Speed of Clear Linux

> I remember that this distribution would often come out the winner in benchmarks. I don’t know whether that’s because Intel did a better job with compiler flags to build for their hardware (as long ago with pgcc) or through a better choice of kernel configuration and C standard library or whatever.

It was a combination of all of those, on top of a substantial amount of upstream work to make those individual components faster. [1]

However, keep in mind that Clear Linux's benchmark lead came largely at the expense of backwards compatibility -- if you didn't have the latest-and-greatest processors (or more accurately, instruction set extensions) then Clear simply wouldn't even *boot*,. This is not a compromise that more mainstream (and more general-purpose) distributions were willing to make... at least not at the time.

Today the situation is different; mainstream distributions increasingly target newer processor families (eg RHEL requiring the 12-year-old x86_64v3) and even without that, more and more performance-sensititve upstreams now detect and switch to advanced ISAs at runtime. So Clear Linux's original raison d'etre[2] is mostly moot.

[1] Intel will presumably be continuing most of this upstream work.
[2] Showcasing the performance gains possible when using Intel's newest processors -- in other words, marketing.