Fedora's Firefox To Stick With GCC Over Clang, Beefed Up By LTO/PGO Optimizations

[pal666]

They are targeting servers

how is that fair to compare to general-use distros?

[kgonzales]

They are targeting servers and you can use CPU's from 2011 onwards. So how is that "no legacy" and "no production"?

They have no SLA and no paying users. If it breaks, you can keep both pieces. Therefore, no production.

The project is barely 3 years old. Therefore no legacy users. Compared to people who have used Fedora for 10+ years, or distros like Debian for longer.

That is how it is "no legacy users" and "no production".

[skeevy420]

.

True but the point of Gentoo is to not do a binary distribution but to actually optimize for each unique installation.

So, make a default config file for your target machines with -march=native and the use flags you want and copy it on each install..

That doesn't change the fact that Gentoo can be setup to compile somewhat optimized binaries for all the architectures a person owns so they don't have to do, for example, 4 different march=native builds or march & mtune=some_arch builds on the fastest/best compiling PC they own (also allows for using distcc since we can't do that safely with native). All they'd have to do is setup the best PC with march=worst_arch (or =generic) mtune=best_arch, maybe use distcc, and they'd have something that works for everything they own (well, should technically work).

[treba]

Just tested the fedora 29 package. Performance on speedometer went from 54.9 to 64.53 which would be a gain of 17.54% if I got the calculation right

[hubicka]

Note that http://hubicka.blogspot.com/2018/12/...lding-and.html has some additional benchmarks to speedometer. One remaining issue is the fact that Skia (a graphic library used to render some stuff) needs to be ported to GCC. Currently it has hand optimized vector rendering code only for Clang. I plan to look into that after finishing some GCC work - Firefox is very good interesting real-world LTO benchmark and there was number of things to fix/improve for GCC 9 which I noticed while looking into its performance.

[brad0]

lol at upstream firefox devs. if only there was way to submit bug report to gcc instead of dancing between archaic gcc and superior competition

There, corrected that for you.

[boltronics]

Typing this from a work computer built in 2010 - an i7 930 with no EFI support... and I've no plans to replace it because it's working fine. Only supporting CPUs from 2011+ would be an epic fail.

[kpedersen]

I mean, Clear Linux has shown us that LITERALLY NO (I mean, except Clear) Linux distros are making efforts for performance. Its sad that Intel had to come in and do the work that every distro should have been doing from the start.

Quite a few Linux desktop users also left Windows to get away from this wasteful culture of consumerism and pointless hardware refreshes. Knowing that my OS will quickly cease to correctly support my hardware is a little bit careless.
Of course Intel is doing this because they want to *sell* you more shite!

Perhaps your want for consumerism and quick deprecation cycle means you would best be suited within the Apple ecosystem.

For example, Debian is the "Universal Operating System". Not just the OS for consumeristic hardware junkies. There is more to computing than raw hardware performance and the internet.

If your interest *is* raw hardware performance, or server grade workloads then please ignore my "consumerism" rant. Its not aimed at you. That said I would say that this interest is a bit more of a niche for the Linux desktop world. Most people just want stuff to work so they can get on with their life

[skeevy420]

Typing this from a work computer built in 2010 - an i7 930 with no EFI support... and I've no plans to replace it because it's working fine. Only supporting CPUs from 2011+ would be an epic fail.

Not really....it's essentially the AVX division line (which, oddly, isn't listed in their requirements). I made/suggested the same division line when I said I'd stop Gen 1 x86_64 with Westmere\AES and would start Gen 2 with Sandy/AVX -- we all might not necessarily agree with it, but it does make perfect sense from Intel's position and with their goals for Clear Linux.

EDIT: Just remember that Clear Linux is open source and that there is one general purpose distribution that includes their tweaks that doesn't have the same target requirements, Solus Linux, and that source based distributions like Gentoo allow us to include the same tweaks without those target requirements.

[skeevy420]

Quite a few Linux desktop users also left Windows to get away from this wasteful culture of consumerism and pointless hardware refreshes. Knowing that my OS will quickly cease to correctly support my hardware is a little bit careless.

If your interest *is* raw hardware performance, or server grade workloads then please ignore my "consumerism" rant. Its not aimed at you. That said I would say that this interest is a bit more of a niche for the Linux desktop world. Most people just want stuff to work so they can get on with their life

It can be both. Being able to get more life and performance out of older hardware means we're not being wasteful consumers by tossing it aside and upgrading as fast as possible. Older hardware is also cheaper; keeping it performing optimally allows poorer people to have decent computing experiences.