Is my computer suitable for gentoo?

[cien]

Hi, I just installed gentoo on a spare old computer. I have been paying attention to gentoo for a long time, and now I have the conditions to install gentoo on a physical machine. But my computer is too old. When I installed the following software, the software package net-libs/webkit-gtk took a long time to compile. It took one night and half a day and it was not compiled yet. This makes me a little confused whether my computer is suitable for installing and using gentoo. Can anyone give me some advice? Thanks a lot.

[NeddySeagoon]

cien,

That system is suitable for Gentoo, just a little slow.
With only 4G RAM, you may want to restrict MAKEOPTS for big packages, as building on fewer cores will be faster than running out of real RAM and swapping.
Most of the content of make.conf can be set on a per package basis.

Gentoo also splits nicely into what you need to run Gentoo and what you need to build Gentoo.
That means that its possible to make your own binhost the way you like it. I keep an old AMD 350 alive that way.

[cien]

[eccerr0r]

There was someone with a Pentium Sandybridge recently that was getting sigills and looks like this is happening. Make sure your CFLAGS do not use -march=sandybridge -- at least you need to disable AVX if you do.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

[cien]

[eccerr0r]

In your COMMON_FLAGS you have "-march=native" which is telling gcc what instruction set it should output. I wonder if gcc can figure out you're using a non-avx sandybridge with -march=native else you'll need to explicitly add -fno-avx after -march=native in your make.conf ... but unfortunately you'll need to rebuild any packages that got the illegal instructions - which at minimum is x11-libs/libdrm .
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

[cien]

[Hu]

Start by showing us what gcc interprets -march=native to mean on your hardware, and show us the contents of /proc/cpuinfo.

You mentioned using binary packages. Are the broken files ones that you built locally or did you get them from a Gentoo binhost?

[eccerr0r]

Run this:

[CaptainBlood]

Please consider setting PORTAGE_TMPDIR to hard disk at package level definition:
wiki
Local example here

[Hu]

Although OP is a bit low on memory to build large packages in tmpfs, at this time we have no reason to believe OP is using a tmpfs for any package, large or small. If /var/tmp/portage is on physically backed storage, there is no need for a per-package override.

[CaptainBlood]

Let's restate my Luddite pov.
It should be convenient & compact to default PORTAGE_TMPDIR to tmpfs.
Such targeted packages would build faster.

However memory hungry package builds should be redirected to HD as described initially.
Such targeted packages could allow higher MAKEOPTS.

Thks 4 ur attention, interest & support.
_________________
USE="-* ..." in /etc/portage/make.conf here, i.e. a countermeasure to portage implicit braces, belt & diaper paradigm
LT: "I've been doing a passable imitation of the Fontana di Trevi, except my medium is mucus. Sooo much mucus. "

[NeddySeagoon]

CaptainBlood,

What you say is factually correct but does not help fix the immediate issue, which is Illegal Instruction exceptions.
The OP should save your advice for later, once their install works properly.

Making it work is the priority now.
Making it leaner, meaner, faster, whatever can come later.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

[eccerr0r]

Okay yes based on some anecdotal evidence from here and on stackexchange, -march=native for gcc on a Celeron or Pentium Sandybridge will produce incorrect results. -march=native will work properly only on Sandybridge CPUs that are i3-2xxx, i5-2xxx, i7-2xxx, and their Xeon counterparts.

So I think that COMMON_FLAGS in /etc/make.conf will need to be changed to something like:

COMMON_FLAGS="-O2 -pipe -march=sandybridge -mno-avx -mno-aes"

and you'll need to

# emerge --oneshot x11-libs/libdrm

This hopefully will fix your X11. Then the other problem is there may be other packages that have this issue. You may be forced to

# emerge --emptytree @world

but don't do this yet unless you have absolutely no way to find which packages were compiled wrong. It may be faster to start over from a fresh stage3 if you haven't invested much time.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

[Hu]

It should be possible to identify bad packages by finding those which have -march=native recorded in their CFLAGS in the vdb. The user would still need to identify which packages actually used those CFLAGS[1], but it would be a smaller set than emptytree.

[1] Packages that only install text files, like documentation or Python scripts, will have recorded their CFLAGS, but never used them, and so do not need to be rebuilt.

[CaptainBlood]

[eccerr0r]

Don't know. I've never had one of these cpus to independently verify that this is true, but there's not just one person that have tried march=native and march=sandybridge with no other flags that end up with unusable avx instructions. I suspect that there are other low end cpus that have this issue.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

[cien]

[cien]

Thank you all for your response. I really appreciate it.
The following error occurred when I ran the following command. Is it caused by the same problem?

[cien]