[Tenacious Dyl]

Hey Everyone,
At the Apple Store in San Francisco... they have 2 Mac Pros set up already. Thought I would post what I could about it for you all!

First, I checked the profiler. The one I'm on is identical to the other:
2x Dual-Core Xeon at 2.66
OS 10.4.7
1 GB Of DDR2 FB-DIMM*
250 GB SATA HD
Sony DVD RW (DW-D150A)
NVidia GeForce 7300GT (30" Hooked Up To It)
No Airport, No Bluetooth

I downloaded xBench, ran it three times... the middle run is below:

Results 49.44
System Info
Xbench Version 1.2
System Version 10.4.7 (8K1079)
Physical RAM 1024 MB
Model MacPro1,1
Drive Type ST3250824AS P
CPU Test 136.85
GCD Loop 316.05 16.66 Mops/sec
Floating Point Basic 152.08 3.61 Gflop/sec
vecLib FFT 105.98 3.50 Gflop/sec
Floating Point Library 99.46 17.32 Mops/sec
Thread Test 526.60
Computation 484.27 9.81 Mops/sec, 4 threads
Lock Contention 577.03 24.82 Mlocks/sec, 4 threads
Memory Test 113.85
System 120.96
Allocate 152.06 558.40 Kalloc/sec
Fill 108.11 5256.55 MB/sec
Copy 111.43 2301.47 MB/sec
Stream 107.53
Copy 102.66 2120.34 MB/sec
Scale 104.14 2151.55 MB/sec
Add 111.28 2370.50 MB/sec
Triad 112.74 2411.82 MB/sec
Quartz Graphics Test 88.93
Line 87.58 5.83 Klines/sec [50% alpha]
Rectangle 86.24 25.75 Krects/sec [50% alpha]
Circle 86.45 7.05 Kcircles/sec [50% alpha]
Bezier 113.20 2.86 Kbeziers/sec [50% alpha]
Text 78.08 4.88 Kchars/sec
OpenGL Graphics Test 185.03
Spinning Squares 185.03 234.72 frames/sec
User Interface Test 12.12
Elements 12.12 55.64 refresh/sec
Disk Test 40.87
Sequential 92.58
Uncached Write 79.71 48.94 MB/sec [4K blocks]
Uncached Write 97.34 55.07 MB/sec [256K blocks]
Uncached Read 88.37 25.86 MB/sec [4K blocks]
Uncached Read 110.24 55.41 MB/sec [256K blocks]
Random 26.22
Uncached Write 8.17 0.87 MB/sec [4K blocks]
Uncached Write 90.39 28.94 MB/sec [256K blocks]
Uncached Read 87.88 0.62 MB/sec [4K blocks]
Uncached Read 129.65 24.06 MB/sec [256K blocks]

*The Ram Set-Up is something new, so I thought I would give more info here too. In this computer, the 1GB is from two 512 sticks. The profiler shows that one is "DIMM Riser A/DIMM 1" It is of the type "DDR2 FB-DIMM" Speed of "667 MHz" with no errors. The second is the same, also on Riser A, only its labled "Dimm 2". I assume this means with 2 risers, and 4 channels, that the ram is on 1 and 2 of the top four slots, putting it in the samd channel.

The computer seems fairly fast, but given the xBench, not as fast as I think people assumed. Perhaps with more HD (or RAID) and more RAM and better graphics... Still, it ties my 17" MacBook Pro.

Hope this helped the curious. The store here has a card out, saying that they have them in stock here, to take hsme today... so they at least did that right!

[chefpastry]

Personally, I don't trust xBench all too much.

[Lateralus]

Xbench is so flaky on benchmarking that I think most people dismiss it now. Something as simple as a reboot or a logout has been known to affect scores by 10-30%.

[harrisjamieh]

'It ties my 17 MBP'....

Do you mean it ties with your MBP, as in its the same...........

If so, surely that should tell you never to trust xbench

[mduell]

Xbench is poo.

Originally Posted by Tenacious Dyl

*The Ram Set-Up is something new, so I thought I would give more info here too. In this computer, the 1GB is from two 512 sticks. The profiler shows that one is "DIMM Riser A/DIMM 1" It is of the type "DDR2 FB-DIMM" Speed of "667 MHz" with no errors. The second is the same, also on Riser A, only its labled "Dimm 2". I assume this means with 2 risers, and 4 channels, that the ram is on 1 and 2 of the top four slots, putting it in the samd channel.

On a given riser, slots 1 and 3 are on the same channel, as are 2 and 4. With memory in slots 1 and 2 (or 0 and 1, depending on how you want to number them), you're getting dual channel.

[masugu]

Well, if Mark thinks XBENCH is Poo, that's good enough for me. THanks Mark. You know your, er...Poo!

I am looking forward to some thorough reviews before I buy....I may also consider waiting until Leopard launches with the Mac Pro. I bought a MacBook and I should wait a tad.

[mduell]

Originally Posted by masugu

Well, if Mark thinks XBENCH is Poo, that's good enough for me. THanks Mark. You know your, er...Poo!

I am looking forward to some thorough reviews before I buy....I may also consider waiting until Leopard launches with the Mac Pro. I bought a MacBook and I should wait a tad.

Here's an example of what utter nonsense Xbench is:



Yes, that's right, and iMac with a nearly identical disk to a PMG5 and Mac Pro gets 2-4 times the score. Oh, and a MBP, with it's little 2.5" disk, gets twice the score of a PMG5.

[mac128k-1984]

Originally Posted by Tenacious Dyl

I downloaded xBench, ran it three times... the middle run is below:
Results 49.44

That cannot be right. My 1.6Ghz intel mini produced an overall result of 53.

[ender78]

The MBP HD is SATA after all, could this account for the performance difference? Then again, why the heck is the iMac the fastest on the HD test.

[chefpastry]

All the computers in that graph uses SATA drives.

[jamil5454]

Run cinebench and possibly some photoshop tests!

[Tenacious Dyl]

I never said anything for xBench's perfection, or lack. It was just a quick thing to run while in the store... I only had a few minutes to spend there. I posted it for those who were curious. I know that my MBP does not = the new Mac Pro... however, even if xBench is a bit flaky, my MBP shouldn't come close to the Mac Pro, and the playing around on it that I did do just wasn't that impressive.

[all2ofme]

I'd put it down to the lack of RAM in the Mac Pro that it didn't feel impressive. I played with one at the Regent St store in London and felt the same. It too had only 1GB.