Something to consider when looking at Geekbench scores

[theSeb]

Geekbench has become the most popular and often quoted benchmark on these forums. People use it for various arguments. Intel has been concentrating a lot on the performance of the mobile CPUs and has, arguably, been ignoring the desktop performance. We see this trend in Geekbench, with Mobile CPUs in the Macbook Pros snapping at the heels of the top of the range iMac and the iMac snapping at the heels of the Mac Pros. However, as I have always said, Geekbench is a sprint and shows mainly how fast the CPU is with maximum Intel Turbo Boost. It does not show how well the system will perform in real use.

The problem is that pmset -g thermlog clearly shows that when we run Geekbench, we receive no events, therefore the CPU is running at 100% during this time with no speed limit. Geekbench’s workload is simply not strenuous enough to see the impact of thermal throttling and makes comparisons skewed towards CPUs with the maximum Turbo Boost clock rate. Basing purchasing decisions on this is not a good idea in my opinion.

The second, less important point, is that I am pretty sure that at least some people are trying to compare 32 bit and 64 bit GB scores.

 

[ha1o2surfer]

Geekbench has become the most popular and often quoted benchmark on these forums. People use it for various arguments. Intel has been concentrating a lot on the performance of the mobile CPUs and has, arguably, been ignoring the desktop performance. We see this trend in Geekbench, with Mobile CPUs in the Macbook Pros snapping at the heals of the top of the range iMac and the iMac snapping at the heal of the Mac Pros. However, as I have always said Geekbench is a sprint and shows mainly how fast the CPU is with maximum Intel Turbo Boost. It does not show how well the system will perform in real use.

The problem is that pmset -g thermlog clearly shows that when we run Geekbench, we receive no events, therefore the CPU is running at 100% during this time with no speed limit. Geekbench’s workload is simply not strenuous enough to see the impact of thermal throttling and makes comparisons skewed towards CPUs with the maximum Turbo Boost clock rate. Basing purchasing decisions on this is not a good idea in my opinion.

The second, less important point, is that I am pretty sure that at least some people are trying to compare 32 bit and 64 bit GB scores.

This guy is EXACTLY correct. It puts a load on the CPU over a small period of time; this is enough for the cooper heat sink to absorb even without any increased fan activity. Once that thermal threshold has been passed,throttling begins on Macbooks

Try running a Linx/Prime benchmark for a couple minutes instead to get better readings for performance/thermal dissipation performance.

 

[FredT2]

And not just Geekbench. There are plenty of other benchmarks that are put up in reviews that are just as unrealistic. Can you imagine running HandBrake encodes that take a minute or two and making judgments based on the results? (Macworld I'm talking to you!)

 

[MacUser2525]

And not just Geekbench. There are plenty of other benchmarks that are put up in reviews that are just as unrealistic. Can you imagine running HandBrake encodes that take a minute or two and making judgments based on the results? (Macworld I'm talking to you!)

Speaking of Geekbench and Handbrake when I upgraded from my 2.8 to 3.33 in my 5,1 I ran before and after tests using them. Geekbench went from just over 8K to just over 15K or just about double. The test encode of h264 video I did before and after about 2 hours long the new processor did it in about half the time just like you would expect from the Geekbench result. You people really make me laugh Geekbench used to be the be all and end all around here until it showed the new shinny Mac Pro in a bad light, don't worry knowing human nature I'm sure you can still come up with the justification to buy one...

 

[mrsavage1]

Exactly geekbench doesn't tell the full story of the mac pro. Its not as simple as looking a number and declaring ha ! the new mac pro is slower than my full config imac.

 

[propower]

+1
People think a MacMini is almost as powerful as a Quad MacPro because geekbench said so... LOL
Thermal control is 1/2 if not 3/4 of the battle -- TDP most of the rest...
My 2013 imac doesn't hit TDP thermal limit till almost 85% constant CPU load. Even then falls back to 3.5. It WILL give a Quad nMP a run for its money!

Merry Christmas!

 

[sidabet]

I upgraded my Mac pro from a Quad to a Hex and geekbench 3 64bit went from 9996 to 16134 and as expected I got a proportional decrease in my encoding times.

But what Geekbench does not include is the impact of the GPU, eg the mercury rendering and playback engine in Adobe premiere.

I believe this is where the new Mac Pro will really score. Final Cut Pro 10.1 has already been updated to support dual GPUs

 

[Tutor]

Geekbench has become the most popular and often quoted benchmark on these forums. People use it for various arguments. Intel has been concentrating a lot on the performance of the mobile CPUs and has, arguably, been ignoring the desktop performance. We see this trend in Geekbench, with Mobile CPUs in the Macbook Pros snapping at the heels of the top of the range iMac and the iMac snapping at the heels of the Mac Pros. However, as I have always said, Geekbench is a sprint and shows mainly how fast the CPU is with maximum Intel Turbo Boost. It does not show how well the system will perform in real use.

The problem is that pmset -g thermlog clearly shows that when we run Geekbench, we receive no events, therefore the CPU is running at 100% during this time with no speed limit. Geekbench’s workload is simply not strenuous enough to see the impact of thermal throttling and makes comparisons skewed towards CPUs with the maximum Turbo Boost clock rate. Basing purchasing decisions on this is not a good idea in my opinion.

The second, less important point, is that I am pretty sure that at least some people are trying to compare 32 bit and 64 bit GB scores.

Also, if you're reviewing results online in the Geekbench Browser, make sure that the Geekbench result submission date(s) and the number and model of CPU(s [only for pre2013s]) listed seem plausible in relationship to the system you are trying to judge.

Here a listing of submissions for so-called MacPro6,1s: http://browser.primatelabs.com/geekbench3/search?dir=asc&page=7&q=macPro6,1&sort=created_at

Can you spot the ones which are clearly bogus in that the system is not likely to be a true MacPro6,1? What's the dividing line? What's indicative of it being a Hackintosh?

One hint - click on the "Uploaded" column heading to order the results by date.