[llvm-branch-commits] [llvm-branch] r223534 - Merging r223170:
Michael Zolotukhin
mzolotukhin at apple.com
Fri Dec 5 14:29:16 PST 2014
Author: mzolotukhin
Date: Fri Dec 5 16:29:16 2014
New Revision: 223534
URL: http://llvm.org/viewvc/llvm-project?rev=223534&view=rev
Log:
Merging r223170:
Apply loop-rotate to several vectorizer tests.
Such loops shouldn't be vectorized due to the loops form.
After applying loop-rotate (+simplifycfg) the tests again start to check
what they are intended to check.
Modified:
llvm/branches/release_35/ (props changed)
llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-address-space.ll
llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-readonly-address-space.ll
llvm/branches/release_35/test/Transforms/LoopVectorize/vect.stats.ll
Propchange: llvm/branches/release_35/
------------------------------------------------------------------------------
--- svn:mergeinfo (original)
+++ svn:mergeinfo Fri Dec 5 16:29:16 2014
@@ -1,3 +1,3 @@
/llvm/branches/Apple/Pertwee:110850,110961
/llvm/branches/type-system-rewrite:133420-134817
-/llvm/trunk:155241,213653,213665,213726,213749,213773,213793,213798-213799,213815,213847,213880,213883-213884,213894-213896,213899,213915,213966,213999,214060,214129,214180,214287,214331,214423,214429,214519,214670,214674,214679,215685,215711,215806,216064,216262,216531,216920,217257,218745,221009,221408,221453,223163
+/llvm/trunk:155241,213653,213665,213726,213749,213773,213793,213798-213799,213815,213847,213880,213883-213884,213894-213896,213899,213915,213966,213999,214060,214129,214180,214287,214331,214423,214429,214519,214670,214674,214679,215685,215711,215806,216064,216262,216531,216920,217257,218745,221009,221408,221453,223163,223170
Modified: llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-address-space.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-address-space.ll?rev=223534&r1=223533&r2=223534&view=diff
==============================================================================
--- llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-address-space.ll (original)
+++ llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-address-space.ll Fri Dec 5 16:29:16 2014
@@ -31,25 +31,23 @@ define void @foo(i32 addrspace(1)* %a, i
; CHECK: ret
entry:
- br label %for.cond
+ %cmp1 = icmp slt i32 0, %n
+ br i1 %cmp1, label %for.body, label %for.end
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp slt i32 %i.0, %n
- br i1 %cmp, label %for.body, label %for.end
-
-for.body: ; preds = %for.cond
- %idxprom = sext i32 %i.0 to i64
+for.body: ; preds = %entry, %for.body
+ %i.02 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+ %idxprom = sext i32 %i.02 to i64
%arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %idxprom
%0 = load i32 addrspace(1)* %arrayidx, align 4
%mul = mul nsw i32 %0, 3
- %idxprom1 = sext i32 %i.0 to i64
+ %idxprom1 = sext i32 %i.02 to i64
%arrayidx2 = getelementptr inbounds i32 addrspace(1)* %a, i64 %idxprom1
store i32 %mul, i32 addrspace(1)* %arrayidx2, align 4
- %inc = add nsw i32 %i.0, 1
- br label %for.cond
+ %inc = add nsw i32 %i.02, 1
+ %cmp = icmp slt i32 %inc, %n
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body, %entry
ret void
}
@@ -60,25 +58,23 @@ define void @bar0(i32* %a, i32 addrspace
; CHECK: ret
entry:
- br label %for.cond
+ %cmp1 = icmp slt i32 0, %n
+ br i1 %cmp1, label %for.body, label %for.end
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp slt i32 %i.0, %n
- br i1 %cmp, label %for.body, label %for.end
-
-for.body: ; preds = %for.cond
- %idxprom = sext i32 %i.0 to i64
+for.body: ; preds = %entry, %for.body
+ %i.02 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+ %idxprom = sext i32 %i.02 to i64
%arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %idxprom
%0 = load i32 addrspace(1)* %arrayidx, align 4
%mul = mul nsw i32 %0, 3
- %idxprom1 = sext i32 %i.0 to i64
+ %idxprom1 = sext i32 %i.02 to i64
%arrayidx2 = getelementptr inbounds i32* %a, i64 %idxprom1
store i32 %mul, i32* %arrayidx2, align 4
- %inc = add nsw i32 %i.0, 1
- br label %for.cond
+ %inc = add nsw i32 %i.02, 1
+ %cmp = icmp slt i32 %inc, %n
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body, %entry
ret void
}
@@ -89,25 +85,23 @@ define void @bar1(i32 addrspace(1)* %a,
; CHECK: ret
entry:
- br label %for.cond
+ %cmp1 = icmp slt i32 0, %n
+ br i1 %cmp1, label %for.body, label %for.end
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp slt i32 %i.0, %n
- br i1 %cmp, label %for.body, label %for.end
-
-for.body: ; preds = %for.cond
- %idxprom = sext i32 %i.0 to i64
+for.body: ; preds = %entry, %for.body
+ %i.02 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+ %idxprom = sext i32 %i.02 to i64
%arrayidx = getelementptr inbounds i32* %b, i64 %idxprom
%0 = load i32* %arrayidx, align 4
%mul = mul nsw i32 %0, 3
- %idxprom1 = sext i32 %i.0 to i64
+ %idxprom1 = sext i32 %i.02 to i64
%arrayidx2 = getelementptr inbounds i32 addrspace(1)* %a, i64 %idxprom1
store i32 %mul, i32 addrspace(1)* %arrayidx2, align 4
- %inc = add nsw i32 %i.0, 1
- br label %for.cond
+ %inc = add nsw i32 %i.02, 1
+ %cmp = icmp slt i32 %inc, %n
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body, %entry
ret void
}
@@ -119,25 +113,23 @@ define void @bar2(i32* noalias %a, i32 a
; CHECK: ret
entry:
- br label %for.cond
-
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp slt i32 %i.0, %n
- br i1 %cmp, label %for.body, label %for.end
+ %cmp1 = icmp slt i32 0, %n
+ br i1 %cmp1, label %for.body, label %for.end
-for.body: ; preds = %for.cond
- %idxprom = sext i32 %i.0 to i64
+for.body: ; preds = %entry, %for.body
+ %i.02 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+ %idxprom = sext i32 %i.02 to i64
%arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %idxprom
%0 = load i32 addrspace(1)* %arrayidx, align 4
%mul = mul nsw i32 %0, 3
- %idxprom1 = sext i32 %i.0 to i64
+ %idxprom1 = sext i32 %i.02 to i64
%arrayidx2 = getelementptr inbounds i32* %a, i64 %idxprom1
store i32 %mul, i32* %arrayidx2, align 4
- %inc = add nsw i32 %i.0, 1
- br label %for.cond
+ %inc = add nsw i32 %i.02, 1
+ %cmp = icmp slt i32 %inc, %n
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body, %entry
ret void
}
@@ -149,25 +141,23 @@ define void @arst0(i32* %b, i32 %n) #0 {
; CHECK: ret
entry:
- br label %for.cond
-
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp slt i32 %i.0, %n
- br i1 %cmp, label %for.body, label %for.end
+ %cmp1 = icmp slt i32 0, %n
+ br i1 %cmp1, label %for.body, label %for.end
-for.body: ; preds = %for.cond
- %idxprom = sext i32 %i.0 to i64
+for.body: ; preds = %entry, %for.body
+ %i.02 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+ %idxprom = sext i32 %i.02 to i64
%arrayidx = getelementptr inbounds i32* %b, i64 %idxprom
%0 = load i32* %arrayidx, align 4
%mul = mul nsw i32 %0, 3
- %idxprom1 = sext i32 %i.0 to i64
+ %idxprom1 = sext i32 %i.02 to i64
%arrayidx2 = getelementptr inbounds [1024 x i32] addrspace(1)* @g_as1, i64 0, i64 %idxprom1
store i32 %mul, i32 addrspace(1)* %arrayidx2, align 4
- %inc = add nsw i32 %i.0, 1
- br label %for.cond
+ %inc = add nsw i32 %i.02, 1
+ %cmp = icmp slt i32 %inc, %n
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body, %entry
ret void
}
@@ -180,25 +170,23 @@ define void @arst1(i32* %b, i32 %n) #0 {
; CHECK: ret
entry:
- br label %for.cond
-
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp slt i32 %i.0, %n
- br i1 %cmp, label %for.body, label %for.end
+ %cmp1 = icmp slt i32 0, %n
+ br i1 %cmp1, label %for.body, label %for.end
-for.body: ; preds = %for.cond
- %idxprom = sext i32 %i.0 to i64
+for.body: ; preds = %entry, %for.body
+ %i.02 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+ %idxprom = sext i32 %i.02 to i64
%arrayidx = getelementptr inbounds [1024 x i32] addrspace(1)* @g_as1, i64 0, i64 %idxprom
%0 = load i32 addrspace(1)* %arrayidx, align 4
%mul = mul nsw i32 %0, 3
- %idxprom1 = sext i32 %i.0 to i64
+ %idxprom1 = sext i32 %i.02 to i64
%arrayidx2 = getelementptr inbounds i32* %b, i64 %idxprom1
store i32 %mul, i32* %arrayidx2, align 4
- %inc = add nsw i32 %i.0, 1
- br label %for.cond
+ %inc = add nsw i32 %i.02, 1
+ %cmp = icmp slt i32 %inc, %n
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body, %entry
ret void
}
@@ -210,25 +198,23 @@ define void @aoeu(i32 %n) #0 {
; CHECK: ret
entry:
- br label %for.cond
+ %cmp1 = icmp slt i32 0, %n
+ br i1 %cmp1, label %for.body, label %for.end
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp slt i32 %i.0, %n
- br i1 %cmp, label %for.body, label %for.end
-
-for.body: ; preds = %for.cond
- %idxprom = sext i32 %i.0 to i64
+for.body: ; preds = %entry, %for.body
+ %i.02 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+ %idxprom = sext i32 %i.02 to i64
%arrayidx = getelementptr inbounds [1024 x i32] addrspace(2)* @q_as2, i64 0, i64 %idxprom
%0 = load i32 addrspace(2)* %arrayidx, align 4
%mul = mul nsw i32 %0, 3
- %idxprom1 = sext i32 %i.0 to i64
+ %idxprom1 = sext i32 %i.02 to i64
%arrayidx2 = getelementptr inbounds [1024 x i32] addrspace(1)* @g_as1, i64 0, i64 %idxprom1
store i32 %mul, i32 addrspace(1)* %arrayidx2, align 4
- %inc = add nsw i32 %i.0, 1
- br label %for.cond
+ %inc = add nsw i32 %i.02, 1
+ %cmp = icmp slt i32 %inc, %n
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body, %entry
ret void
}
Modified: llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-readonly-address-space.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-readonly-address-space.ll?rev=223534&r1=223533&r2=223534&view=diff
==============================================================================
--- llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-readonly-address-space.ll (original)
+++ llvm/branches/release_35/test/Transforms/LoopVectorize/runtime-check-readonly-address-space.ll Fri Dec 5 16:29:16 2014
@@ -8,26 +8,24 @@ define void @add_ints_1_1_1(i32 addrspac
; CHECK-LABEL: @add_ints_1_1_1(
; CHECK: <4 x i32>
; CHECK: ret
-entry:
- br label %for.cond
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp ult i64 %i.0, 200
- br i1 %cmp, label %for.body, label %for.end
+entry:
+ br label %for.body
-for.body: ; preds = %for.cond
- %arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %i.0
+for.body: ; preds = %entry, %for.body
+ %i.01 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
+ %arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %i.01
%0 = load i32 addrspace(1)* %arrayidx, align 4
- %arrayidx1 = getelementptr inbounds i32 addrspace(1)* %c, i64 %i.0
+ %arrayidx1 = getelementptr inbounds i32 addrspace(1)* %c, i64 %i.01
%1 = load i32 addrspace(1)* %arrayidx1, align 4
%add = add nsw i32 %0, %1
- %arrayidx2 = getelementptr inbounds i32 addrspace(1)* %a, i64 %i.0
+ %arrayidx2 = getelementptr inbounds i32 addrspace(1)* %a, i64 %i.01
store i32 %add, i32 addrspace(1)* %arrayidx2, align 4
- %inc = add i64 %i.0, 1
- br label %for.cond
+ %inc = add i64 %i.01, 1
+ %cmp = icmp ult i64 %inc, 200
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body
ret void
}
@@ -35,26 +33,24 @@ define void @add_ints_as_1_0_0(i32 addrs
; CHECK-LABEL: @add_ints_as_1_0_0(
; CHECK-NOT: <4 x i32>
; CHECK: ret
-entry:
- br label %for.cond
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp ult i64 %i.0, 200
- br i1 %cmp, label %for.body, label %for.end
+entry:
+ br label %for.body
-for.body: ; preds = %for.cond
- %arrayidx = getelementptr inbounds i32* %b, i64 %i.0
+for.body: ; preds = %entry, %for.body
+ %i.01 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
+ %arrayidx = getelementptr inbounds i32* %b, i64 %i.01
%0 = load i32* %arrayidx, align 4
- %arrayidx1 = getelementptr inbounds i32* %c, i64 %i.0
+ %arrayidx1 = getelementptr inbounds i32* %c, i64 %i.01
%1 = load i32* %arrayidx1, align 4
%add = add nsw i32 %0, %1
- %arrayidx2 = getelementptr inbounds i32 addrspace(1)* %a, i64 %i.0
+ %arrayidx2 = getelementptr inbounds i32 addrspace(1)* %a, i64 %i.01
store i32 %add, i32 addrspace(1)* %arrayidx2, align 4
- %inc = add i64 %i.0, 1
- br label %for.cond
+ %inc = add i64 %i.01, 1
+ %cmp = icmp ult i64 %inc, 200
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body
ret void
}
@@ -62,26 +58,24 @@ define void @add_ints_as_0_1_0(i32* %a,
; CHECK-LABEL: @add_ints_as_0_1_0(
; CHECK-NOT: <4 x i32>
; CHECK: ret
-entry:
- br label %for.cond
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp ult i64 %i.0, 200
- br i1 %cmp, label %for.body, label %for.end
+entry:
+ br label %for.body
-for.body: ; preds = %for.cond
- %arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %i.0
+for.body: ; preds = %entry, %for.body
+ %i.01 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
+ %arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %i.01
%0 = load i32 addrspace(1)* %arrayidx, align 4
- %arrayidx1 = getelementptr inbounds i32* %c, i64 %i.0
+ %arrayidx1 = getelementptr inbounds i32* %c, i64 %i.01
%1 = load i32* %arrayidx1, align 4
%add = add nsw i32 %0, %1
- %arrayidx2 = getelementptr inbounds i32* %a, i64 %i.0
+ %arrayidx2 = getelementptr inbounds i32* %a, i64 %i.01
store i32 %add, i32* %arrayidx2, align 4
- %inc = add i64 %i.0, 1
- br label %for.cond
+ %inc = add i64 %i.01, 1
+ %cmp = icmp ult i64 %inc, 200
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body
ret void
}
@@ -89,26 +83,24 @@ define void @add_ints_as_0_1_1(i32* %a,
; CHECK-LABEL: @add_ints_as_0_1_1(
; CHECK-NOT: <4 x i32>
; CHECK: ret
-entry:
- br label %for.cond
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp ult i64 %i.0, 200
- br i1 %cmp, label %for.body, label %for.end
+entry:
+ br label %for.body
-for.body: ; preds = %for.cond
- %arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %i.0
+for.body: ; preds = %entry, %for.body
+ %i.01 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
+ %arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %i.01
%0 = load i32 addrspace(1)* %arrayidx, align 4
- %arrayidx1 = getelementptr inbounds i32 addrspace(1)* %c, i64 %i.0
+ %arrayidx1 = getelementptr inbounds i32 addrspace(1)* %c, i64 %i.01
%1 = load i32 addrspace(1)* %arrayidx1, align 4
%add = add nsw i32 %0, %1
- %arrayidx2 = getelementptr inbounds i32* %a, i64 %i.0
+ %arrayidx2 = getelementptr inbounds i32* %a, i64 %i.01
store i32 %add, i32* %arrayidx2, align 4
- %inc = add i64 %i.0, 1
- br label %for.cond
+ %inc = add i64 %i.01, 1
+ %cmp = icmp ult i64 %inc, 200
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body
ret void
}
@@ -116,26 +108,24 @@ define void @add_ints_as_0_1_2(i32* %a,
; CHECK-LABEL: @add_ints_as_0_1_2(
; CHECK-NOT: <4 x i32>
; CHECK: ret
-entry:
- br label %for.cond
-for.cond: ; preds = %for.body, %entry
- %i.0 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
- %cmp = icmp ult i64 %i.0, 200
- br i1 %cmp, label %for.body, label %for.end
+entry:
+ br label %for.body
-for.body: ; preds = %for.cond
- %arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %i.0
+for.body: ; preds = %entry, %for.body
+ %i.01 = phi i64 [ 0, %entry ], [ %inc, %for.body ]
+ %arrayidx = getelementptr inbounds i32 addrspace(1)* %b, i64 %i.01
%0 = load i32 addrspace(1)* %arrayidx, align 4
- %arrayidx1 = getelementptr inbounds i32 addrspace(2)* %c, i64 %i.0
+ %arrayidx1 = getelementptr inbounds i32 addrspace(2)* %c, i64 %i.01
%1 = load i32 addrspace(2)* %arrayidx1, align 4
%add = add nsw i32 %0, %1
- %arrayidx2 = getelementptr inbounds i32* %a, i64 %i.0
+ %arrayidx2 = getelementptr inbounds i32* %a, i64 %i.01
store i32 %add, i32* %arrayidx2, align 4
- %inc = add i64 %i.0, 1
- br label %for.cond
+ %inc = add i64 %i.01, 1
+ %cmp = icmp ult i64 %inc, 200
+ br i1 %cmp, label %for.body, label %for.end
-for.end: ; preds = %for.cond
+for.end: ; preds = %for.body
ret void
}
Modified: llvm/branches/release_35/test/Transforms/LoopVectorize/vect.stats.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/branches/release_35/test/Transforms/LoopVectorize/vect.stats.ll?rev=223534&r1=223533&r2=223534&view=diff
==============================================================================
--- llvm/branches/release_35/test/Transforms/LoopVectorize/vect.stats.ll (original)
+++ llvm/branches/release_35/test/Transforms/LoopVectorize/vect.stats.ll Fri Dec 5 16:29:16 2014
@@ -13,53 +13,47 @@ target triple = "x86_64-unknown-linux-gn
define void @vectorized(float* nocapture %a, i64 %size) {
entry:
- %cmp1 = icmp sgt i64 %size, 0
- br i1 %cmp1, label %for.header, label %for.end
-
-for.header:
- %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
- %cmp2 = icmp sgt i64 %indvars.iv, %size
- br i1 %cmp2, label %for.end, label %for.body
-
-for.body:
-
- %arrayidx = getelementptr inbounds float* %a, i64 %indvars.iv
+ %cmp1 = icmp sle i64 %size, 0
+ %cmp21 = icmp sgt i64 0, %size
+ %or.cond = or i1 %cmp1, %cmp21
+ br i1 %or.cond, label %for.end, label %for.body
+
+for.body: ; preds = %entry, %for.body
+ %indvars.iv2 = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
+ %arrayidx = getelementptr inbounds float* %a, i64 %indvars.iv2
%0 = load float* %arrayidx, align 4
%mul = fmul float %0, %0
store float %mul, float* %arrayidx, align 4
+ %indvars.iv.next = add nuw nsw i64 %indvars.iv2, 1
+ %cmp2 = icmp sgt i64 %indvars.iv.next, %size
+ br i1 %cmp2, label %for.end, label %for.body
- %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
- br label %for.header
-
-for.end:
+for.end: ; preds = %entry, %for.body
ret void
}
define void @not_vectorized(float* nocapture %a, i64 %size) {
entry:
- %cmp1 = icmp sgt i64 %size, 0
- br i1 %cmp1, label %for.header, label %for.end
-
-for.header:
- %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
- %cmp2 = icmp sgt i64 %indvars.iv, %size
- br i1 %cmp2, label %for.end, label %for.body
-
-for.body:
-
- %0 = add nsw i64 %indvars.iv, -5
+ %cmp1 = icmp sle i64 %size, 0
+ %cmp21 = icmp sgt i64 0, %size
+ %or.cond = or i1 %cmp1, %cmp21
+ br i1 %or.cond, label %for.end, label %for.body
+
+for.body: ; preds = %entry, %for.body
+ %indvars.iv2 = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
+ %0 = add nsw i64 %indvars.iv2, -5
%arrayidx = getelementptr inbounds float* %a, i64 %0
%1 = load float* %arrayidx, align 4
- %2 = add nsw i64 %indvars.iv, 2
+ %2 = add nsw i64 %indvars.iv2, 2
%arrayidx2 = getelementptr inbounds float* %a, i64 %2
%3 = load float* %arrayidx2, align 4
%mul = fmul float %1, %3
- %arrayidx4 = getelementptr inbounds float* %a, i64 %indvars.iv
+ %arrayidx4 = getelementptr inbounds float* %a, i64 %indvars.iv2
store float %mul, float* %arrayidx4, align 4
+ %indvars.iv.next = add nuw nsw i64 %indvars.iv2, 1
+ %cmp2 = icmp sgt i64 %indvars.iv.next, %size
+ br i1 %cmp2, label %for.end, label %for.body
- %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
- br label %for.header
-
-for.end:
+for.end: ; preds = %entry, %for.body
ret void
}
\ No newline at end of file
More information about the llvm-branch-commits
mailing list