[PeteWK]

Hello everyone. I just upgraded my 20 inch iMac G5 2.0 from the stock 512mb to the stated max of 2048. I used a matched set of Crucial chips costing exactly 238.14 shipped. The overall XBench score went from 174.72 to 183.73. The specific memory scores went as follows:

THE FOLLOWING LIST OF NUMBERS SHOWS THE CHANGE FROM 512 TO 2048 MB OF RAM.



Memory Text 229.36 to 302.55
System 301.04 to 366.10
Allocate 660.29 to 682.15
Fill 422.74 to 456.10
Copy 164.33 to 220.46
Stream 185.25 to 257.80
Copy 161.07 to 225.05
Scale 160.81 to 233.10
Add 211.18 to 287.56
Triad 225.70 to 302.57



The 2 gig complete XBench test is as follows:

Results 183.73
System Info
Xbench Version 1.1.3
System Version 10.4.1 (8B15)
Physical RAM 2048 MB
Model PowerMac8,2
Processor PowerPC G5 @ 2.00 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 2.00 GHz
Bus Frequency 667 MHz
Video Card ATY,RV351
Drive Type WDC WD2500JD-40HBC0
CPU Test 187.28
GCD Loop 110.52 4.32 Mops/sec
Floating Point Basic 317.40 1.15 Gflop/sec
AltiVec Basic 136.46 3.96 Gflop/sec
vecLib FFT 219.71 3.41 Gflop/sec
Floating Point Library 381.62 15.28 Mops/sec
Thread Test 109.23
Computation 74.84 1.01 Mops/sec, 4 threads
Lock Contention 202.04 2.54 Mlocks/sec, 4 threads
Memory Test 302.55
System 366.10
Allocate 682.15 444.97 Kalloc/sec
Fill 456.10 3630.55 MB/sec
Copy 220.46 1102.31 MB/sec
Stream 257.80
Copy 225.05 1645.15 MB/sec [G5]
Scale 233.10 1720.25 MB/sec [G5]
Add 287.56 1840.41 MB/sec [G5]
Triad 302.57 1848.71 MB/sec [G5]
Quartz Graphics Test 211.92
Line 191.85 4.88 Klines/sec [50% alpha]
Rectangle 192.06 13.51 Krects/sec [50% alpha]
Circle 210.03 4.84 Kcircles/sec [50% alpha]
Bezier 220.05 2.39 Kbeziers/sec [50% alpha]
Text 258.43 4.21 Kchars/sec
OpenGL Graphics Test 210.35
Spinning Squares 210.35 147.20 frames/sec
User Interface Test 310.23
Elements 310.23 99.78 refresh/sec
Disk Test 131.51
Sequential 159.03
Uncached Write 159.32 66.41 MB/sec [4K blocks]
Uncached Write 138.29 56.63 MB/sec [256K blocks]
Uncached Read 224.98 35.62 MB/sec [4K blocks]
Uncached Read 138.90 56.12 MB/sec [256K blocks]
Random 112.11
Uncached Write 92.97 1.39 MB/sec [4K blocks]
Uncached Write 146.01 32.93 MB/sec [256K blocks]
Uncached Read 100.73 0.66 MB/sec [4K blocks]
Uncached Read 122.76 25.26 MB/sec [256K blocks]


I hope this helps anyone thinking about upgrading their iMac G5. If you have any other questions let me know. Regards,

PeteWK

[MaxPower2k3]

I used my new 20" iMac for about a day with the stock 512mb RAM, and, while individual apps ran great, it did bog down a couple of times with a lot of stuff open. popped in an extra 1GB of RAM that came the next day, and it's running fantastic, even with lots of programs open.

Glad to see that the 20" iMac gets over half of the performance of a Dual 2.7GHz G5 (in overall score). And the CPU score of the iMac is over 70% of the score of the PowerMac. Makes me a little bit happier with my decision to get the iMac (besides fiscal reasons)

[Tomacorno]

I added a 1gig stick to my 17" iMac. When I went to see all that idle memory in my system monitor widget I noticed that the CPU was bogging down like crazy. I had already noticed it running hot, that is, the fans were coming on full and staying on. Turned out to be an issue with Virex not working properly with Tiger. Uninstalled Virex and now this machine just runs great no matter how many things I have open (so far).

[iBorg]

Originally Posted by PeteWK

The overall XBench score went from 174.72 to 183.73.

I haven't used XBench, but am I undestanding you to say that after quadrupling your memory, you only gained 5% in speed? I have 1 GB RAM using 512MB x 2 in a matched set, and my 20" iMac G5 2 GHz seems VERY fast, and I've wondered at times if I should have just taken the stock 512 RAM and bought 2 x 1 GB RAM, but after seeing your figures is this a waste of money?



iBorg

[Podolsky]

If you had run XBench 4 times in a row with your stock memory or with your upagrade you would have gotten a greater range of values than you report here for before and after. XBench is notoriously sensitive to what else is going on on your system. However, I am not saying that there isn't a slight difference by adding RAM. But a better way to detect that would be to reboot and run XBench as the first and only app and take the mean of say 5 or better 10 runs of XBench and than do the same with the upgraded RAM. That would give you a more robust test than what you did but by no means a definitive test. The developers of XBench may not want to hear that that is how their "product" should be used but that has been my experience.

[PeteWK]

Originally Posted by iBorg

I haven't used XBench, but am I undestanding you to say that after quadrupling your memory, you only gained 5% in speed? I have 1 GB RAM using 512MB x 2 in a matched set, and my 20" iMac G5 2 GHz seems VERY fast, and I've wondered at times if I should have just taken the stock 512 RAM and bought 2 x 1 GB RAM, but after seeing your figures is this a waste of money?



iBorg

You also need to know that XBench tests the entire system of which memory is only one part. The test then puts the average of all those values into one number. That's why I gave the memory results first and as a seperate section. Those values show increases as follows:

32%
22%
3%
8%
34%
39%
40%
45%
36%
34%

An example of this idea would be that while memory doesn't make the hard drive faster, it swaps out the information in greater amounts. Etc.

The speed increase of just adding the memory has been huge. My iMac comes on in half the time, pages open twice as fast and I can run several programs without lagging the computer. If you're thinking about the upgrade I would give it two thumbs up without hesitation.

PeteWK

[Hodog16]

Thanks for the info. I got my 1 GB x 2 RAM from Crucial also but installed it before I ever turned on my iMac for the 1st time .