macOS 10.14 Mojave vs. Ubuntu 18.04 LTS vs. Clear Linux Benchmarks

[cen1]

Well, at least you have ethernet port, I have 2 usb on my Mac, one for ethernet dongle, second one for mouse adapter. And that's it, I am out of ports. It does have 2xminiDp and HDMI out tho so that is pretty good. Still need a converter to connect to monitors but that is to be expected.

Agree on those USB port placements tho.. horrible. You'd think in 2018 laptop makers would figure simple things like this out.

[RomuloP]

Kidding aside: The tests involving IO are admittedly not that great but beyond that, it is basically a compiler benchmark. Apple LLVM is probably not even based on vanilla LLVM 6 and due to their programming paradigms, it still doesn't support OpenMP. If you actually care about GraphicsMagick etc., fire up brew, get clang 7 or gcc 8, and you're good to go. That works quite well for my code. Additionally, macOS tends to be very power conscious when it comes to scheduling. At least in the past, the battery benchmarks seemed to indicate this. All in all, the OS simply isn't made for raw throughput.

It is understandable, different paradigms. I've had one for 1 entire year, they are not bad, but, summed up, the aesthetic and interaction difference was pointless to my work and so not enough to keep me in so I moved out.

[caligula]

[QUOTE=cen1;n1050399
Agree on those USB port placements tho.. horrible. You'd think in 2018 laptop makers would figure simple things like this out.[/QUOTE]
It's called aesthetics. The UI/UX designers don't care about expert opinions from techies.

[RussianNeuroMancer]

Friend of mine fed up with OS X on his MacBook, so I googled up a little Linux drivers situation. Mostly fine, as I understand, but some stuff like fan management and web camera require external tools/drivers. It's good that mbpfan is packaged, and it's possible to create dkms package for web camera driver, but I want to know if there is other issues I should be aware of? How suspend works? Is battery consumption with tlp is acceptable? Stuff like hotkeys, touchpad, wifi and other stuff works without issues?

[torsionbar28]

Additionally, macOS tends to be very power conscious when it comes to scheduling. At least in the past, the battery benchmarks seemed to indicate this. All in all, the OS simply isn't made for raw throughput.

Most likely because all their chassis designs, aside from the Mac Pro, have total crap for cooling. The price you pay for being fashionably slim, it seems.

[GruenSein]

Most likely because all their chassis designs, aside from the Mac Pro, have total crap for cooling. The price you pay for being fashionably slim, it seems.

Might have something to do with it but I am not sure if this is the main reason. My 2014 Retina MBP maintains at least base clock on all eight threads. Yes, it gets warm and the fans spin up, but it doesn't seem like it is throttling much. My guess is that the thermal management in macOS is just more conservative when it comes to bumping up the frequency, which would show mostly in intermittent loads. But anyway, if you are fine with a much thicker, heavier machine, you can get more performance (particularly on the GPU side) than in an MBP. I simply like the form factor, display quality etc. pp. while rendering times are of minor interest to me.

[Scellow]

Again, what's the point of comparing two different machines ?
If we normalize the specs, macOS beat linux

Please setup a hackintosh, and test with the same machine

[Michael]

Again, what's the point of comparing two different machines ?
If we normalize the specs, macOS beat linux

Please setup a hackintosh, and test with the same machine

The same system is used the entire time, as stated. Any reported difference in the automatically generated system table just comes down to how each OS exposes the software/hardware.

[Royi]

This is just Apple LLVM vs GCC test.
Why wouldn't you use the same compiler on all platforms?
Those tests do not test the OS and shouldn't be treated as such.

[Michael]

This is just Apple LLVM vs GCC test.
Why wouldn't you use the same compiler on all platforms?
Those tests do not test the OS and shouldn't be treated as such.

Because on each OS most users/developers rely upon the default toolchain... But there are also a number of tests here where that toolchain doesn't come into play anyhow from Go to Perl to Python to PHP, etc.