This adds in some code (currently disabled unless you pass
-enable-smarter-addr-folding to llc) that gives CGP a better
cost model for when to sink computations into addressing modes.
The basic observation is that sinking increases register
pressure when part of the addr computation has to be available
for other reasons, such as having a use that is a non-memory
operation. In cases where it works, it can substantially reduce
register pressure.
This code is currently an overall win on 403.gcc and 255.vortex
(the two things I've been looking at), but there are several
things I want to do before enabling it by default:
1. This isn't doing any caching of results, so it is much slower
than it could be. It currently slows down release-asserts llc
by 1.7% on 176.gcc: 27.12s -> 27.60s.
2. This doesn't think about inline asm memory operands yet.
3. The cost model botches the case when the needed value is live
across the computation for other reasons.
I'll continue poking at this, and eventually turn it on as llcbeta.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60074 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/test/CodeGen/X86/isel-sink3.ll b/test/CodeGen/X86/isel-sink3.ll
new file mode 100644
index 0000000..a0fba3a
--- /dev/null
+++ b/test/CodeGen/X86/isel-sink3.ll
@@ -0,0 +1,25 @@
+; RUN: llvm-as < %s | llc -enable-smarter-addr-folding | grep {addl.(%eax), %ecx}
+; RUN: llvm-as < %s | llc -enable-smarter-addr-folding | not grep leal
+; this should not sink %1 into bb1, that would increase reg pressure.
+
+; rdar://6399178
+
+target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
+target triple = "i386-apple-darwin7"
+
+define i32 @bar(i32** %P) nounwind {
+entry:
+ %0 = load i32** %P, align 4 ; <i32*> [#uses=2]
+ %1 = getelementptr i32* %0, i32 1 ; <i32*> [#uses=1]
+ %2 = icmp ugt i32* %1, inttoptr (i64 1233 to i32*) ; <i1> [#uses=1]
+ br i1 %2, label %bb1, label %bb
+
+bb: ; preds = %entry
+ store i32* inttoptr (i64 123 to i32*), i32** %P, align 4
+ br label %bb1
+
+bb1: ; preds = %entry, %bb
+ %3 = getelementptr i32* %1, i32 1 ; <i32*> [#uses=1]
+ %4 = load i32* %3, align 4 ; <i32> [#uses=1]
+ ret i32 %4
+}