[eggman]

Primate Labs has posted their benchmarks for the new eight-core Mac Pros... and the results have me this close to canceling my order for a 3.0 system, stepping down to the 2.8 and saving myself $800.

Here are the results.

Of course I don't expect the percentage difference in scores to map to the percentage difference in price... but still... if looking at the basic system cost, there's greater than a 28% premium on what works out to be less than a 2% differential in performance.

I'm wondering what other folks are thinking, based on these numbers.

[Big Mac]

I haven't been focusing too much on the new MPs (or the old MPs, for that matter), but it seems to me that the entry level 2.8 model is the sweet spot in the lineup.

[ninahagen]

Originally Posted by eggman 

Primate Labs has posted their benchmarks for the new eight-core Mac Pros... and the results have me this close to canceling my order for a 3.0 system, stepping down to the 2.8 and saving myself $800.

Here are the results.

Of course I don't expect the percentage difference in scores to map to the percentage difference in price... but still... if looking at the basic system cost, there's greater than a 28% premium on what works out to be less than a 2% differential in performance.

I'm wondering what other folks are thinking, based on these numbers.

If money is an issue, it seems like you would get more value with faster drives, more RAM and/or a RAID setup.

How much extra money is your 3.0 costing you?

[eggman]

Originally Posted by ninahagen 

If money is an issue, it seems like you would get more value with faster drives, more RAM and/or a RAID setup.

That's what I'm thinking.

How much extra money is your 3.0 costing you?

As mentioned, there's an $800 differential at the Apple store between the 2.8 and the 3.0... and another $800 to go to the 3.2. And there's a more substantial increase in performance between the 3.2 and the 3.0... overall, about 5%. So it seems like if money is no object, get the 3.2 - and if money makes any difference at all, get a 2.8... but it's hard for me to see the circumstance in which the 3.0 makes much a lot of sense, at least for me.

That's assuming these scores are reliable... and at all reflective of real world performance.

Still, I wonder why there's such a small increment between the 2.8 and the 3.0 compared with the 3.0 and 3.2. I understand that it's not going to scale in a linear fashion, but still.

On some of these scores the 2.8 actually outperformed the 3.0, and that's a head-scratcher.

[MallyMal]

Here's the MacWorld tests if anyone is interested. Macworld | From the Lab: More Mac Pro benchmark testing

[eggman]

Originally Posted by MallyMal 

Here's the MacWorld tests if anyone is interested. Macworld | From the Lab: More Mac Pro benchmark testing

This makes the 3.0 look a bit more of an improvement over the 2.8, looking more like a 4% benefit... and that's starting to get to the threshold of being a more interesting option. Still, I expect that putting the money into RAM or faster drives would still buy you better price/performance.

Too bad the 3.2 GHz unit was DOA and not included in their testing.

[Tesselator]

Nice! The new Mac Pro MB and FSB speeds make a nice difference over the 1.0 and 1.1 MB

Here's my 8-core Mac Pro v1.1 (x5355): Geekbench Result Browser: Mac Pro

[Tesselator]

Or... maybe it doesn't... Dang - Geekbench gives wildly different
results each time it's run.

The next time I ran it I got: Geekbench Result Browser: Mac Pro
And after a quit and rerun: Geekbench Result Browser: Mac Pro
And another entry is here: Geekbench Result Browser: Mac Pro

Really makes me think Geekbench is total crap and not to be trusted.

[SierraDragon]

The question gets to apps used and how much money is available. IMO with limited funds for 95+% of workflows the $800 is better spent on faster/bigger/buffered hard drives as well as additional RAM. However of course best is doing it all: faster CPU, faster/bigger/buffered hard drives as well as additional RAM.

IMO we have a ways (years maybe) to go before apps really take full advantage of the spectacular power available in Mac Pro CPUs, so marginal improvements in CPU speed are just that, marginal. Even with essentially unlimited client funds I specify the 2.8 GHz box, +8 GB OWC RAM to start, and put money into hard drives and RAID.

Adding the 8800 GT card can be a very big help for apps like Aperture, little help for Adobe apps.

-Allen Wicks

[Big Mac]

Originally Posted by Tesselator 

Or... maybe it doesn't... Dang - Geekbench gives wildly different
results each time it's run.

The next time I ran it I got: Geekbench Result Browser: Mac Pro
And after a quit and rerun: Geekbench Result Browser: Mac Pro
And another entry is here: Geekbench Result Browser: Mac Pro

Really makes me think Geekbench is total crap and not to be trusted.

I don't much of anything about Geekbench, but it's impressive to me that you have to go down to the third page of highest scores to find a non Mac Pro.