Cinebench results OS X vs Win10

[Hene1]

When I ran Cinebench in both OS X and Windows 10, graphics performance was slighty better in Windows (100 in Windows vs 96 fps in OS X), but CPU score was much lower (440 in Windows vs 560 in OS X). Do anyone has any idea why CPU performance in Cinebench is so much worse in Windows 10 compared to OS X?

Late 2015 iMac with i5-6500 and R9 M390.

 

[h9826790]

When I ran Cinebench in both OS X and Windows 10, graphics performance was slighty better in Windows (100 in Windows vs 96 fps in OS X), but CPU score was much lower (440 in Windows vs 560 in OS X). Do anyone has any idea why CPU performance in Cinebench is so much worse in Windows 10 compared to OS X?

Late 2015 iMac with i5-6500 and R9 M390.

May be there was something running in the background in windows 10 when you benchmarking your CPU.

The result is consistent?

 

[Trebuin]

When I ran Cinebench in both OS X and Windows 10, graphics performance was slighty better in Windows (100 in Windows vs 96 fps in OS X), but CPU score was much lower (440 in Windows vs 560 in OS X). Do anyone has any idea why CPU performance in Cinebench is so much worse in Windows 10 compared to OS X?

Late 2015 iMac with i5-6500 and R9 M390.

Mac's Proprietary hardware is the reason. Mac's software is built for very few hardware variations. Windows builds multiple hardware variations and has code to deal with that. Because of this, they lose on the CPU side of the world. This is also driven by bus configurations, motherboards, sound cards, etc. GPU is different as there's only ATI and nVidia as the main market players. Apple video drivers are not constantly updated compared to the microsoft side of the world.

As h9826790 pointed out, your background processes will be different as well which will play a big role.

 

[maflynn]

Different platform (windows vs osx), and so there will always be slight variations because we're dealing with two different applications. The output is the same and the measurements may be the same but how that is encoded in the program is completely different.

 

[joema2]

Mac's Proprietary hardware is the reason. Mac's software is built for very few hardware variations. Windows builds multiple hardware variations and has code to deal with that...

If he was running it on the exact same hardware (e.g, Boot Camp) then the only difference between Windows and OS X Cinebench is software. If the Cinebench CPU score is higher on OS X vs Windows 10 on the same machine, there are various possible explanations including background tasks, OS thread scheduler differences, Cinebench could be built with different compiler optimization flags on OS X vs Windows, or just random variation if he only ran it once.

If he was running it on different machines (IOW iMac vs some Windows PC) then you're right -- the test has no meaning since the hardware is different.

 

[Trebuin]

If he was running it on the exact same hardware (e.g, Boot Camp) then the only difference between Windows and OS X Cinebench is software. If the Cinebench CPU score is higher on OS X vs Windows 10 on the same machine, there are various possible explanations including background tasks, OS thread scheduler differences, Cinebench could be built with different compiler optimization flags on OS X vs Windows, or just random variation if he only ran it once.

If he was running it on different machines (IOW iMac vs some Windows PC) then you're right -- the test has no meaning since the hardware is different.

I'm referring to hardware being limited on the Mac side allows for Apple to keep the software tuned to run on less configurations, giving them the edge over Windows. Proprietary hardware is what allows them to achieve what you and I both listed over Windows, who has to make sure so many configurations are built into their software. In the end, yes software is the cause, but ultimately not the root cause. Keep asking why until you can ask no more to find the root cause.

 

[joema2]

I'm referring to hardware being limited on the Mac side allows for Apple to keep the software tuned to run on less configurations, giving them the edge over Windows. Proprietary hardware is what allows them to achieve what you and I both listed over Windows, who has to make sure so many configurations are built into their software. In the end, yes software is the cause, but ultimately not the root cause. Keep asking why until you can ask no more to find the root cause.

I just ran both OS X and Windows versions of Cinebench on my top-spec 2015 MacBook Pro using Boot Camp and Windows 8.1. The CPU test results were essentially identical between OS X and Windows -- 622 vs 623 (see attached). On the OpenGL GPU test the Windows version was faster -- 74.4 fps vs 62.98 fps.

The OP asked why the Cinebench CPU benchmark produced faster results on OS X than on Windows -- presumably on the same machine, although he didn't specify.

If both OS X and Windows Cinebench tests were run on the same machine, I don't see how "proprietary" Mac hardware could have any difference in the Cinebench CPU test. It's a small CPU benchmark running on the exact same hardware, just OS X vs Windows. It's not doing a lot of I/O or other system calls. It is possible whoever compiled and linked the benchmark might have used different CPU optimization flags on the OS X vs Windows version but if so that's not due to the underlying hardware difference -- hardware is the same. However on my MBP there was no real difference in Cinebench CPU performance between OS X and Windows, so it's more likely the OP's results were spurious or due to background tasks, not due to proprietary hardware.

I also ran OS X Cinebench on my 2015 top-spec iMac 27 and it produced 868 for CPU and 107.49 fps for GPU. I don't have Boot Camp installed on that machine so I can't run a Windows test. However I'd expect it to be similar to the Windows vs OS X Cinebench CPU test on m 2015 MBP, which showed essentially no CPU difference.