4 ms·

This is why you always have to benchmark YOUR software, your usecase. If Linux works better for your use of "dav1d" or whatever else is on this list, then it ma

This is why you always have to benchmark YOUR software, your usecase. If Linux works better for your use of "dav1d" or whatever else is on this list, then it may be what you should be using. I use nothing on this list of benchmarks other than git, and I think I'm good on the trade offs here to move to desktop Linux. I personally love Windows 11 and typing this on it, with my 11900K. I won't be switching.

Sesse__ 5y ago
Not defending Phoronix' benchmarks, but do you ever play YouTube videos? If so, you're most likely using dav1d.
marcodiego 5y ago
These benchmarks are quite general and good measurement of how high performant apps behave when running on different kernels. This advantage probably translates to other workloads as well. At the least, these results show that software vendors could improve the performance of their applications by porting to linux.
- BuckRogers 5y ago
  You need to actually prove that for each piece of software being ported, otherwise it's an assumption for potentially a million dollar investment for what would be a very poor business decision. It always depends on who is paying, you? If it were an outright win, we'd all be on desktop Linux ages ago. It's a trade off, and you lose a lot. Including customers, by wasting money on a desktop Linux port when you can deliver value where 95% of your customers are. The other 5% is macOS. Your competition is most likely focused on Windows. Other than server usage that is, Linux is a fine server. For my money, if I did find a performance loss on Windows, I'd start by working on either improving my implementation, or if it is a kernel issue to collaborate with Microsoft to equalize that. Rather than kick off years of support, and potentially millions of dollars on what is comparatively a barren wasteland.

4 ms·

This is why you always have to benchmark YOUR software, your usecase. If Linux works better for your use of "dav1d" or whatever else is on this list, then it ma

by BuckRogers 5y ago

Sesse__ 5y ago
Not defending Phoronix' benchmarks, but do you ever play YouTube videos? If so, you're most likely using dav1d.
marcodiego 5y ago
These benchmarks are quite general and good measurement of how high performant apps behave when running on different kernels. This advantage probably translates to other workloads as well. At the least, these results show that software vendors could improve the performance of their applications by porting to linux.
- BuckRogers 5y ago
  You need to actually prove that for each piece of software being ported, otherwise it's an assumption for potentially a million dollar investment for what would be a very poor business decision. It always depends on who is paying, you? If it were an outright win, we'd all be on desktop Linux ages ago. It's a trade off, and you lose a lot. Including customers, by wasting money on a desktop Linux port when you can deliver value where 95% of your customers are. The other 5% is macOS. Your competition is most likely focused on Windows. Other than server usage that is, Linux is a fine server. For my money, if I did find a performance loss on Windows, I'd start by working on either improving my implementation, or if it is a kernel issue to collaborate with Microsoft to equalize that. Rather than kick off years of support, and potentially millions of dollars on what is comparatively a barren wasteland.