@morphmex: I'm very happy that you've seen the differencies from a precompiled "generic" distro to the Magical Funtoo :P :P:P
Funtoo must never Dies :D :!!!
I am willing to invite anyone to my house to let him try a precompiled vs Funtoo :P
There is no battle: D
Especially if you have a processor that can benefit from the various new flags implemented (ssse3 se4.x avx avx2 etc...)
I take a trivial example: AMD invented AMD64 technology, soon Integrated by Intel with EM64T technology.
Now there is talk of "amd64" or "Intel64" where machine-level there are little differencies (eg AMD already has a Physical adress of 48 bits while Intel 39-bit (48 in compatibility mode).
When he was born the technology x86_64, gcc started to support this technology.
So with a 64 AMD processor generation I would have gotten more or less the same results between the use -mtune=generic & mtune="cpu_type".
Clearly more processors are technologically improved by adding many instructions (SSE3 SSSE3 SSE4.1 SSE4.2 then aes-ni AVX AVX2 etc.) the difference between a prebuilt package for "generic" and one compiled for Your cpu always does a greater difference (obviously depend very much on the ability of GCC to use algorithms that use the various new SIMD).
I know people very good and with a gret knowledge about Linux systems ... but for them is not important compile from sources.
I also said in the face of one Red Hat developer (who argued the goodness of their package-manager and "denigrating Portage):that "For me the package management system more eclectic and brilliant on this planet, form me iss Portage". :D
not confirmed anything .... maybe because he only "saw Portage" but never used it seriously :)
Good for All :)
PS: I've a 4771 cpu; comparing with phoronix benchs (using a 9360 with 6 cores / 12 thread) with Funtoo and his optimizations, in some tests win .... in other tests the values are "similar".
For example with my cpu:
$ phoronix-test-suite benchmark c-ray
give me 22 seconds.and in "Apache benchmark" is more efficient than a *BSD : http://www.phoronix.com/scan.php?page=article&item=linux_11_way&num=3
About "more fun" ? mmmm I don't know..... for me Funtoo is already fantastic at this time..... I have no idea how to improve Funtoo ....
but .... I think that only Funtoo can beat Funtoo.
Maybe the implementation of a branch based on * BSD kernel?
I do not know .... for me is already excellent as well :)
I hope not to have bored with this post
I am going to be making some changes to the project soon, to fulfill the vision of a user-centric community and also make Funtoo more suitable for enterprise systems.
Here is what I have found. If you have a single Funtoo system, then the continual changes that we receive from Gentoo are probably not a big deal. You do a system update daily or maybe once a week, resolve any issues, and your system continues to work well with a little hand-holding.
However, there is a lot of work being done behind the scenes by angry_vincent and myself to keep things working well. Our tree is created in a totally automated way, but we "fix" things that break as various things in Gentoo are updated.
And there is the problem. The automated Gentoo updates are fine for a single system, but if you like me have 5+ Funtoo systems, you do not update your systems regularly and then the various hand-holding, perl-cleaner, revdep-rebuild and blocker resolution is a huge pain.
And there is another problem... Angry_vincent and I probably spend 90%+ of our Funtoo time on fixing things that are broken, which means we have no time left to actually create new awesome technology to make Funtoo better. Like a boot-update rewrite, or a metro rewrite, all things that I want to get to. Yes, I find a way to sneak in time to do various things but it is not very ideal.
So, for any problem there is a solution. I have a vision of a solution that will give us the ability to have more control over quality, without doing a wholesale fork of all of Gentoo. I don't see the point of forking all of Gentoo... without better planning and technology, we will just duplicate the same problem of continuous rolling release, but have it happen on the Funtoo side. Rolling release is a good thing, but it is not suitable for all users of Funtoo or Gentoo... so we need something that will make Funtoo more capable without working against the benefits of rolling release.
Since I am not only a developer, not only BDFL, but also a *user* of Funtoo Linux (and *that* is what I consider myself first and foremost,) I want Funtoo Linux to work for larger scale deployments. And I want developers and *users* to not be continually burdened with unending ebuild updates and break/fix situations.
So, over the next few weeks, I am going to be reaching out to our small team of staff and get them up to speed on regular Funtoo maintenance so that I can focus my 'Funtoo' time on developing this new system. Not a new distro, but a new way that Funtoo will manage updates that will give us much more control and the ability to focus on strategic initiatives much, much more than we have in the past.