Someone had posted this elsewhere, although only to the summary page where "Of 86 benchmarks carried out across all three operating systems, Ubuntu 19.10 was the fastest 48% of the time followed by Windows 10 with leads 31% of the time and macOS 10.15 at just under 20%."
I used to be pretty involved at some prior jobs doing various benchmarks, from porting of distributed load testing code, building test farms, automating results as well as adjusting HW and SW specs to see the net results, plus being a long-time Unix/Linux developer, had inherent interest as well as for setting up various infrastructure for my own servers and for datacenter, etc.
OS X historically has been at a disadvantage in areas of thread and filesystem performance. The first due to the inherent nature of OS X, wrapping NeXT 'mach' threads with it's own thread layer/library/interface - adding additional calls in place, which may be relatively trivial in normal use, but add up for sustainer performance or benchmark purposes. Back when Apple had tried to do 'OSX Servers,' I had run some tests running apache and some others across RHEL, Fedora, Win Server, etc... and OS X was surprisingly awful at that, as well. The other historic challenge was in their somewhat archaic filesystem, not to say it is inherently terrible, but Linux filesystems are in constant evolution, while I'm quite sure MS has put some real time into their filesystem performance as well, due to being under 'server OS attack' constantly by Linux.
So I initially went to the bookmark link expecting to see some not-surprising-to-me results, but what I found interesting is yes, the number of times MacOS was 'in the middle' was a good # of times, but when you dig into the tests being done, excluding OpenGL and game-based testing where I'd expect MS to hold a strong lead (not to mention Apple deprecating OpenGL)...macOS did pretty well. It's no real surprise that Ubuntu won out on some historically developed on/for Linux tests (e.g. PHP, git execution times), but macOS was pretty close behind in many cases - which I'd personally consider 'not bad at all.'
Some of the tests selected were questionable IMO, but kind of interesting - MacOS 'won' at Selenium tests using Firefox but performed poorly with Selenium tests on Chrome - this is more indicative of the underlying code optimizations, or not, that have been done within those respective browsers, IMO, across each platform.
I'd love to see another 'Snow Leopard' type release, focused on performance and optimization instead of thinner + more emojis, of course, but anyways - thought it was an interesting read..
I used to be pretty involved at some prior jobs doing various benchmarks, from porting of distributed load testing code, building test farms, automating results as well as adjusting HW and SW specs to see the net results, plus being a long-time Unix/Linux developer, had inherent interest as well as for setting up various infrastructure for my own servers and for datacenter, etc.
OS X historically has been at a disadvantage in areas of thread and filesystem performance. The first due to the inherent nature of OS X, wrapping NeXT 'mach' threads with it's own thread layer/library/interface - adding additional calls in place, which may be relatively trivial in normal use, but add up for sustainer performance or benchmark purposes. Back when Apple had tried to do 'OSX Servers,' I had run some tests running apache and some others across RHEL, Fedora, Win Server, etc... and OS X was surprisingly awful at that, as well. The other historic challenge was in their somewhat archaic filesystem, not to say it is inherently terrible, but Linux filesystems are in constant evolution, while I'm quite sure MS has put some real time into their filesystem performance as well, due to being under 'server OS attack' constantly by Linux.
So I initially went to the bookmark link expecting to see some not-surprising-to-me results, but what I found interesting is yes, the number of times MacOS was 'in the middle' was a good # of times, but when you dig into the tests being done, excluding OpenGL and game-based testing where I'd expect MS to hold a strong lead (not to mention Apple deprecating OpenGL)...macOS did pretty well. It's no real surprise that Ubuntu won out on some historically developed on/for Linux tests (e.g. PHP, git execution times), but macOS was pretty close behind in many cases - which I'd personally consider 'not bad at all.'
Some of the tests selected were questionable IMO, but kind of interesting - MacOS 'won' at Selenium tests using Firefox but performed poorly with Selenium tests on Chrome - this is more indicative of the underlying code optimizations, or not, that have been done within those respective browsers, IMO, across each platform.
I'd love to see another 'Snow Leopard' type release, focused on performance and optimization instead of thinner + more emojis, of course, but anyways - thought it was an interesting read..
