So, let’s imagine you’re a happy owner of Raspberry Pi, you either just unpacked it, or may be used it for some time.
And for some reason you’ve decided to install not just your favourite binary distribution, but mighty source-based Gentoo linux.
First of all checkout Gentoo wiki pages dedicated to Raspberry Pi: , .
Also checkout Beyond Linux From Scratch on the Raspberry Pi as it also covers the topics we discuss here.
While Gentoo is already a lot about utilizing much of your processor power and I/O bandwidth for the compiling packages, Gentoo on a Raspberry Pi faces a real bottleneck when it comes on compiling on the same machine.
There are several ways to speed the things up.
- Hardware-related: speed up the RPi itself, i.e.
build RPi build farm, buy smth else overclock it.
- Software related:
- Distributed compilation: e.g. cross-compilation with distcc.
- Building binary packages inside VM:
qemu-system-arm, for QEMU+RPi initial configuration see ,  (in Russian).
- Building binary packages inside special environment: chroot + static
Just found out that Fullprof ebuild appeared in gentoo overlay science about a year ago.
These are the good news :-)
However, GUI part of the fullprof package is still better in windows version, so use it under wine.
Finally switched back from funtoo to gentoo after 2 years.
Git-based portage tree, concept of the flora overlay and flavored profiles are really nice & cool ideas!
However, outdated portage tree is a pain in the ass for me… :-(
Guys, why not just importing everything from gentoo portage, patching whatever you need to patch, and mask whatever you need to mask?
GCC-4.7.4 and 4.8.1 are too unstable? Just do not keyword or simply mask them. But let them be in a tree.
Currently I’m a PhD student, finishing my thesis. During the preparation of the manuscript I used LaTeX for the typesetting and git for the revision control.
I would like to share certain ideas and techniques which for the preparation of the manuscript and data visualization, so following posts of this blog will cover (but won’t be limited to) this topic. And I hope that this experience will be useful for anybody else.
Here are some execution time benchmarks of Maxima computer algebra system.
Maxima 5.29.1 was compiled with various Lisp compilers: CLISP, ECL and SBCL. For each case function run_testsuite(display_all = true, time=all); was used for comparison of execution speed of Maxima with these compilers.
Benchmarks were run on Funtoo runnning 3.7.5-pf kernel on i7-2630QM CPU @ 2.00GHz 4Gb RAM. The system was running single shell without X,networking, etc. No unexpected errors (there are tests resulting in error on technical purpose) were found out of 9,519 tests with each compiler. I show the timing results under the cut. Continue reading