This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[patch] function inlining when optimize for size (PR11830)


I have examined if it is possible to use the function inlining when
optimizing for size. I've found we can use it without any risk because
when the max-inline-insns-* parameters were set with a small value
the function inlining doesn't inline large functions.

It is also fix the PR11830 (

I measuerd the size effect on CSiBE and this patch made about 1% code
size save (arm:1.212%, i386:1.024%, i686:1.029%, m68k:0.795%,
mips:1.096%, ppc:1.106%).

Bootstrapped and regtested on i686-pc-linux-gnu and regtested also on
{arm,mips,ppc}-elf with no new failures.

Is it OK for mainline?

Regards, Gábor Lóki

2004-03-17 Gábor Lóki <>

	* opts.c (decode_options): Do function inlining with very small
	max-inline-insns-* parameters when optimizing for size

Index: gcc/gcc/opts.c
RCS file: /cvs/gcc/gcc/gcc/opts.c,v
retrieving revision 1.61
diff -u -r1.61 opts.c
--- gcc/gcc/opts.c	10 Mar 2004 06:02:53 -0000	1.61
+++ gcc/gcc/opts.c	16 Mar 2004 16:24:29 -0000
@@ -593,6 +593,16 @@
 	 use more short jumps instead of long jumps.  */
       flag_reorder_blocks = 0;
+  if (optimize_size)
+    {
+      /* Do the functions inlining with very small max-inline-insns-*. 
+         handle_options could overwrite if it's necessary.  */
+      set_param_value ("max-inline-insns-single", 5);
+      set_param_value ("max-inline-insns-auto", 5);
+      set_param_value ("max-inline-insns-rtl", 10);
+      flag_inline_functions = 1;
+    }
   /* Initialize whether `char' is signed.  */
   flag_signed_char = DEFAULT_SIGNED_CHAR;

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]