Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

MacPro Mid 2010 First Impressions?

N

Nautigar

macrumors regular
Original poster
Oct 4, 2007
121
7
Hello,

as the first few lucky ones here get their new MacPros, I thougt it would be interesting to hear about their first impressions and maybe even see some pictures or benchmark results.

Those of you who owned a 2009 MacPro before, can you tell differences other than the processor tray and graphics card?

Best wishes,

Matthias.
 
I am also very interested in hearing your first impressions of the new mac pros (especially the hexacore which I plan to order in the next 10 days or so)
 
Nautigar said:
Hello,

as the first few lucky ones here get their new MacPros, I thougt it would be interesting to hear about their first impressions and maybe even see some pictures or benchmark results.

Those of you who owned a 2009 MacPro before, can you tell differences other than the processor tray and graphics card?

Best wishes,

Matthias.
Click to expand...

There are some benchmarks of the hexcore and the base 2.4 quad here:
https://forums.macrumors.com/showthread.php?p=10879619&posted=1#post10879619

On the processor benchmarks the hexcore beats the 2009 2.93 8-core as well as the 2010 2.4 8-core
 
drml said:
There are some benchmarks of the hexcore and the base 2.4 quad here:
...
On the processor benchmarks the hexcore beats the 2009 2.93 8-core as well as the 2010 2.4 8-core
Click to expand...

Yes, you get a 9,78% higher Geekbench score by spending 12,50% more money for the SP MP 3.33 GHz with 6 GB RAM.
I compared the 64-bit results and the prices in the german Apple Online Store.
 
Fiete5401 said:
Yes, you get a 9,78% higher Geekbench score by spending 12,50% more money for the SP MP 3.33 GHz with 6 GB RAM.
I compared the 64-bit results and the prices in the german Apple Online Store.
Click to expand...

What you say is true, but somewhat misleading. The Geekbench score is multithreaded. If you focus on the single-threaded results you will find:

Integer performance:
6-core 34.47% faster

Floating point performance:
6-core 35.52% faster

Memory performance:
6-core 34.09% faster

Stream performance:
6-core 39.62% faster

The reality is that most people will see much better performance from the 6-core than the 2.4 8-core. The vast majority of tasks that users perform will not fully tax the 6 or 8 cores. Even under full load (like Handbrake) the 6-core will be faster due to the higher clock speed.
 
Nautigar said:
Hello,

as the first few lucky ones here get their new MacPros, I thougt it would be interesting to hear about their first impressions and maybe even see some pictures or benchmark results.

Those of you who owned a 2009 MacPro before, can you tell differences other than the processor tray and graphics card?

Best wishes,

Matthias.
Click to expand...

Does the Kernel run in 64 bit mode?
 
Can anyone who has already received their Mac Pro that didn't BTO an SSD drive tell us if 2.5" hard drive adapters are included?
 
drml said:
What you say is true, but somewhat misleading. The Geekbench score is multithreaded. If you focus on the single-threaded results you will find:

[...]

The reality is that most people will see much better performance from the 6-core than the 2.4 8-core. The vast majority of tasks that users perform will not fully tax the 6 or 8 cores. Even under full load (like Handbrake) the 6-core will be faster due to the higher clock speed.
Click to expand...

Completely agree.

The reasons why I got the octo were specific: my workload nowadays is more about lots of simultaneous processes and VMs, not raw powerful computing. So the 8-core seemed like a good fit due to:

- Greater RAM capacity (and cheaper to expand)
- More physical cores benefits virtualization
- Cheaper with the potential to try upgrading the CPUs later
- I didn't pay for it, so I was more willing to risk a different approach than the 6-core


So far I'm loving it. But I still have tons of work left to get it set up. :D
 
drml said:
What you say is true, but somewhat misleading. The Geekbench score is multithreaded. If you focus on the single-threaded results you will find:
...
The reality is that most people will see much better performance from the 6-core than the 2.4 8-core. The vast majority of tasks that users perform will not fully tax the 6 or 8 cores. Even under full load (like Handbrake) the 6-core will be faster due to the higher clock speed.
Click to expand...

I completely agree with your conclusion. How did you do your calculations?
Did your extract the single threaded results from the integer and floating point sections and compared these numbers? I'm currently putting an excel sheet together by copy&paste.
 
Fiete5401 said:
I completely agree with your conclusion. How did you do your calculations?
Did your extract the single threaded results from the integer and floating point sections and compared these numbers? I'm currently putting an excel sheet together by copy&paste.
Click to expand...

Yes, I just pulled the single threaded results from all four sections from two geekbench runs by other forum members (my 6-core is still in Honk Kong :(). Then, I put the numbers into a spreadsheet and calculated the averages by section.

Here are the 3.33 6-core results:
http://browse.geekbench.ca/geekbench2/view/280534

and here are the 2.4 8-core results:
http://browse.geekbench.ca/geekbench2/view/280564
 
drml said:
Here are the 3.33 6-core results:
http://browse.geekbench.ca/geekbench2/view/280534

and here are the 2.4 8-core results:
http://browse.geekbench.ca/geekbench2/view/280564
Click to expand...

That's interesting! The 3.3 GHz 6-core is even a bit faster on average with multithreaded applications than the 2.4 GHz 8-core. Well, actually not so surprising since 8*2.4 = 19.2 whereas 6*3.33 = 20.

Strange though that Geekbench lists the 8-core as 16 cores/16 threads and the 6-core as 6 cores/12 threads ....?

Thanks for posting!!
 
My 8-core Nehalem 2.93 is 20% faster than the 6-core Westmere 3.33 running Geekbench 64-bit. Specifically the Integer and FP scores are higher but the Westmere memory scores win.

http://browse.geekbench.ca/geekbench2/view/280193

It will be interesting to see how they compare on real world apps like AE CS5, Photoshop CS5, FCP, Aperture, etc.
 
Fiete5401 said:
How do the memory scores translate into real world performance?
Click to expand...

The 8-core Nehalem 2.93 machine is a great machine. It is close in speed to the 3.33 6-core. If you look at the single-threaded scores, the 3.33 is 10-15% faster than the 8-core. If you load the computer completely (like with Handbrake) the 8-core will probably be a bit faster because it has 33% more cores (vs. the 6-core) but is only 10-15% slower on average.

So, there probably won't be a huge rush to upgrade the 2.93 machines unless they go for the 12 core machines. The 12-core should be 5-10% faster on single threaded tasks, but will be much faster (30%+) when you fully load the machine.

In my experience, running your machine fully loaded very rarely happens for most users. Digilloyd recommends the 6-core machine as the sweet spot in terms of performance and I have to agree.
 
It's not bad (I have the 8-core). There's a pretty constant low fan noise which I think is coming from the GPU.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back