I'm in the process of setting up my workstation with Gentoo, and I have no idea why the compile times are as long as they are. I have used Gentoo in the past(over a year ago), and switched to Redhat distros because it seemed more practical at the time. Systemd has driven me back to Gentoo. The only problem is how long the compiles are taking. I'm running a 4930k with 32GB of RAM and an SSD for the OS drive. MAKEOPTS is set at -j12 and -march=native. I've been compiling xorg-server and kdebase-meta since 11:30 AM EST, and am not close to done. I remember the last time I installed Gentoo on my AMD 8350, and it didn't take this long. Has something changed? What am I missing?

Well, the code bases will have increased, so maybe that's part of the issue.

-j also is simply the number of jobs you want to use, not cores. So you could try going higher, especially since, I assume, you are no really using the computer for much else as you compile X and KDE. Try -j24 or -j32

That makes sense. It just seems like a really steep drop in performance. I will try increasing the jobs. It isn't doing anything but compiling.

Hard to say what's going on.

You can check cpu and memory usage and see if there's something strange going on besides gcc and python/emerge. Maybe some defunct process is doing something in the background. That should at most keep one core busy, not all of them, usually. If the ram fills up and you start paging to disk, then any system will be brought to its knees. It really doesn't matter how amazing your cpu is.

Sorted that out, I'll say that the amount of packages and code to compile can vary greatly depending on the architecture, USE flags, package versions, etc.

Most definitely, the kde code base has surely increased from 3.x to 4.13 (or .12), and if you changed from x86 to x86_64 then you need to compile two toolchains unless you are in no-multilib.

I'm not really sure exactly what is happening or why, but it seems to run faster with X running. In any case the CPU is maxed on big compiles(like firefox), but the RAM is nowhere close. My guess is that it must just be the larger codebase.

If it runs faster in X, there's a chance that your framebuffer drivers is eating an insane amount of cpu time. Still, that shouldn't affect more than one core unless there's some insane bug involved.