I'm playing around with optimizing compiled packages to run fine on Ångström, but the question isn't really Ångström-specific.

I've seen on Meanie's blog on building Firefox natively that he used the -O4 option in CFLAGS.

I've looked but not found anything substantive on this flag. One Finnish page on the web states that if -On has n higher than 3, then the flag is automatically interpreted as -O3. Is this so? Or is there a difference between these?

that's using the Spinal Tap version of the compiler - everyone else's goes up to O3, but the 'tap' version goes up to O4.


---

I wonder whether anyone will understand this joke?

11

oh man, these geek jokes are way over my head. can someone explain it so some dummy like me can understand it?

If you haven't seen the film "Spinal Tap", the joke won't make any sense.

Didn't get the joke either...

How about the original issue though, is there really an -O4 flag that furthers optimization, or is it just the same as -O3?

Just keep in mind that you're reading from *slow* media into a tiny amount of RAM, so 'optimized' binaries will execute slower in the end. I heard the people at nokia use -Os for certain applications because execution time is IO-limited.

Well then, i have another curiosity question: how would a Firefox build with -O2 optimization (which seems to be very common for compiling Z apps) compared with the one with -O4?
When do we use which level of optimization?
And how do we know? :?

just stick with -O2, -O3 or higer dont always produce faster binaries and somtimes will not compile (as i have found out)

if you really need spped -O640K should be good enough for anyone

What does this one do? And how does it compare in terms of speed, subjectively what you feel when using the compiled app, with -O2 or -O3 ?

Wasn't that MS humor?