-O4 a.k.a. link-time optimization
-O4 a.k.a. link-time optimization
- Subject: -O4 a.k.a. link-time optimization
- From: Glen Low <email@hidden>
- Date: Sat, 04 Jun 2011 19:43:34 +0800
I'm exploring ways to make my app Instaviz run faster. In Xcode 4.0.2
there's the option to use the clang 2.0 compiler which appears much
more stable and performant than Xcode 3.2.x's clang 1.5/1.6. There's
also the option to turn on link-time optimization with -O4, and my
initial tests with x64 code seem good.
I'd like any tips/experience/caveats with using link-time optimization.
1. Does it actually work on armv6/armv7/iOS builds as of Xcode
4.0.2?
2. Since -O4 apparently produces LLVM object files, can I freely
link both LLVM and native object files to produce a native app? i.e.
do -O4 on some compiles and leave it out on other compiles?
3. On command-line builds do I have to pass anything special to
the ld linker? Or does it automatically do LTO when it sees LLVM
object files?
4. Is it still incompatible with debug symbols? What is the end-
user result of shipping an app with LTO -- will the crash reports I
get back from users be unsymbolicatable??
Cheers,
Glen Low
Pixelglow Software
http://instaviz.com
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden