Understanding benchmarking

Hi, can anyone refer me to helpful links explaining benchmarking and how to properly compare Mac Pros from different years? Or if you can explain a thing or two about it yourself, that'd be cool


Where I'm stuck:
I can roughly compare CPUs and memory and other speed info, but when it comes to benchmarking, I'm not sure how to interpret things.

I understand that benchmarking are synthetic scores and don't always evaluate performance correctly. Also, as far as I can tell (at least for Macs) that the later the year the better the product, but I know there's some overlap between higher and lower end models, between two or more years.

For example, I see that a 2008 8-core 3.0GHz gets better benchmarks than a 2010 4-core 2.8GHz. To me, the 2010 should be the better model regardless, but besides being newer, I don't know how to compare properly. Confused. Thanks for any help.

markusaurileus said:

Hi, can anyone refer me to helpful links explaining benchmarking and how to properly compare Mac Pros from different years? Or if you can explain a thing or two about it yourself, that'd be cool

Click to expand...

It's really pretty straightforward. Basically the higher the score the faster the computer is. Yes...real world results will vary. But everyone's "real world" situation is different. So a benchmark is as best a relative comparison that can be done (system vs. system).

markusaurileus said:

Where I'm stuck:
I can roughly compare CPUs and memory and other speed info, but when it comes to benchmarking, I'm not sure how to interpret things.

Click to expand...

The higher the score...the faster the computer. Just remember...many benchmark apps just test the speed of the cpu/cpus. Fewer benchmarks apps also test the graphics.

markusaurileus said:

For example, I see that a 2008 8-core 3.0GHz gets better benchmarks than a 2010 4-core 2.8GHz. To me, the 2010 should be the better model regardless, but besides being newer, I don't know how to compare properly.

Click to expand...

In this example...I think that you're missing the fact that the 2008 Mac Pro has 8-cores...and the 2010 has 4 cores. You are correct...usually the newer computer is faster (a single 2010 core is faster than a single 2008 core).

But when comparing a 2008 Mac Pro with 8 cores (and running at 3.0ghz)...it's going to be faster than a 2010 Mac Pro with only 4 cores...AND running at a slower clock speed of 2.8ghz.

- Nick

Thank you for the reply, pigoo. I think I overlooked the fact that most benchmarks test only cpu or cpus.

Either way, it's just surprising to me that a 2008 machine can be better than a 2010. I always was under the impression that a newer cpu even at a lower clock could do much better than two older cpus... because of newer architecture? I don't know -__ -

markusaurileus said:

I always was under the impression that a newer cpu even at a lower clock could do much better than two older cpus... because of newer architecture? I don't know -__ -

Click to expand...

In many/most situations you are absolutely correct. And generally speaking...computing power can and does increase dramatically with time (Moore's Law).

If we look at things much more specifically...the benchmarks for a:

- 2008 8-core 3.0ghz Mac Pro = 8858
- 2010 4-core 2.8ghz Mac Pro = 8673

Thus the 2010 4-core Mac Pro has shown some terrific improvement in just 2 years (4 cores almost as fast a 8 cores was in 2008).

Now if we compare "Apples to Apples" (8-core vs. 8-core):

- 2008 8-core 3.0ghz Mac Pro = 8858
- 2010 8-core 2.4ghz Mac Pro = 12788

The 8-core 2010 Mac Pro does much better (approx. 50% better).

But even this comparison is not as fair as it could be...since the cpu clock speed of the 2010 8-core Mac Pro is only 2.4ghz (compared to the 3.0ghz of the 2008 Mac Pro).

If both 8-core computers had similar cpu clock rates...we could be talking closer to a 75% benchmark score improvement.

- Nick

Ahh, this makes a lot of sense. So there's definitely some performance overlap over the years, depending on the amount of cores, clock speeds, and changes in architecture.

I think I understand how to compare specs with each other now. Much simpler.

Btw, where did you pull those benchmarks from?

Apple Mac Benchmarks (All Macintosh Benchmarks) @ EveryMac.com

or
http://browser.primatelabs.com/mac-benchmarks

Thanks, I'm familiar with the geekbench 3 chart, but didn't notice the geekbench 2 data before. Helps to reinforce comparisons.