).

Benchmarks are only tools, and if you are not clear as to what they measure, you can be disappointed with the results. Real performance is not the same as benchmark scores.

Try timed tests with multithreaded applications. Oh, and the more RAM you add, the poorer the xBench result; try yanking 2.5Gigs out and see how that helps the "score". It won't help performance (depending on the task, it might harm it) but, hey, if it's all about the number then you will be happier if you cripple the memory.

But, any computer does more work with more memory, or at least adequate memory. xBench is just a number, and there are a lot of xBenchers who deliberately run a certain config for the score, not for real-world performance. It's "Bench Racing", not road racing.
 
#5 ·
Here is my Cinebench results.

CINEBENCH 9.5
****************************************************

Tester : john Main

Processor : imac 17"
MHz : 1.83 Ghz
Number of CPUs : 2
Operating System : 10.4.5

Graphics Card : ati x1600
Resolution : <1440X900>
Color Depth : <32 Bit>

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

Rendering (Single CPU): 283 CB-CPU
Rendering (Multiple CPU): 518 CB-CPU

Multiprocessor Speedup: 1.83

Shading (CINEMA 4D) : 330 CB-GFX
Shading (OpenGL Software Lighting) : 892 CB-GFX
Shading (OpenGL Hardware Lighting) : 1578 CB-GFX

OpenGL Speedup: 4.78

****************************************************
 
#6 ·
Here are my geek bench scores as well

System Information
OS: Version 10.4.5 (Build 8G1454)
Model: iMac4,1
Motherboard: iMac4,1
CPU: Intel Core Duo
CPU ID: 7, 4
CPU Count (Physical): 2
CPU Count (Logical): 2
CPU Frequency: 1830 MHz
Bus Frequency: 664 MHz
Memory: 1536 MB

CPU Integer Performance
Emulate 6502 168 (1 thread, 298.7 megahertz)
Emulate 6502 322 (4 threads, 589.7 megahertz)
Blowfish 322 (1 thread, 443.5 megabytes/sec)
Blowfish 570 (4 threads, 865.5 megabytes/sec)
bzip2 Compress 127 (1 thread, 22.92 megabytes/sec)
bzip2 Compress 241 (4 threads, 44.92 megabytes/sec)
bzip2 Decompress 128 (1 thread, 53.24 megabytes/sec)
bzip2 Decompress 242 (4 threads, 102.7 megabytes/sec)

CPU Floating Point Performance
Mandelbrot 136 (1 thread, 915.4 megaflops)
Mandelbrot 257 (4 threads, 1.796 gigaflops)

Memory Performance
Latency 597 (1 thread, 17.53 nanoseconds/load)
Read Sequential 363 (1 thread, 2.637 gigabytes/sec)
Write Sequential 160 (1 thread, 942.8 megabytes/sec)
Stdlib Allocate 130 (1 thread, 101.9 kiloallocs/sec)
Stdlib Allocate 149 (4 threads, 117.5 kiloallocs/sec)
Stdlib Write 137 (1 thread, 2.166 gigabytes/sec)
Stdlib Copy 166 (1 thread, 1.243 gigabytes/sec)

Stream Performance
Stream Copy 120 (1 thread, 1.638 gigabytes/sec)
Stream Scale 119 (1 thread, 1.631 gigabytes/sec)
Stream Add 182 (1 thread, 2.521 gigabytes/sec)
Stream Triad 181 (1 thread, 2.545 gigabytes/sec)
 
#7 · (Edited)
My MacBook Pro 1.83 Ghz:


System Information
OS: Version 10.4.5 (Build 8H1455)
Model: MacBookPro1,1
Motherboard: MacBookPro1,1
CPU: Intel Core Duo
CPU ID: 7, 4
CPU Count (Physical): 2
CPU Count (Logical): 2
CPU Frequency: 1830 MHz
Bus Frequency: 664 MHz
Memory: 2048 MB

CPU Integer Performance
Emulate 6502 162 (1 thread, 289.4 megahertz)
Emulate 6502 321 (4 threads, 587.7 megahertz)
Blowfish 277 (1 thread, 381 megabytes/sec)
Blowfish 529 (4 threads, 801.9 megabytes/sec)
bzip2 Compress 119 (1 thread, 21.45 megabytes/sec)
bzip2 Compress 239 (4 threads, 44.61 megabytes/sec)
bzip2 Decompress 123 (1 thread, 51.46 megabytes/sec)
bzip2 Decompress 244 (4 threads, 103.4 megabytes/sec)

CPU Floating Point Performance
Mandelbrot 133 (1 thread, 899.6 megaflops)
Mandelbrot 254 (4 threads, 1.776 gigaflops)

Memory Performance
Latency 488 (1 thread, 21.44 nanoseconds/load)
Read Sequential 303 (1 thread, 2.201 gigabytes/sec)
Write Sequential 139 (1 thread, 821.3 megabytes/sec)
Stdlib Allocate 124 (1 thread, 96.81 kiloallocs/sec)
Stdlib Allocate 142 (4 threads, 112 kiloallocs/sec)
Stdlib Write 132 (1 thread, 2.088 gigabytes/sec)
Stdlib Copy 161 (1 thread, 1.204 gigabytes/sec)

Stream Performance
Stream Copy 116 (1 thread, 1.583 gigabytes/sec)
Stream Scale 116 (1 thread, 1.578 gigabytes/sec)
Stream Add 182 (1 thread, 2.523 gigabytes/sec)
Stream Triad 177 (1 thread, 2.497 gigabytes/sec)
 
#8 ·
Thanks for the info everyone

I will check out those benchmarks and post them for everyone as well... I was just worried that something was malfunctioning in the Quad.... Thanks again to everyone for the help and the schooling on how benchmarks work...

haven2020
 
#9 ·
Here's what I got

CINEBENCH 9.5
****************************************************

Tester :

Processor : Powermac G5 Quad
MHz : 2.5
Number of CPUs : 4
Operating System : 10.4.5

Graphics Card : geforce 7800
Resolution : <1152x870>
Color Depth : <32 bit>

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

Rendering (Single CPU): 385 CB-CPU
Rendering (Multiple CPU): 1189 CB-CPU

Multiprocessor Speedup: 3.09

Shading (CINEMA 4D) : 387 CB-GFX
Shading (OpenGL Software Lighting) : 1309 CB-GFX
Shading (OpenGL Hardware Lighting) : 2705 CB-GFX

OpenGL Speedup: 6.98

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

Geekbench Information
Version: Geekbench Preview 2 (r73)
Compiler: GCC 4.0.1 (Apple Computer, Inc. build 5250)

System Information
OS: Version 10.4.5 (Build 8H14)
Model: PowerMac11,2
Motherboard: PowerMac11,2
CPU: PowerPC G5 (970)
CPU ID: 18, 100
CPU Count (Physical): 4
CPU Count (Logical): 4
CPU Frequency: 2500 MHz
Bus Frequency: 1250 MHz
Memory: 4608 MB

CPU Integer Performance
Emulate 6502 201 (1 thread, 358.9 megahertz)
Emulate 6502 754 (4 threads, 1.378 gigahertz)
Blowfish 180 (1 thread, 247.8 megabytes/sec)
Blowfish 613 (4 threads, 929.7 megabytes/sec)
bzip2 Compress 169 (1 thread, 30.35 megabytes/sec)
bzip2 Compress 608 (4 threads, 113.2 megabytes/sec)
bzip2 Decompress 178 (1 thread, 74.1 megabytes/sec)
bzip2 Decompress 674 (4 threads, 285.7 megabytes/sec)

CPU Floating Point Performance
Mandelbrot 165 (1 thread, 1.11 gigaflops)
Mandelbrot 625 (4 threads, 4.369 gigaflops)
 
#12 ·
I am not sure why you would need Bench Marks to verify this. I would assume that (In Darth Vader Voice) YOU WOULD FEEL THE POWER OF THE QUAD!

I know I felt a huge difference going from my G4 933 to my Intel CoreDuo. I imagine that doing any Media Encoding with a Quad would eat through it like my dog eats table scraps.
 
G
#13 ·
It will really shine doing big rendering/number crunching/filtering sort of stuff. :) For day to day OS operation you may not notice all that much. Think I'll have to track down a quad for my studio one of these years .. imagine all the altiverb instances I could run LOL.
 
#14 ·
Jmain said:
I am not sure why you would need Bench Marks to verify this. I would assume that (In Darth Vader Voice) YOU WOULD FEEL THE POWER OF THE QUAD!

I know I felt a huge difference going from my G4 933 to my Intel CoreDuo. I imagine that doing any Media Encoding with a Quad would eat through it like my dog eats table scraps.
Thing is, again, it depends on what you run on the Quad to compare. If you run a non-SMP game like UT2004 (hey, I'm a gamin' guy) you'll be disappointed, because it won't use the second processor (and the two cores in that processor) for much except the sound and maybe some OS tasks. If however you take out an SMP aware game like Quake III, it'll wipe the pants off any other system, guaranteed. Now that Intel has gone multi-cores, we should see more and more applications which support multi-core and multi-processor computers since most applications that are cross-platform are written for Windows machines first and Mac machines second. If the SMP support is there on the Windows side it makes it that much easier to port that over to the Mac side.

If you can find an app that is Altivec optimized and optimized for SMP the Quad will fly by any other machine.

The Quad machine is a hard machine to keep busy unless your doing video rendering or some other intensive things.
 
#15 ·
The Quad machine is a hard machine to keep busy unless your doing video rendering or some other intensive things.
We've found that as well - the 2.5 duals do well more most clients but when you really need crunch the Quads are delightful - we've traded a number of video editors up and really the gain for the money is superb.
If you figure it's 45-70% faster than a 2.7 dual then the gain over say a 2.0 dual is enormous.
 
#16 ·
geekbench...

i googled my way here looking for someone who could help me find a replacement apple hardware disk, as mine is scratched to hell and wont even load anymore, and found this post, so i tried geekbench, but after running it it come up saying that ive got a 32-bit ppc instead of 64bit? did i get ripped off or is geekbench wrong?.
 
#17 ·
i googled my way here looking for someone who could help me find a replacement apple hardware disk, as mine is scratched to hell and wont even load anymore, and found this post, so i tried geekbench, but after running it it come up saying that ive got a 32-bit ppc instead of 64bit? did i get ripped off or is geekbench wrong?.
What kind of Mac do you have?

Trev
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top