Re: [tree-ssa] decls in tree-ssa (fwd)

[law at redhat dot com]

In message <20030210180736.GA24649@tornado.toronto.redhat.com>, Diego Novillo w
rites:
 >Dan sent me this analysis this morning.  Both problems are
 >to be expected: (1) the gimplifier creates a truckload of
 >VAR_DECLs, (2) the RTL expander is emitting lots of RTL.
Yes.  We're going to have to dive into these at some point in the not
too distant future.

 >> Another thing is the RTL generated when using SSA: 
 >> 
 >> wc -l combine.i.00.rtl ->  235816 
 >> when not using the SSA optimizations:
 >> wc -l combine.i.00.rtl ->  179817
 >> 
 >> in the SSA case there's a lot of RTL that looks like this:
 >> 
 >> (barrier 17 16 18)
 >> 
 >> (jump_insn 18 17 19 (nil) (set (pc)
 >>         (label_ref 21)) -1 (nil)
 >>     (nil))
 >> 
 >> (barrier 19 18 20)
 >> 
 >> (code_label 20 19 21 2 "" [0 uses])
 >> 
 >> (code_label 21 20 23 3 "" [0 uses])
 >> 
 >>         
 >> I estimate that this type of RTL accounts for about 50% of the size
 >> increase of the 00.rtl file generated in the SSA case.
 >> Maybe there's a way to avoid generating this type of RTL...
 >> 
 >> I thought you might want to know about this stuff.
Yup.  I've run into this stuff as well.  It's rather annoying.  At some
level it also plays into the regression using the tree-ssa code for this
test:

FAIL: gcc.dg/i386-cadd.c scan-assembler sbb

We may be able to clean some of this up, some might have to hang around
(unless we have a step where we completely linearize the tree structures
at which point we can correct some of this sillyness).

Jeff