From: Chen Zheng <czhengsz@cn.ibm.com>
Date: Tue, 4 Jun 2019 06:48:14 +0000 (+0000)
Subject: [PowerPC] add testcases for reordering LSR and PPCCTRLoops - NFC
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=477a7cd43583aa11f51daa1f4605cc14c3f5a99f;p=llvm

[PowerPC] add testcases for reordering LSR and PPCCTRLoops - NFC


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@362468 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/test/CodeGen/PowerPC/lsr-ctrloop.ll b/test/CodeGen/PowerPC/lsr-ctrloop.ll
new file mode 100644
index 00000000000..14fd4c96dea
--- /dev/null
+++ b/test/CodeGen/PowerPC/lsr-ctrloop.ll
@@ -0,0 +1,217 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mcpu=pwr9 -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu < %s | FileCheck %s
+
+; void foo(float *data, float d) {
+;   long i;
+;   for (i = 0; i < 8000; i++)
+;     data[i] = d;
+; }
+; 
+; This loop will be unrolled by 96 and vectorized on power9.
+; icmp for loop iteration index and loop trip count(384) has LSRUse for 'reg({0,+,384})'.
+; Make sure above icmp does not impact LSR choose best formulae sets based on 'reg({(192 + %0),+,384})' 
+
+define void @foo(float* nocapture %data, float %d) {
+; CHECK-LABEL: foo:
+; CHECK:  .LBB0_1: # %vector.body
+; CHECK:         add 5, 3, 4
+; CHECK-NEXT:    stxvx 0, 3, 4
+; CHECK-NEXT:    addi 4, 4, 384
+; CHECK-NEXT:    stxv 0, 16(5)
+; CHECK-NEXT:    stxv 0, 32(5)
+; CHECK-NEXT:    stxv 0, 48(5)
+; CHECK-NEXT:    stxv 0, 64(5)
+; CHECK-NEXT:    stxv 0, 80(5)
+; CHECK-NEXT:    stxv 0, 96(5)
+; CHECK-NEXT:    stxv 0, 112(5)
+; CHECK-NEXT:    stxv 0, 128(5)
+; CHECK-NEXT:    stxv 0, 144(5)
+; CHECK-NEXT:    stxv 0, 160(5)
+; CHECK-NEXT:    stxv 0, 176(5)
+; CHECK-NEXT:    stxv 0, 192(5)
+; CHECK-NEXT:    stxv 0, 208(5)
+; CHECK-NEXT:    stxv 0, 224(5)
+; CHECK-NEXT:    stxv 0, 240(5)
+; CHECK-NEXT:    stxv 0, 256(5)
+; CHECK-NEXT:    stxv 0, 272(5)
+; CHECK-NEXT:    stxv 0, 288(5)
+; CHECK-NEXT:    stxv 0, 304(5)
+; CHECK-NEXT:    stxv 0, 320(5)
+; CHECK-NEXT:    stxv 0, 336(5)
+; CHECK-NEXT:    stxv 0, 352(5)
+; CHECK-NEXT:    stxv 0, 368(5)
+; CHECK-NEXT:    bdnz .LBB0_1
+
+entry:
+  %broadcast.splatinsert16 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat17 = shufflevector <4 x float> %broadcast.splatinsert16, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert18 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat19 = shufflevector <4 x float> %broadcast.splatinsert18, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert20 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat21 = shufflevector <4 x float> %broadcast.splatinsert20, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert22 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat23 = shufflevector <4 x float> %broadcast.splatinsert22, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert24 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat25 = shufflevector <4 x float> %broadcast.splatinsert24, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert26 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat27 = shufflevector <4 x float> %broadcast.splatinsert26, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert28 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat29 = shufflevector <4 x float> %broadcast.splatinsert28, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert30 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat31 = shufflevector <4 x float> %broadcast.splatinsert30, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert32 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat33 = shufflevector <4 x float> %broadcast.splatinsert32, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert34 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat35 = shufflevector <4 x float> %broadcast.splatinsert34, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert36 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat37 = shufflevector <4 x float> %broadcast.splatinsert36, <4 x float> undef, <4 x i32> zeroinitializer
+  %broadcast.splatinsert38 = insertelement <4 x float> undef, float %d, i32 0
+  %broadcast.splat39 = shufflevector <4 x float> %broadcast.splatinsert38, <4 x float> undef, <4 x i32> zeroinitializer
+  br label %vector.body
+
+vector.body:                                      ; preds = %vector.body, %entry
+  %index = phi i64 [ 0, %entry ], [ %index.next.1, %vector.body ]
+  %0 = getelementptr inbounds float, float* %data, i64 %index
+  %1 = bitcast float* %0 to <4 x float>*
+  store <4 x float> %broadcast.splat17, <4 x float>* %1, align 4
+  %2 = getelementptr inbounds float, float* %0, i64 4
+  %3 = bitcast float* %2 to <4 x float>*
+  store <4 x float> %broadcast.splat19, <4 x float>* %3, align 4
+  %4 = getelementptr inbounds float, float* %0, i64 8
+  %5 = bitcast float* %4 to <4 x float>*
+  store <4 x float> %broadcast.splat21, <4 x float>* %5, align 4
+  %6 = getelementptr inbounds float, float* %0, i64 12
+  %7 = bitcast float* %6 to <4 x float>*
+  store <4 x float> %broadcast.splat23, <4 x float>* %7, align 4
+  %8 = getelementptr inbounds float, float* %0, i64 16
+  %9 = bitcast float* %8 to <4 x float>*
+  store <4 x float> %broadcast.splat25, <4 x float>* %9, align 4
+  %10 = getelementptr inbounds float, float* %0, i64 20
+  %11 = bitcast float* %10 to <4 x float>*
+  store <4 x float> %broadcast.splat27, <4 x float>* %11, align 4
+  %12 = getelementptr inbounds float, float* %0, i64 24
+  %13 = bitcast float* %12 to <4 x float>*
+  store <4 x float> %broadcast.splat29, <4 x float>* %13, align 4
+  %14 = getelementptr inbounds float, float* %0, i64 28
+  %15 = bitcast float* %14 to <4 x float>*
+  store <4 x float> %broadcast.splat31, <4 x float>* %15, align 4
+  %16 = getelementptr inbounds float, float* %0, i64 32
+  %17 = bitcast float* %16 to <4 x float>*
+  store <4 x float> %broadcast.splat33, <4 x float>* %17, align 4
+  %18 = getelementptr inbounds float, float* %0, i64 36
+  %19 = bitcast float* %18 to <4 x float>*
+  store <4 x float> %broadcast.splat35, <4 x float>* %19, align 4
+  %20 = getelementptr inbounds float, float* %0, i64 40
+  %21 = bitcast float* %20 to <4 x float>*
+  store <4 x float> %broadcast.splat37, <4 x float>* %21, align 4
+  %22 = getelementptr inbounds float, float* %0, i64 44
+  %23 = bitcast float* %22 to <4 x float>*
+  store <4 x float> %broadcast.splat39, <4 x float>* %23, align 4
+  %index.next = add nuw nsw i64 %index, 48
+  %24 = getelementptr inbounds float, float* %data, i64 %index.next
+  %25 = bitcast float* %24 to <4 x float>*
+  store <4 x float> %broadcast.splat17, <4 x float>* %25, align 4
+  %26 = getelementptr inbounds float, float* %24, i64 4
+  %27 = bitcast float* %26 to <4 x float>*
+  store <4 x float> %broadcast.splat19, <4 x float>* %27, align 4
+  %28 = getelementptr inbounds float, float* %24, i64 8
+  %29 = bitcast float* %28 to <4 x float>*
+  store <4 x float> %broadcast.splat21, <4 x float>* %29, align 4
+  %30 = getelementptr inbounds float, float* %24, i64 12
+  %31 = bitcast float* %30 to <4 x float>*
+  store <4 x float> %broadcast.splat23, <4 x float>* %31, align 4
+  %32 = getelementptr inbounds float, float* %24, i64 16
+  %33 = bitcast float* %32 to <4 x float>*
+  store <4 x float> %broadcast.splat25, <4 x float>* %33, align 4
+  %34 = getelementptr inbounds float, float* %24, i64 20
+  %35 = bitcast float* %34 to <4 x float>*
+  store <4 x float> %broadcast.splat27, <4 x float>* %35, align 4
+  %36 = getelementptr inbounds float, float* %24, i64 24
+  %37 = bitcast float* %36 to <4 x float>*
+  store <4 x float> %broadcast.splat29, <4 x float>* %37, align 4
+  %38 = getelementptr inbounds float, float* %24, i64 28
+  %39 = bitcast float* %38 to <4 x float>*
+  store <4 x float> %broadcast.splat31, <4 x float>* %39, align 4
+  %40 = getelementptr inbounds float, float* %24, i64 32
+  %41 = bitcast float* %40 to <4 x float>*
+  store <4 x float> %broadcast.splat33, <4 x float>* %41, align 4
+  %42 = getelementptr inbounds float, float* %24, i64 36
+  %43 = bitcast float* %42 to <4 x float>*
+  store <4 x float> %broadcast.splat35, <4 x float>* %43, align 4
+  %44 = getelementptr inbounds float, float* %24, i64 40
+  %45 = bitcast float* %44 to <4 x float>*
+  store <4 x float> %broadcast.splat37, <4 x float>* %45, align 4
+  %46 = getelementptr inbounds float, float* %24, i64 44
+  %47 = bitcast float* %46 to <4 x float>*
+  store <4 x float> %broadcast.splat39, <4 x float>* %47, align 4
+  %index.next.1 = add nuw nsw i64 %index, 96
+  %48 = icmp eq i64 %index.next.1, 7968
+  br i1 %48, label %for.body, label %vector.body
+
+for.body:                                         ; preds = %vector.body
+  %arrayidx = getelementptr inbounds float, float* %data, i64 7968
+  store float %d, float* %arrayidx, align 4
+  %arrayidx.1 = getelementptr inbounds float, float* %data, i64 7969
+  store float %d, float* %arrayidx.1, align 4
+  %arrayidx.2 = getelementptr inbounds float, float* %data, i64 7970
+  store float %d, float* %arrayidx.2, align 4
+  %arrayidx.3 = getelementptr inbounds float, float* %data, i64 7971
+  store float %d, float* %arrayidx.3, align 4
+  %arrayidx.4 = getelementptr inbounds float, float* %data, i64 7972
+  store float %d, float* %arrayidx.4, align 4
+  %arrayidx.5 = getelementptr inbounds float, float* %data, i64 7973
+  store float %d, float* %arrayidx.5, align 4
+  %arrayidx.6 = getelementptr inbounds float, float* %data, i64 7974
+  store float %d, float* %arrayidx.6, align 4
+  %arrayidx.7 = getelementptr inbounds float, float* %data, i64 7975
+  store float %d, float* %arrayidx.7, align 4
+  %arrayidx.8 = getelementptr inbounds float, float* %data, i64 7976
+  store float %d, float* %arrayidx.8, align 4
+  %arrayidx.9 = getelementptr inbounds float, float* %data, i64 7977
+  store float %d, float* %arrayidx.9, align 4
+  %arrayidx.10 = getelementptr inbounds float, float* %data, i64 7978
+  store float %d, float* %arrayidx.10, align 4
+  %arrayidx.11 = getelementptr inbounds float, float* %data, i64 7979
+  store float %d, float* %arrayidx.11, align 4
+  %arrayidx.12 = getelementptr inbounds float, float* %data, i64 7980
+  store float %d, float* %arrayidx.12, align 4
+  %arrayidx.13 = getelementptr inbounds float, float* %data, i64 7981
+  store float %d, float* %arrayidx.13, align 4
+  %arrayidx.14 = getelementptr inbounds float, float* %data, i64 7982
+  store float %d, float* %arrayidx.14, align 4
+  %arrayidx.15 = getelementptr inbounds float, float* %data, i64 7983
+  store float %d, float* %arrayidx.15, align 4
+  %arrayidx.16 = getelementptr inbounds float, float* %data, i64 7984
+  store float %d, float* %arrayidx.16, align 4
+  %arrayidx.17 = getelementptr inbounds float, float* %data, i64 7985
+  store float %d, float* %arrayidx.17, align 4
+  %arrayidx.18 = getelementptr inbounds float, float* %data, i64 7986
+  store float %d, float* %arrayidx.18, align 4
+  %arrayidx.19 = getelementptr inbounds float, float* %data, i64 7987
+  store float %d, float* %arrayidx.19, align 4
+  %arrayidx.20 = getelementptr inbounds float, float* %data, i64 7988
+  store float %d, float* %arrayidx.20, align 4
+  %arrayidx.21 = getelementptr inbounds float, float* %data, i64 7989
+  store float %d, float* %arrayidx.21, align 4
+  %arrayidx.22 = getelementptr inbounds float, float* %data, i64 7990
+  store float %d, float* %arrayidx.22, align 4
+  %arrayidx.23 = getelementptr inbounds float, float* %data, i64 7991
+  store float %d, float* %arrayidx.23, align 4
+  %arrayidx.24 = getelementptr inbounds float, float* %data, i64 7992
+  store float %d, float* %arrayidx.24, align 4
+  %arrayidx.25 = getelementptr inbounds float, float* %data, i64 7993
+  store float %d, float* %arrayidx.25, align 4
+  %arrayidx.26 = getelementptr inbounds float, float* %data, i64 7994
+  store float %d, float* %arrayidx.26, align 4
+  %arrayidx.27 = getelementptr inbounds float, float* %data, i64 7995
+  store float %d, float* %arrayidx.27, align 4
+  %arrayidx.28 = getelementptr inbounds float, float* %data, i64 7996
+  store float %d, float* %arrayidx.28, align 4
+  %arrayidx.29 = getelementptr inbounds float, float* %data, i64 7997
+  store float %d, float* %arrayidx.29, align 4
+  %arrayidx.30 = getelementptr inbounds float, float* %data, i64 7998
+  store float %d, float* %arrayidx.30, align 4
+  %arrayidx.31 = getelementptr inbounds float, float* %data, i64 7999
+  store float %d, float* %arrayidx.31, align 4
+  ret void
+}