LPIA is Low-Power Intel Architechture. If you use the prior nomenclature they used at Intel, it really ought to be IA-LP.
Hehe...that's the hand wavy description. I really want details since I run gentoo (and I use an n330 running 64bit as a workstation). I also have an acer aspire one that's running arch linux.
Originally Posted by Svartalf
I'm pretty sure there's a bit more to it than those settings. GCC probably doesn't target the whole range of optimizations needed to actually "target" the LPIA microcode type of x86 machines, based on that link you provided. It's interesting that they're reccomending -O3 (I wouldn't in all cases...) as well as using the Nocona optimizations.
Well being pretty sure is pretty hand wavy again. The link I found there was posted by an intel employee for their suggested LPIA compile flags. I did some searching and came up with nothing more.
I'm speculating the nacona is a more basic architecture than the core2 with less reliance on out of order optimizations, which may incidentally run better on the in order atom.
Please could you elaborate on how did you get those graphs with phoronix-test-suite? I would like to repeat the experiment, but I don't get any graph generated in my report. Any special option has to be enabled in user-config.xml?