• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1; REQUIRES: asserts
2; RUN: llc -march=hexagon -mcpu=hexagonv5 -disable-hsdr --stats -o - 2>&1 < %s | FileCheck %s
3
4; Check that store is post-incremented.
5; CHECK-NOT: extractu(r{{[0-9]+}},#32,
6; CHECK-NOT: insert
7; CHECK-NOT: Number of node order issues found
8; CHECK: Number of loops software pipelined
9; CHECK-NOT: Number of node order issues found
10target datalayout = "e-p:32:32:32-i64:64:64-i32:32:32-i16:16:16-i1:32:32-f64:64:64-f32:32:32-v64:64:64-v32:32:32-a0:0-n16:32"
11target triple = "hexagon"
12
13define i32 @foo(i16* noalias nocapture %src, i16* noalias nocapture %dstImg, i32 %width, i32 %idx, i32 %flush) #0 {
14entry:
15  %0 = tail call i64 @llvm.hexagon.A2.combinew(i32 %flush, i32 %flush)
16  %1 = bitcast i64 %0 to <2 x i32>
17  br label %polly.loop_body
18
19polly.loop_after:                                 ; preds = %polly.loop_body
20  ret i32 0
21
22polly.loop_body:                                  ; preds = %entry, %polly.loop_body
23  %p_arrayidx35.phi = phi i16* [ %dstImg, %entry ], [ %p_arrayidx35.inc, %polly.loop_body ]
24  %p_arrayidx.phi = phi i16* [ %src, %entry ], [ %p_arrayidx.inc, %polly.loop_body ]
25  %polly.loopiv56 = phi i32 [ 0, %entry ], [ %polly.next_loopiv, %polly.loop_body ]
26  %polly.next_loopiv = add nsw i32 %polly.loopiv56, 4
27  %vector_ptr = bitcast i16* %p_arrayidx.phi to <4 x i16>*
28  %_p_vec_full = load <4 x i16>, <4 x i16>* %vector_ptr, align 2
29  %_high_half = shufflevector <4 x i16> %_p_vec_full, <4 x i16> undef, <2 x i32> <i32 2, i32 3>
30  %_low_half = shufflevector <4 x i16> %_p_vec_full, <4 x i16> undef, <2 x i32> <i32 0, i32 1>
31  %2 = zext <2 x i16> %_low_half to <2 x i32>
32  %3 = zext <2 x i16> %_high_half to <2 x i32>
33  %add33p_vec = add <2 x i32> %2, %1
34  %add33p_vec48 = add <2 x i32> %3, %1
35  %4 = trunc <2 x i32> %add33p_vec to <2 x i16>
36  %5 = trunc <2 x i32> %add33p_vec48 to <2 x i16>
37  %_combined_vec = shufflevector <2 x i16> %4, <2 x i16> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
38  %vector_ptr49 = bitcast i16* %p_arrayidx35.phi to <4 x i16>*
39  store <4 x i16> %_combined_vec, <4 x i16>* %vector_ptr49, align 2
40  %6 = icmp slt i32 %polly.next_loopiv, 1024
41  %p_arrayidx35.inc = getelementptr i16, i16* %p_arrayidx35.phi, i32 4
42  %p_arrayidx.inc = getelementptr i16, i16* %p_arrayidx.phi, i32 4
43  br i1 %6, label %polly.loop_body, label %polly.loop_after
44}
45
46declare i64 @llvm.hexagon.A2.combinew(i32, i32) #1
47
48attributes #0 = { nounwind "less-precise-fpmad"="false" "frame-pointer"="non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "unsafe-fp-math"="false" "use-soft-float"="false" }
49attributes #1 = { nounwind readnone }
50
51
52