When trying to run gentoo, if you’re emerging with fewer than a few dozen cores (either in a single machine with something like a threadripper, or in a cluster with distcc), then I highly recommend using the binary versions of certain packages. This can be done either with -bin versions of packages, or something like the Gentoo Binary Host Project.
Packages that particularly benifit from using binary versions would primarily web browser or web browser adjacent packages such as Firefox, Chrome, QTWebEngine, but really any particularly large compile that doesn’t benifit from compiling locally (eg: not that many use flags, not likely to use any additional CPU features you might have such as avx512). In fact, bin versions of Web browsers often will perform much better than locally compiled versions since they are compiled with additional optimisations that either make the compile time even longer (O3 and LTO), or require additional manual steps (such as PGO where the unoptomised browser is compiled and ran through real-world workloads with a profiler attached to identify code hotpaths so the compiler can optimise more efficiently during a second complete compile run).
When trying to run gentoo, if you’re emerging with fewer than a few dozen cores (either in a single machine with something like a threadripper, or in a cluster with
distcc), then I highly recommend using the binary versions of certain packages. This can be done either with-binversions of packages, or something like the Gentoo Binary Host Project.Packages that particularly benifit from using binary versions would primarily web browser or web browser adjacent packages such as Firefox, Chrome, QTWebEngine, but really any particularly large compile that doesn’t benifit from compiling locally (eg: not that many use flags, not likely to use any additional CPU features you might have such as avx512). In fact, bin versions of Web browsers often will perform much better than locally compiled versions since they are compiled with additional optimisations that either make the compile time even longer (O3 and LTO), or require additional manual steps (such as PGO where the unoptomised browser is compiled and ran through real-world workloads with a profiler attached to identify code hotpaths so the compiler can optimise more efficiently during a second complete compile run).