[llvm] 9ca9fee - [SystemZ] Don't shrink 64-bit FP constants.
Jonas Paulsson via llvm-commits
llvm-commits at lists.llvm.org
Thu Jan 27 14:15:40 PST 2022
Author: Jonas Paulsson
Date: 2022-01-27T16:14:53-06:00
New Revision: 9ca9fee6e8440d31b121773c9aa9513296a3ca26
URL: https://github.com/llvm/llvm-project/commit/9ca9fee6e8440d31b121773c9aa9513296a3ca26
DIFF: https://github.com/llvm/llvm-project/commit/9ca9fee6e8440d31b121773c9aa9513296a3ca26.diff
LOG: [SystemZ] Don't shrink 64-bit FP constants.
Return false from ShouldShrinkFPConstant(), so that these constants are stored
in their full size on the constant pool, even if they could have been shrunk
and used with an extending load.
This is better since LD is faster than LDE, and it also enables reg/mem opcodes.
Review: Ulrich Weigand
Differential Revision: https://reviews.llvm.org/D117927
Added:
Modified:
llvm/lib/Target/SystemZ/SystemZISelLowering.h
llvm/test/CodeGen/SystemZ/args-07.ll
llvm/test/CodeGen/SystemZ/fp-const-04.ll
llvm/test/CodeGen/SystemZ/fp-conv-10.ll
llvm/test/CodeGen/SystemZ/fp-conv-12.ll
llvm/test/CodeGen/SystemZ/fp-strict-conv-10.ll
llvm/test/CodeGen/SystemZ/fp-strict-conv-12.ll
llvm/test/CodeGen/SystemZ/vector-constrained-fp-intrinsics.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/SystemZ/SystemZISelLowering.h b/llvm/lib/Target/SystemZ/SystemZISelLowering.h
index a8ddb8c62d180..de446f33f5f18 100644
--- a/llvm/lib/Target/SystemZ/SystemZISelLowering.h
+++ b/llvm/lib/Target/SystemZ/SystemZISelLowering.h
@@ -443,6 +443,11 @@ class SystemZTargetLowering : public TargetLowering {
EVT VT) const override;
bool isFPImmLegal(const APFloat &Imm, EVT VT,
bool ForCodeSize) const override;
+ bool ShouldShrinkFPConstant(EVT VT) const override {
+ // Do not shrink 64-bit FP constpool entries since LDEB is slower than
+ // LD, and having the full constant in memory enables reg/mem opcodes.
+ return VT != MVT::f64;
+ }
bool hasInlineStackProbe(MachineFunction &MF) const override;
bool isLegalICmpImmediate(int64_t Imm) const override;
bool isLegalAddImmediate(int64_t Imm) const override;
diff --git a/llvm/test/CodeGen/SystemZ/args-07.ll b/llvm/test/CodeGen/SystemZ/args-07.ll
index 7dfdab6bbe64d..b20707b32a524 100644
--- a/llvm/test/CodeGen/SystemZ/args-07.ll
+++ b/llvm/test/CodeGen/SystemZ/args-07.ll
@@ -31,13 +31,13 @@ define { i64, i64, i64, i64, i64 } @f2() {
define { double, double, double, double } @f3() {
; CHECK-LABEL: f3:
; CHECK: larl [[TMP:%r[0-5]]], .LCPI
-; CHECK: ldeb %f0, 0([[TMP]])
+; CHECK: ld %f0, 0([[TMP]])
; CHECK: larl [[TMP:%r[0-5]]], .LCPI
-; CHECK: ldeb %f2, 0([[TMP]])
+; CHECK: ld %f2, 0([[TMP]])
; CHECK: larl [[TMP:%r[0-5]]], .LCPI
-; CHECK: ldeb %f4, 0([[TMP]])
+; CHECK: ld %f4, 0([[TMP]])
; CHECK: larl [[TMP:%r[0-5]]], .LCPI
-; CHECK: ldeb %f6, 0([[TMP]])
+; CHECK: ld %f6, 0([[TMP]])
; CHECK: br %r14
; SOFT-FLOAT-LABEL: f3:
diff --git a/llvm/test/CodeGen/SystemZ/fp-const-04.ll b/llvm/test/CodeGen/SystemZ/fp-const-04.ll
index 44817c2c67ec3..4e96aac8d748c 100644
--- a/llvm/test/CodeGen/SystemZ/fp-const-04.ll
+++ b/llvm/test/CodeGen/SystemZ/fp-const-04.ll
@@ -1,5 +1,5 @@
-; Test loads of 64-bit floating-point constants that can be represented
-; as 32-bit constants.
+; Test loads of 64-bit floating-point constants that could be represented
+; as 32-bit constants, but should not be.
;
; RUN: llc < %s -mtriple=s390x-linux-gnu | FileCheck %s
; RUN: llc < %s -mtriple=s390x-linux-gnu | FileCheck %s -check-prefix=CONST
@@ -7,9 +7,9 @@
define double @f1() {
; CHECK-LABEL: f1:
; CHECK: larl [[REGISTER:%r[1-5]]], {{.*}}
-; CHECK: ldeb %f0, 0([[REGISTER]])
+; CHECK: ld %f0, 0([[REGISTER]])
; CHECK: br %r14
;
-; CONST: .long 0x3f800001
+; CONST: .quad 0x3ff0000020000000 # double 1.0000001192092896
ret double 0x3ff0000020000000
}
diff --git a/llvm/test/CodeGen/SystemZ/fp-conv-10.ll b/llvm/test/CodeGen/SystemZ/fp-conv-10.ll
index f897743ef11aa..d94ac8bf99939 100644
--- a/llvm/test/CodeGen/SystemZ/fp-conv-10.ll
+++ b/llvm/test/CodeGen/SystemZ/fp-conv-10.ll
@@ -33,7 +33,7 @@ define i32 @f2(double %f) {
; CHECK-LABEL: f2:
; CHECK: # %bb.0:
; CHECK-NEXT: larl %r1, .LCPI1_0
-; CHECK-NEXT: ldeb %f1, 0(%r1)
+; CHECK-NEXT: ld %f1, 0(%r1)
; CHECK-NEXT: cdbr %f0, %f1
; CHECK-NEXT: jnl .LBB1_2
; CHECK-NEXT: # %bb.1:
diff --git a/llvm/test/CodeGen/SystemZ/fp-conv-12.ll b/llvm/test/CodeGen/SystemZ/fp-conv-12.ll
index 91c377fa3e270..ef933ee3049d5 100644
--- a/llvm/test/CodeGen/SystemZ/fp-conv-12.ll
+++ b/llvm/test/CodeGen/SystemZ/fp-conv-12.ll
@@ -32,7 +32,7 @@ define i64 @f2(double %f) {
; CHECK-LABEL: f2:
; CHECK: # %bb.0:
; CHECK-NEXT: larl %r1, .LCPI1_0
-; CHECK-NEXT: ldeb %f1, 0(%r1)
+; CHECK-NEXT: ld %f1, 0(%r1)
; CHECK-NEXT: cdbr %f0, %f1
; CHECK-NEXT: jnl .LBB1_2
; CHECK-NEXT: # %bb.1:
diff --git a/llvm/test/CodeGen/SystemZ/fp-strict-conv-10.ll b/llvm/test/CodeGen/SystemZ/fp-strict-conv-10.ll
index 5327eccdcce6e..e40d90c2c0fbc 100644
--- a/llvm/test/CodeGen/SystemZ/fp-strict-conv-10.ll
+++ b/llvm/test/CodeGen/SystemZ/fp-strict-conv-10.ll
@@ -42,7 +42,7 @@ define i32 @f2(double %f) #0 {
; CHECK-LABEL: f2:
; CHECK: # %bb.0:
; CHECK-NEXT: larl %r1, .LCPI1_0
-; CHECK-NEXT: ldeb %f1, 0(%r1)
+; CHECK-NEXT: ld %f1, 0(%r1)
; CHECK-NEXT: kdbr %f0, %f1
; CHECK-NEXT: jnl .LBB1_2
; CHECK-NEXT: # %bb.1:
diff --git a/llvm/test/CodeGen/SystemZ/fp-strict-conv-12.ll b/llvm/test/CodeGen/SystemZ/fp-strict-conv-12.ll
index e7ed6af330cfe..a9de3ba6ec042 100644
--- a/llvm/test/CodeGen/SystemZ/fp-strict-conv-12.ll
+++ b/llvm/test/CodeGen/SystemZ/fp-strict-conv-12.ll
@@ -41,7 +41,7 @@ define i64 @f2(double %f) #0 {
; CHECK-LABEL: f2:
; CHECK: # %bb.0:
; CHECK-NEXT: larl %r1, .LCPI1_0
-; CHECK-NEXT: ldeb %f1, 0(%r1)
+; CHECK-NEXT: ld %f1, 0(%r1)
; CHECK-NEXT: kdbr %f0, %f1
; CHECK-NEXT: jnl .LBB1_2
; CHECK-NEXT: # %bb.1:
diff --git a/llvm/test/CodeGen/SystemZ/vector-constrained-fp-intrinsics.ll b/llvm/test/CodeGen/SystemZ/vector-constrained-fp-intrinsics.ll
index fcaca553f1e12..e6559381132d7 100644
--- a/llvm/test/CodeGen/SystemZ/vector-constrained-fp-intrinsics.ll
+++ b/llvm/test/CodeGen/SystemZ/vector-constrained-fp-intrinsics.ll
@@ -31,11 +31,11 @@ define <2 x double> @constrained_vector_fdiv_v2f64() #0 {
; S390X-LABEL: constrained_vector_fdiv_v2f64:
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI1_0
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI1_1
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI1_2
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: ddbr %f2, %f1
; S390X-NEXT: ddbr %f0, %f1
; S390X-NEXT: br %r14
@@ -101,11 +101,11 @@ define void @constrained_vector_fdiv_v3f64(<3 x double>* %a) #0 {
; S390X-LABEL: constrained_vector_fdiv_v3f64:
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI3_1
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI3_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI3_0
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ddb %f1, 0(%r2)
; S390X-NEXT: ddb %f0, 8(%r2)
; S390X-NEXT: ddb %f2, 16(%r2)
@@ -117,7 +117,7 @@ define void @constrained_vector_fdiv_v3f64(<3 x double>* %a) #0 {
; SZ13-LABEL: constrained_vector_fdiv_v3f64:
; SZ13: # %bb.0: # %entry
; SZ13-NEXT: larl %r1, .LCPI3_0
-; SZ13-NEXT: ldeb %f1, 0(%r1)
+; SZ13-NEXT: ld %f1, 0(%r1)
; SZ13-NEXT: ddb %f1, 16(%r2)
; SZ13-NEXT: larl %r1, .LCPI3_1
; SZ13-NEXT: vl %v0, 0(%r2), 4
@@ -141,15 +141,15 @@ define <4 x double> @constrained_vector_fdiv_v4f64() #0 {
; S390X-LABEL: constrained_vector_fdiv_v4f64:
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI4_0
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI4_1
-; S390X-NEXT: ldeb %f6, 0(%r1)
+; S390X-NEXT: ld %f6, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI4_2
-; S390X-NEXT: ldeb %f4, 0(%r1)
+; S390X-NEXT: ld %f4, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI4_3
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI4_4
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: ddbr %f6, %f1
; S390X-NEXT: ddbr %f4, %f1
; S390X-NEXT: ddbr %f2, %f1
@@ -231,13 +231,13 @@ define <2 x double> @constrained_vector_frem_v2f64() #0 {
; S390X-NEXT: .cfi_offset %f8, -168
; S390X-NEXT: .cfi_offset %f9, -176
; S390X-NEXT: larl %r1, .LCPI6_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI6_1
-; S390X-NEXT: ldeb %f8, 0(%r1)
+; S390X-NEXT: ld %f8, 0(%r1)
; S390X-NEXT: ldr %f2, %f8
; S390X-NEXT: brasl %r14, fmod at PLT
; S390X-NEXT: larl %r1, .LCPI6_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f9, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: ldr %f2, %f8
@@ -258,7 +258,7 @@ define <2 x double> @constrained_vector_frem_v2f64() #0 {
; SZ13-NEXT: std %f8, 176(%r15) # 8-byte Folded Spill
; SZ13-NEXT: .cfi_offset %f8, -168
; SZ13-NEXT: larl %r1, .LCPI6_0
-; SZ13-NEXT: ldeb %f8, 0(%r1)
+; SZ13-NEXT: ld %f8, 0(%r1)
; SZ13-NEXT: vgmg %v0, 1, 1
; SZ13-NEXT: ldr %f2, %f8
; SZ13-NEXT: brasl %r14, fmod at PLT
@@ -384,17 +384,17 @@ define void @constrained_vector_frem_v3f64(<3 x double>* %a) #0 {
; S390X-NEXT: ld %f8, 0(%r2)
; S390X-NEXT: ld %f2, 16(%r2)
; S390X-NEXT: larl %r1, .LCPI8_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: ld %f9, 8(%r2)
; S390X-NEXT: brasl %r14, fmod at PLT
; S390X-NEXT: larl %r1, .LCPI8_1
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: ldr %f2, %f9
; S390X-NEXT: brasl %r14, fmod at PLT
; S390X-NEXT: larl %r1, .LCPI8_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f9, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: ldr %f2, %f8
@@ -437,7 +437,7 @@ define void @constrained_vector_frem_v3f64(<3 x double>* %a) #0 {
; SZ13-NEXT: vmrhg %v0, %v1, %v0
; SZ13-NEXT: larl %r1, .LCPI8_0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: ldr %f2, %f8
; SZ13-NEXT: brasl %r14, fmod at PLT
; SZ13-NEXT: std %f0, 16(%r13)
@@ -474,25 +474,25 @@ define <4 x double> @constrained_vector_frem_v4f64() #0 {
; S390X-NEXT: .cfi_offset %f10, -184
; S390X-NEXT: .cfi_offset %f11, -192
; S390X-NEXT: larl %r1, .LCPI9_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI9_1
-; S390X-NEXT: ldeb %f8, 0(%r1)
+; S390X-NEXT: ld %f8, 0(%r1)
; S390X-NEXT: ldr %f2, %f8
; S390X-NEXT: brasl %r14, fmod at PLT
; S390X-NEXT: larl %r1, .LCPI9_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f9, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: ldr %f2, %f8
; S390X-NEXT: brasl %r14, fmod at PLT
; S390X-NEXT: larl %r1, .LCPI9_3
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: ldr %f2, %f8
; S390X-NEXT: brasl %r14, fmod at PLT
; S390X-NEXT: larl %r1, .LCPI9_4
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f11, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: ldr %f2, %f8
@@ -517,7 +517,7 @@ define <4 x double> @constrained_vector_frem_v4f64() #0 {
; SZ13-NEXT: std %f8, 192(%r15) # 8-byte Folded Spill
; SZ13-NEXT: .cfi_offset %f8, -168
; SZ13-NEXT: larl %r1, .LCPI9_0
-; SZ13-NEXT: ldeb %f8, 0(%r1)
+; SZ13-NEXT: ld %f8, 0(%r1)
; SZ13-NEXT: vgmg %v0, 1, 1
; SZ13-NEXT: ldr %f2, %f8
; SZ13-NEXT: brasl %r14, fmod at PLT
@@ -531,13 +531,13 @@ define <4 x double> @constrained_vector_frem_v4f64() #0 {
; SZ13-NEXT: vmrhg %v0, %v0, %v1
; SZ13-NEXT: larl %r1, .LCPI9_1
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: ldr %f2, %f8
; SZ13-NEXT: brasl %r14, fmod at PLT
; SZ13-NEXT: larl %r1, .LCPI9_2
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 176(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: ldr %f2, %f8
; SZ13-NEXT: brasl %r14, fmod at PLT
; SZ13-NEXT: vl %v1, 176(%r15), 3 # 16-byte Folded Reload
@@ -586,13 +586,12 @@ define <2 x double> @constrained_vector_fmul_v2f64() #0 {
; S390X-LABEL: constrained_vector_fmul_v2f64:
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI11_0
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI11_1
-; S390X-NEXT: ld %f1, 0(%r1)
+; S390X-NEXT: ldr %f2, %f0
+; S390X-NEXT: mdb %f2, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI11_2
-; S390X-NEXT: ldeb %f0, 0(%r1)
-; S390X-NEXT: mdbr %f2, %f1
-; S390X-NEXT: mdbr %f0, %f1
+; S390X-NEXT: mdb %f0, 0(%r1)
; S390X-NEXT: br %r14
;
; SZ13-LABEL: constrained_vector_fmul_v2f64:
@@ -694,19 +693,18 @@ define <4 x double> @constrained_vector_fmul_v4f64() #0 {
; S390X-LABEL: constrained_vector_fmul_v4f64:
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI14_0
-; S390X-NEXT: ldeb %f6, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI14_1
-; S390X-NEXT: ld %f1, 0(%r1)
+; S390X-NEXT: ldr %f6, %f0
+; S390X-NEXT: mdb %f6, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI14_2
-; S390X-NEXT: ldeb %f4, 0(%r1)
+; S390X-NEXT: ldr %f4, %f0
+; S390X-NEXT: mdb %f4, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI14_3
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ldr %f2, %f0
+; S390X-NEXT: mdb %f2, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI14_4
-; S390X-NEXT: ldeb %f0, 0(%r1)
-; S390X-NEXT: mdbr %f6, %f1
-; S390X-NEXT: mdbr %f4, %f1
-; S390X-NEXT: mdbr %f2, %f1
-; S390X-NEXT: mdbr %f0, %f1
+; S390X-NEXT: mdb %f0, 0(%r1)
; S390X-NEXT: br %r14
;
; SZ13-LABEL: constrained_vector_fmul_v4f64:
@@ -760,13 +758,12 @@ define <2 x double> @constrained_vector_fadd_v2f64() #0 {
; S390X-LABEL: constrained_vector_fadd_v2f64:
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI16_0
-; S390X-NEXT: ld %f1, 0(%r1)
-; S390X-NEXT: larl %r1, .LCPI16_2
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI16_1
-; S390X-NEXT: ldr %f2, %f1
+; S390X-NEXT: ldr %f2, %f0
; S390X-NEXT: adb %f2, 0(%r1)
-; S390X-NEXT: adbr %f0, %f1
+; S390X-NEXT: larl %r1, .LCPI16_2
+; S390X-NEXT: adb %f0, 0(%r1)
; S390X-NEXT: br %r14
;
; SZ13-LABEL: constrained_vector_fadd_v2f64:
@@ -866,19 +863,18 @@ define <4 x double> @constrained_vector_fadd_v4f64() #0 {
; S390X-LABEL: constrained_vector_fadd_v4f64:
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI19_0
-; S390X-NEXT: ld %f1, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI19_1
-; S390X-NEXT: ldr %f2, %f1
-; S390X-NEXT: ldr %f6, %f1
+; S390X-NEXT: ldr %f6, %f0
; S390X-NEXT: adb %f6, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI19_2
-; S390X-NEXT: ldeb %f4, 0(%r1)
-; S390X-NEXT: larl %r1, .LCPI19_4
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ldr %f4, %f0
+; S390X-NEXT: adb %f4, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI19_3
+; S390X-NEXT: ldr %f2, %f0
; S390X-NEXT: adb %f2, 0(%r1)
-; S390X-NEXT: adbr %f4, %f1
-; S390X-NEXT: adbr %f0, %f1
+; S390X-NEXT: larl %r1, .LCPI19_4
+; S390X-NEXT: adb %f0, 0(%r1)
; S390X-NEXT: br %r14
;
; SZ13-LABEL: constrained_vector_fadd_v4f64:
@@ -933,12 +929,11 @@ define <2 x double> @constrained_vector_fsub_v2f64() #0 {
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI21_0
; S390X-NEXT: ld %f0, 0(%r1)
-; S390X-NEXT: larl %r1, .LCPI21_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI21_1
; S390X-NEXT: ldr %f2, %f0
; S390X-NEXT: sdb %f2, 0(%r1)
-; S390X-NEXT: sdbr %f0, %f1
+; S390X-NEXT: larl %r1, .LCPI21_2
+; S390X-NEXT: sdb %f0, 0(%r1)
; S390X-NEXT: br %r14
;
; SZ13-LABEL: constrained_vector_fsub_v2f64:
@@ -1043,15 +1038,13 @@ define <4 x double> @constrained_vector_fsub_v4f64() #0 {
; S390X-NEXT: ldr %f6, %f0
; S390X-NEXT: sdb %f6, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI24_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
-; S390X-NEXT: larl %r1, .LCPI24_4
-; S390X-NEXT: ldeb %f3, 0(%r1)
+; S390X-NEXT: ldr %f4, %f0
+; S390X-NEXT: sdb %f4, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI24_3
; S390X-NEXT: ldr %f2, %f0
; S390X-NEXT: sdb %f2, 0(%r1)
-; S390X-NEXT: ldr %f4, %f0
-; S390X-NEXT: sdbr %f4, %f1
-; S390X-NEXT: sdbr %f0, %f3
+; S390X-NEXT: larl %r1, .LCPI24_4
+; S390X-NEXT: sdb %f0, 0(%r1)
; S390X-NEXT: br %r14
;
; SZ13-LABEL: constrained_vector_fsub_v4f64:
@@ -1100,10 +1093,9 @@ define <2 x double> @constrained_vector_sqrt_v2f64() #0 {
; S390X-LABEL: constrained_vector_sqrt_v2f64:
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI26_0
-; S390X-NEXT: larl %r2, .LCPI26_1
-; S390X-NEXT: ldeb %f0, 0(%r2)
; S390X-NEXT: sqdb %f2, 0(%r1)
-; S390X-NEXT: sqdbr %f0, %f0
+; S390X-NEXT: larl %r1, .LCPI26_1
+; S390X-NEXT: sqdb %f0, 0(%r1)
; S390X-NEXT: br %r14
;
; SZ13-LABEL: constrained_vector_sqrt_v2f64:
@@ -1187,11 +1179,10 @@ define <4 x double> @constrained_vector_sqrt_v4f64() #0 {
; S390X-NEXT: sqdb %f6, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI29_1
; S390X-NEXT: sqdb %f4, 0(%r1)
-; S390X-NEXT: larl %r1, .LCPI29_3
-; S390X-NEXT: ldeb %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI29_2
; S390X-NEXT: sqdb %f2, 0(%r1)
-; S390X-NEXT: sqdbr %f0, %f0
+; S390X-NEXT: larl %r1, .LCPI29_3
+; S390X-NEXT: sqdb %f0, 0(%r1)
; S390X-NEXT: br %r14
;
; SZ13-LABEL: constrained_vector_sqrt_v4f64:
@@ -1268,7 +1259,7 @@ define <2 x double> @constrained_vector_pow_v2f64() #0 {
; S390X-NEXT: larl %r1, .LCPI31_0
; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI31_1
-; S390X-NEXT: ldeb %f8, 0(%r1)
+; S390X-NEXT: ld %f8, 0(%r1)
; S390X-NEXT: ldr %f2, %f8
; S390X-NEXT: brasl %r14, pow at PLT
; S390X-NEXT: larl %r1, .LCPI31_2
@@ -1295,7 +1286,7 @@ define <2 x double> @constrained_vector_pow_v2f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI31_0
; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI31_1
-; SZ13-NEXT: ldeb %f8, 0(%r1)
+; SZ13-NEXT: ld %f8, 0(%r1)
; SZ13-NEXT: ldr %f2, %f8
; SZ13-NEXT: brasl %r14, pow at PLT
; SZ13-NEXT: larl %r1, .LCPI31_2
@@ -1425,7 +1416,7 @@ define void @constrained_vector_pow_v3f64(<3 x double>* %a) #0 {
; S390X-NEXT: ld %f8, 0(%r2)
; S390X-NEXT: ld %f0, 16(%r2)
; S390X-NEXT: larl %r1, .LCPI33_0
-; S390X-NEXT: ldeb %f9, 0(%r1)
+; S390X-NEXT: ld %f9, 0(%r1)
; S390X-NEXT: ld %f10, 8(%r2)
; S390X-NEXT: ldr %f2, %f9
; S390X-NEXT: brasl %r14, pow at PLT
@@ -1460,7 +1451,7 @@ define void @constrained_vector_pow_v3f64(<3 x double>* %a) #0 {
; SZ13-NEXT: .cfi_offset %f8, -168
; SZ13-NEXT: .cfi_offset %f9, -176
; SZ13-NEXT: larl %r1, .LCPI33_0
-; SZ13-NEXT: ldeb %f9, 0(%r1)
+; SZ13-NEXT: ld %f9, 0(%r1)
; SZ13-NEXT: vl %v0, 0(%r2), 4
; SZ13-NEXT: ld %f8, 16(%r2)
; SZ13-NEXT: ldr %f2, %f9
@@ -1519,7 +1510,7 @@ define <4 x double> @constrained_vector_pow_v4f64() #0 {
; S390X-NEXT: larl %r1, .LCPI34_0
; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI34_1
-; S390X-NEXT: ldeb %f8, 0(%r1)
+; S390X-NEXT: ld %f8, 0(%r1)
; S390X-NEXT: ldr %f2, %f8
; S390X-NEXT: brasl %r14, pow at PLT
; S390X-NEXT: larl %r1, .LCPI34_2
@@ -1562,7 +1553,7 @@ define <4 x double> @constrained_vector_pow_v4f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI34_0
; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI34_1
-; SZ13-NEXT: ldeb %f8, 0(%r1)
+; SZ13-NEXT: ld %f8, 0(%r1)
; SZ13-NEXT: ldr %f2, %f8
; SZ13-NEXT: brasl %r14, pow at PLT
; SZ13-NEXT: larl %r1, .LCPI34_2
@@ -1787,7 +1778,7 @@ define void @constrained_vector_powi_v3f64(<3 x double>* %a) #0 {
; S390X-NEXT: .cfi_offset %f8, -168
; S390X-NEXT: .cfi_offset %f9, -176
; S390X-NEXT: larl %r1, .LCPI38_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: lgr %r13, %r2
; S390X-NEXT: lghi %r2, 3
; S390X-NEXT: brasl %r14, __powidf2 at PLT
@@ -1827,7 +1818,7 @@ define void @constrained_vector_powi_v3f64(<3 x double>* %a) #0 {
; SZ13-NEXT: larl %r1, .LCPI38_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: lghi %r2, 3
; SZ13-NEXT: brasl %r14, __powidf2 at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
@@ -1996,7 +1987,7 @@ define <2 x double> @constrained_vector_sin_v2f64() #0 {
; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: brasl %r14, sin at PLT
; S390X-NEXT: larl %r1, .LCPI41_1
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, sin at PLT
@@ -2018,7 +2009,7 @@ define <2 x double> @constrained_vector_sin_v2f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI41_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, sin at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -2208,7 +2199,7 @@ define <4 x double> @constrained_vector_sin_v4f64() #0 {
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, sin at PLT
; S390X-NEXT: larl %r1, .LCPI44_3
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, sin at PLT
@@ -2234,7 +2225,7 @@ define <4 x double> @constrained_vector_sin_v4f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI44_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, sin at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -2313,7 +2304,7 @@ define <2 x double> @constrained_vector_cos_v2f64() #0 {
; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: brasl %r14, cos at PLT
; S390X-NEXT: larl %r1, .LCPI46_1
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, cos at PLT
@@ -2335,7 +2326,7 @@ define <2 x double> @constrained_vector_cos_v2f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI46_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, cos at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -2525,7 +2516,7 @@ define <4 x double> @constrained_vector_cos_v4f64() #0 {
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, cos at PLT
; S390X-NEXT: larl %r1, .LCPI49_3
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, cos at PLT
@@ -2551,7 +2542,7 @@ define <4 x double> @constrained_vector_cos_v4f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI49_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, cos at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -2630,7 +2621,7 @@ define <2 x double> @constrained_vector_exp_v2f64() #0 {
; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: brasl %r14, exp at PLT
; S390X-NEXT: larl %r1, .LCPI51_1
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, exp at PLT
@@ -2652,7 +2643,7 @@ define <2 x double> @constrained_vector_exp_v2f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI51_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, exp at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -2842,7 +2833,7 @@ define <4 x double> @constrained_vector_exp_v4f64() #0 {
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, exp at PLT
; S390X-NEXT: larl %r1, .LCPI54_3
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, exp at PLT
@@ -2868,7 +2859,7 @@ define <4 x double> @constrained_vector_exp_v4f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI54_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, exp at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -2944,7 +2935,7 @@ define <2 x double> @constrained_vector_exp2_v2f64() #0 {
; S390X-NEXT: std %f8, 160(%r15) # 8-byte Folded Spill
; S390X-NEXT: .cfi_offset %f8, -168
; S390X-NEXT: larl %r1, .LCPI56_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: brasl %r14, exp2 at PLT
; S390X-NEXT: larl %r1, .LCPI56_1
; S390X-NEXT: ld %f1, 0(%r1)
@@ -2964,7 +2955,7 @@ define <2 x double> @constrained_vector_exp2_v2f64() #0 {
; SZ13-NEXT: aghi %r15, -176
; SZ13-NEXT: .cfi_def_cfa_offset 336
; SZ13-NEXT: larl %r1, .LCPI56_0
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, exp2 at PLT
; SZ13-NEXT: larl %r1, .LCPI56_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -3264,7 +3255,7 @@ define <2 x double> @constrained_vector_log_v2f64() #0 {
; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: brasl %r14, log at PLT
; S390X-NEXT: larl %r1, .LCPI61_1
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, log at PLT
@@ -3286,7 +3277,7 @@ define <2 x double> @constrained_vector_log_v2f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI61_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, log at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -3476,7 +3467,7 @@ define <4 x double> @constrained_vector_log_v4f64() #0 {
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, log at PLT
; S390X-NEXT: larl %r1, .LCPI64_3
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, log at PLT
@@ -3502,7 +3493,7 @@ define <4 x double> @constrained_vector_log_v4f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI64_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, log at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -3581,7 +3572,7 @@ define <2 x double> @constrained_vector_log10_v2f64() #0 {
; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: brasl %r14, log10 at PLT
; S390X-NEXT: larl %r1, .LCPI66_1
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, log10 at PLT
@@ -3603,7 +3594,7 @@ define <2 x double> @constrained_vector_log10_v2f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI66_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, log10 at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -3793,7 +3784,7 @@ define <4 x double> @constrained_vector_log10_v4f64() #0 {
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, log10 at PLT
; S390X-NEXT: larl %r1, .LCPI69_3
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, log10 at PLT
@@ -3819,7 +3810,7 @@ define <4 x double> @constrained_vector_log10_v4f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI69_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, log10 at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -3898,7 +3889,7 @@ define <2 x double> @constrained_vector_log2_v2f64() #0 {
; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: brasl %r14, log2 at PLT
; S390X-NEXT: larl %r1, .LCPI71_1
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, log2 at PLT
@@ -3920,7 +3911,7 @@ define <2 x double> @constrained_vector_log2_v2f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI71_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, log2 at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -4110,7 +4101,7 @@ define <4 x double> @constrained_vector_log2_v4f64() #0 {
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, log2 at PLT
; S390X-NEXT: larl %r1, .LCPI74_3
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, log2 at PLT
@@ -4136,7 +4127,7 @@ define <4 x double> @constrained_vector_log2_v4f64() #0 {
; SZ13-NEXT: larl %r1, .LCPI74_1
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: brasl %r14, log2 at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -4192,7 +4183,7 @@ define <2 x double> @constrained_vector_rint_v2f64() #0 {
; S390X-LABEL: constrained_vector_rint_v2f64:
; S390X: # %bb.0: # %entry
; S390X-NEXT: larl %r1, .LCPI76_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI76_1
; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: fidbr %f2, 0, %f0
@@ -4358,7 +4349,7 @@ define <2 x double> @constrained_vector_nearbyint_v2f64() #0 {
; S390X-NEXT: std %f8, 160(%r15) # 8-byte Folded Spill
; S390X-NEXT: .cfi_offset %f8, -168
; S390X-NEXT: larl %r1, .LCPI81_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: brasl %r14, nearbyint at PLT
; S390X-NEXT: larl %r1, .LCPI81_1
; S390X-NEXT: ld %f1, 0(%r1)
@@ -4602,14 +4593,14 @@ define <2 x double> @constrained_vector_maxnum_v2f64() #0 {
; S390X-NEXT: std %f8, 160(%r15) # 8-byte Folded Spill
; S390X-NEXT: .cfi_offset %f8, -168
; S390X-NEXT: larl %r1, .LCPI86_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI86_1
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: brasl %r14, fmax at PLT
; S390X-NEXT: larl %r1, .LCPI86_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI86_3
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, fmax at PLT
@@ -4626,16 +4617,16 @@ define <2 x double> @constrained_vector_maxnum_v2f64() #0 {
; SZ13-NEXT: aghi %r15, -176
; SZ13-NEXT: .cfi_def_cfa_offset 336
; SZ13-NEXT: larl %r1, .LCPI86_0
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI86_1
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmax at PLT
; SZ13-NEXT: larl %r1, .LCPI86_2
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI86_3
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmax at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -4754,16 +4745,16 @@ define void @constrained_vector_log10_maxnum_v3f64(<3 x double>* %a) #0 {
; S390X-NEXT: ld %f8, 0(%r2)
; S390X-NEXT: ld %f0, 16(%r2)
; S390X-NEXT: larl %r1, .LCPI88_0
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ld %f9, 8(%r2)
; S390X-NEXT: brasl %r14, fmax at PLT
; S390X-NEXT: larl %r1, .LCPI88_1
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f9
; S390X-NEXT: brasl %r14, fmax at PLT
; S390X-NEXT: larl %r1, .LCPI88_2
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f9, %f0
; S390X-NEXT: ldr %f0, %f8
; S390X-NEXT: brasl %r14, fmax at PLT
@@ -4787,24 +4778,24 @@ define void @constrained_vector_log10_maxnum_v3f64(<3 x double>* %a) #0 {
; SZ13-NEXT: std %f8, 192(%r15) # 8-byte Folded Spill
; SZ13-NEXT: .cfi_offset %f8, -168
; SZ13-NEXT: larl %r1, .LCPI88_0
-; SZ13-NEXT: ldeb %f2, 0(%r1)
; SZ13-NEXT: vl %v0, 0(%r2), 4
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: ld %f8, 16(%r2)
; SZ13-NEXT: lgr %r13, %r2
; SZ13-NEXT: vst %v0, 176(%r15), 3 # 16-byte Folded Spill
; SZ13-NEXT: # kill: def $f0d killed $f0d killed $v0
; SZ13-NEXT: brasl %r14, fmax at PLT
; SZ13-NEXT: larl %r1, .LCPI88_1
-; SZ13-NEXT: ldeb %f2, 0(%r1)
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
; SZ13-NEXT: vl %v0, 176(%r15), 3 # 16-byte Folded Reload
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: vrepg %v0, %v0, 1
; SZ13-NEXT: # kill: def $f0d killed $f0d killed $v0
; SZ13-NEXT: brasl %r14, fmax at PLT
-; SZ13-NEXT: larl %r1, .LCPI88_2
-; SZ13-NEXT: ldeb %f2, 0(%r1)
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
+; SZ13-NEXT: larl %r1, .LCPI88_2
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vmrhg %v0, %v1, %v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
@@ -4841,28 +4832,28 @@ define <4 x double> @constrained_vector_maxnum_v4f64() #0 {
; S390X-NEXT: .cfi_offset %f9, -176
; S390X-NEXT: .cfi_offset %f10, -184
; S390X-NEXT: larl %r1, .LCPI89_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI89_1
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: brasl %r14, fmax at PLT
; S390X-NEXT: larl %r1, .LCPI89_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI89_3
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, fmax at PLT
; S390X-NEXT: larl %r1, .LCPI89_4
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI89_5
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f9, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, fmax at PLT
; S390X-NEXT: larl %r1, .LCPI89_6
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI89_7
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, fmax at PLT
@@ -4883,32 +4874,32 @@ define <4 x double> @constrained_vector_maxnum_v4f64() #0 {
; SZ13-NEXT: aghi %r15, -192
; SZ13-NEXT: .cfi_def_cfa_offset 352
; SZ13-NEXT: larl %r1, .LCPI89_0
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI89_1
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmax at PLT
; SZ13-NEXT: larl %r1, .LCPI89_2
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI89_3
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmax at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vmrhg %v0, %v0, %v1
; SZ13-NEXT: larl %r1, .LCPI89_4
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI89_5
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmax at PLT
; SZ13-NEXT: larl %r1, .LCPI89_6
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 176(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI89_7
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmax at PLT
; SZ13-NEXT: vl %v1, 176(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: vl %v24, 160(%r15), 3 # 16-byte Folded Reload
@@ -4976,14 +4967,14 @@ define <2 x double> @constrained_vector_minnum_v2f64() #0 {
; S390X-NEXT: std %f8, 160(%r15) # 8-byte Folded Spill
; S390X-NEXT: .cfi_offset %f8, -168
; S390X-NEXT: larl %r1, .LCPI91_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI91_1
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: brasl %r14, fmin at PLT
; S390X-NEXT: larl %r1, .LCPI91_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI91_3
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, fmin at PLT
@@ -5000,16 +4991,16 @@ define <2 x double> @constrained_vector_minnum_v2f64() #0 {
; SZ13-NEXT: aghi %r15, -176
; SZ13-NEXT: .cfi_def_cfa_offset 336
; SZ13-NEXT: larl %r1, .LCPI91_0
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI91_1
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmin at PLT
; SZ13-NEXT: larl %r1, .LCPI91_2
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI91_3
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmin at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
@@ -5130,7 +5121,7 @@ define void @constrained_vector_minnum_v3f64(<3 x double>* %a) #0 {
; S390X-NEXT: ld %f8, 0(%r2)
; S390X-NEXT: ld %f0, 16(%r2)
; S390X-NEXT: larl %r1, .LCPI93_0
-; S390X-NEXT: ldeb %f9, 0(%r1)
+; S390X-NEXT: ld %f9, 0(%r1)
; S390X-NEXT: ld %f10, 8(%r2)
; S390X-NEXT: ldr %f2, %f9
; S390X-NEXT: brasl %r14, fmin at PLT
@@ -5165,7 +5156,7 @@ define void @constrained_vector_minnum_v3f64(<3 x double>* %a) #0 {
; SZ13-NEXT: .cfi_offset %f8, -168
; SZ13-NEXT: .cfi_offset %f9, -176
; SZ13-NEXT: larl %r1, .LCPI93_0
-; SZ13-NEXT: ldeb %f9, 0(%r1)
+; SZ13-NEXT: ld %f9, 0(%r1)
; SZ13-NEXT: vl %v0, 0(%r2), 4
; SZ13-NEXT: ld %f8, 16(%r2)
; SZ13-NEXT: ldr %f2, %f9
@@ -5219,28 +5210,28 @@ define <4 x double> @constrained_vector_minnum_v4f64() #0 {
; S390X-NEXT: .cfi_offset %f9, -176
; S390X-NEXT: .cfi_offset %f10, -184
; S390X-NEXT: larl %r1, .LCPI94_0
-; S390X-NEXT: ldeb %f0, 0(%r1)
+; S390X-NEXT: ld %f0, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI94_1
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: brasl %r14, fmin at PLT
; S390X-NEXT: larl %r1, .LCPI94_2
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI94_3
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f8, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, fmin at PLT
; S390X-NEXT: larl %r1, .LCPI94_4
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI94_5
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f9, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, fmin at PLT
; S390X-NEXT: larl %r1, .LCPI94_6
-; S390X-NEXT: ldeb %f1, 0(%r1)
+; S390X-NEXT: ld %f1, 0(%r1)
; S390X-NEXT: larl %r1, .LCPI94_7
-; S390X-NEXT: ldeb %f2, 0(%r1)
+; S390X-NEXT: ld %f2, 0(%r1)
; S390X-NEXT: ldr %f10, %f0
; S390X-NEXT: ldr %f0, %f1
; S390X-NEXT: brasl %r14, fmin at PLT
@@ -5261,32 +5252,32 @@ define <4 x double> @constrained_vector_minnum_v4f64() #0 {
; SZ13-NEXT: aghi %r15, -192
; SZ13-NEXT: .cfi_def_cfa_offset 352
; SZ13-NEXT: larl %r1, .LCPI94_0
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI94_1
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmin at PLT
; SZ13-NEXT: larl %r1, .LCPI94_2
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI94_3
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmin at PLT
; SZ13-NEXT: vl %v1, 160(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vmrhg %v0, %v0, %v1
; SZ13-NEXT: larl %r1, .LCPI94_4
; SZ13-NEXT: vst %v0, 160(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI94_5
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmin at PLT
; SZ13-NEXT: larl %r1, .LCPI94_6
; SZ13-NEXT: # kill: def $f0d killed $f0d def $v0
; SZ13-NEXT: vst %v0, 176(%r15), 3 # 16-byte Folded Spill
-; SZ13-NEXT: ldeb %f0, 0(%r1)
+; SZ13-NEXT: ld %f0, 0(%r1)
; SZ13-NEXT: larl %r1, .LCPI94_7
-; SZ13-NEXT: ldeb %f2, 0(%r1)
+; SZ13-NEXT: ld %f2, 0(%r1)
; SZ13-NEXT: brasl %r14, fmin at PLT
; SZ13-NEXT: vl %v1, 176(%r15), 3 # 16-byte Folded Reload
; SZ13-NEXT: vl %v24, 160(%r15), 3 # 16-byte Folded Reload
More information about the llvm-commits
mailing list