[llvm-branch-commits] [llvm] f47bac5 - [ARM] Extra vecreduce tests with smaller than legal types. NFC
David Green via llvm-branch-commits
llvm-branch-commits at lists.llvm.org
Sun Dec 20 13:25:17 PST 2020
Author: David Green
Date: 2020-12-20T21:20:39Z
New Revision: f47bac5dd207c672e803e08685e084e5d66f8bce
URL: https://github.com/llvm/llvm-project/commit/f47bac5dd207c672e803e08685e084e5d66f8bce
DIFF: https://github.com/llvm/llvm-project/commit/f47bac5dd207c672e803e08685e084e5d66f8bce.diff
LOG: [ARM] Extra vecreduce tests with smaller than legal types. NFC
Added:
Modified:
llvm/test/CodeGen/Thumb2/mve-vecreduce-add.ll
llvm/test/CodeGen/Thumb2/mve-vecreduce-addpred.ll
llvm/test/CodeGen/Thumb2/mve-vecreduce-mla.ll
llvm/test/CodeGen/Thumb2/mve-vecreduce-mlapred.ll
Removed:
################################################################################
diff --git a/llvm/test/CodeGen/Thumb2/mve-vecreduce-add.ll b/llvm/test/CodeGen/Thumb2/mve-vecreduce-add.ll
index 995926a1502e..f882582bf148 100644
--- a/llvm/test/CodeGen/Thumb2/mve-vecreduce-add.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-vecreduce-add.ll
@@ -236,6 +236,53 @@ entry:
ret i64 %z
}
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_zext(<4 x i16> %x) {
+; CHECK-LABEL: add_v4i16_v4i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmovlb.u16 q0, q0
+; CHECK-NEXT: vaddlv.u32 r0, r1, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <4 x i16> %x to <4 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_sext(<4 x i16> %x) {
+; CHECK-LABEL: add_v4i16_v4i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.f32 s4, s0
+; CHECK-NEXT: vmov.f32 s6, s1
+; CHECK-NEXT: vmov r1, s0
+; CHECK-NEXT: vmov r0, s6
+; CHECK-NEXT: sxth r1, r1
+; CHECK-NEXT: sxth r0, r0
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: asrs r2, r0, #31
+; CHECK-NEXT: asrs r1, r1, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r2
+; CHECK-NEXT: vmov r2, s6
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r1, s5
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: adc.w r0, r1, r0, asr #31
+; CHECK-NEXT: vmov r1, s4
+; CHECK-NEXT: sxth r1, r1
+; CHECK-NEXT: adds r2, r2, r1
+; CHECK-NEXT: adc.w r1, r0, r1, asr #31
+; CHECK-NEXT: vmov r0, s6
+; CHECK-NEXT: sxth r3, r0
+; CHECK-NEXT: adds r0, r2, r3
+; CHECK-NEXT: adc.w r1, r1, r3, asr #31
+; CHECK-NEXT: bx lr
+entry:
+ %xx = sext <4 x i16> %x to <4 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
+ ret i64 %z
+}
+
define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_zext(<2 x i16> %x) {
; CHECK-LABEL: add_v2i16_v2i64_zext:
; CHECK: @ %bb.0: @ %entry
@@ -294,6 +341,46 @@ entry:
ret i32 %z
}
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_zext(<8 x i8> %x) {
+; CHECK-LABEL: add_v8i8_v8i32_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vaddv.u16 r0, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <8 x i8> %x to <8 x i32>
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
+ ret i32 %z
+}
+
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_sext(<8 x i8> %x) {
+; CHECK-LABEL: add_v8i8_v8i32_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.u16 r0, q0[6]
+; CHECK-NEXT: vmov.u16 r1, q0[4]
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[7]
+; CHECK-NEXT: vmov.u16 r1, q0[5]
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[2]
+; CHECK-NEXT: vmov.u16 r1, q0[0]
+; CHECK-NEXT: vmovlb.s8 q1, q1
+; CHECK-NEXT: vmov q2[2], q2[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[3]
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vmovlb.s16 q1, q1
+; CHECK-NEXT: vmov q2[3], q2[1], r1, r0
+; CHECK-NEXT: vmovlb.s8 q0, q2
+; CHECK-NEXT: vmovlb.s16 q0, q0
+; CHECK-NEXT: vadd.i32 q0, q0, q1
+; CHECK-NEXT: vaddv.u32 r0, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = sext <8 x i8> %x to <8 x i32>
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
+ ret i32 %z
+}
+
define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_zext(<4 x i8> %x) {
; CHECK-LABEL: add_v4i8_v4i32_zext:
; CHECK: @ %bb.0: @ %entry
@@ -599,6 +686,165 @@ entry:
ret i64 %z
}
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_zext(<8 x i8> %x) {
+; CHECK-LABEL: add_v8i8_v8i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vmov.i64 q1, #0xffff
+; CHECK-NEXT: vmov.u16 r0, q0[1]
+; CHECK-NEXT: vmov.u16 r1, q0[0]
+; CHECK-NEXT: vmov q2[2], q2[0], r1, r0
+; CHECK-NEXT: vmov.u16 r2, q0[2]
+; CHECK-NEXT: vand q2, q2, q1
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: vmov r1, s8
+; CHECK-NEXT: add r0, r1
+; CHECK-NEXT: vmov.u16 r1, q0[3]
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r1
+; CHECK-NEXT: vmov.u16 r2, q0[4]
+; CHECK-NEXT: vand q3, q3, q1
+; CHECK-NEXT: vmov r1, s12
+; CHECK-NEXT: add r0, r1
+; CHECK-NEXT: vmov r1, s14
+; CHECK-NEXT: add r0, r1
+; CHECK-NEXT: vmov.u16 r1, q0[5]
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r1
+; CHECK-NEXT: vmov r2, s11
+; CHECK-NEXT: vand q3, q3, q1
+; CHECK-NEXT: vmov r1, s12
+; CHECK-NEXT: vmov r3, s14
+; CHECK-NEXT: add r0, r1
+; CHECK-NEXT: vmov r1, s15
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: vmov.u16 r3, q0[6]
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: vmov.u16 r2, q0[7]
+; CHECK-NEXT: vmov q0[2], q0[0], r3, r2
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r2, s1
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: vmov r3, s2
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: vmov r2, s3
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <8 x i8> %x to <8 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_sext(<8 x i8> %x) {
+; CHECK-LABEL: add_v8i8_v8i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.u16 r0, q0[1]
+; CHECK-NEXT: vmov.u16 r1, q0[0]
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: asrs r2, r0, #31
+; CHECK-NEXT: asrs r1, r1, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r2
+; CHECK-NEXT: vmov r2, s6
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r1, s5
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov.u16 r3, q0[2]
+; CHECK-NEXT: adc.w r12, r1, r0, asr #31
+; CHECK-NEXT: vmov.u16 r1, q0[3]
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: vmov q1[2], q1[0], r3, r1
+; CHECK-NEXT: asrs r0, r1, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r3, r0
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r0, s5
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: adc.w r0, r0, r12
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov.u16 r3, q0[4]
+; CHECK-NEXT: adc.w r12, r0, r1, asr #31
+; CHECK-NEXT: vmov.u16 r1, q0[5]
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: vmov q1[2], q1[0], r3, r1
+; CHECK-NEXT: asrs r0, r1, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r3, r0
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r0, s5
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: adc.w r0, r0, r12
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: adc.w r0, r0, r1, asr #31
+; CHECK-NEXT: vmov.u16 r1, q0[6]
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: adds r2, r2, r1
+; CHECK-NEXT: adc.w r1, r0, r1, asr #31
+; CHECK-NEXT: vmov.u16 r0, q0[7]
+; CHECK-NEXT: sxtb r3, r0
+; CHECK-NEXT: adds r0, r2, r3
+; CHECK-NEXT: adc.w r1, r1, r3, asr #31
+; CHECK-NEXT: bx lr
+entry:
+ %xx = sext <8 x i8> %x to <8 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_zext(<4 x i8> %x) {
+; CHECK-LABEL: add_v4i8_v4i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.i32 q1, #0xff
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vaddlv.u32 r0, r1, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <4 x i8> %x to <4 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_sext(<4 x i8> %x) {
+; CHECK-LABEL: add_v4i8_v4i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.f32 s4, s0
+; CHECK-NEXT: vmov.f32 s6, s1
+; CHECK-NEXT: vmov r1, s0
+; CHECK-NEXT: vmov r0, s6
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: asrs r2, r0, #31
+; CHECK-NEXT: asrs r1, r1, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r2
+; CHECK-NEXT: vmov r2, s6
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r1, s5
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: adc.w r0, r1, r0, asr #31
+; CHECK-NEXT: vmov r1, s4
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: adds r2, r2, r1
+; CHECK-NEXT: adc.w r1, r0, r1, asr #31
+; CHECK-NEXT: vmov r0, s6
+; CHECK-NEXT: sxtb r3, r0
+; CHECK-NEXT: adds r0, r2, r3
+; CHECK-NEXT: adc.w r1, r1, r3, asr #31
+; CHECK-NEXT: bx lr
+entry:
+ %xx = sext <4 x i8> %x to <4 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
+ ret i64 %z
+}
+
define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_zext(<2 x i8> %x) {
; CHECK-LABEL: add_v2i8_v2i64_zext:
; CHECK: @ %bb.0: @ %entry
@@ -911,6 +1157,59 @@ entry:
ret i64 %r
}
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_acc_zext(<4 x i16> %x, i64 %a) {
+; CHECK-LABEL: add_v4i16_v4i64_acc_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmovlb.u16 q0, q0
+; CHECK-NEXT: vaddlva.u32 r0, r1, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <4 x i16> %x to <4 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
+ %r = add i64 %z, %a
+ ret i64 %r
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_acc_sext(<4 x i16> %x, i64 %a) {
+; CHECK-LABEL: add_v4i16_v4i64_acc_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r7, lr}
+; CHECK-NEXT: push {r7, lr}
+; CHECK-NEXT: vmov.f32 s4, s0
+; CHECK-NEXT: vmov.f32 s6, s1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r2, s6
+; CHECK-NEXT: sxth r3, r3
+; CHECK-NEXT: sxth r2, r2
+; CHECK-NEXT: vmov q1[2], q1[0], r3, r2
+; CHECK-NEXT: asr.w r12, r2, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r3, r12
+; CHECK-NEXT: vmov lr, s6
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r12, s5
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: adds.w r3, r3, lr
+; CHECK-NEXT: adc.w r12, r12, r2, asr #31
+; CHECK-NEXT: vmov r2, s4
+; CHECK-NEXT: sxth r2, r2
+; CHECK-NEXT: adds r3, r3, r2
+; CHECK-NEXT: adc.w r12, r12, r2, asr #31
+; CHECK-NEXT: vmov r2, s6
+; CHECK-NEXT: sxth r2, r2
+; CHECK-NEXT: adds r3, r3, r2
+; CHECK-NEXT: adc.w r2, r12, r2, asr #31
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: pop {r7, pc}
+entry:
+ %xx = sext <4 x i16> %x to <4 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
+ %r = add i64 %z, %a
+ ret i64 %r
+}
+
define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_acc_zext(<2 x i16> %x, i64 %a) {
; CHECK-LABEL: add_v2i16_v2i64_acc_zext:
; CHECK: @ %bb.0: @ %entry
@@ -977,6 +1276,48 @@ entry:
ret i32 %r
}
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_acc_zext(<8 x i8> %x, i32 %a) {
+; CHECK-LABEL: add_v8i8_v8i32_acc_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vaddva.u16 r0, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <8 x i8> %x to <8 x i32>
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
+ %r = add i32 %z, %a
+ ret i32 %r
+}
+
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_acc_sext(<8 x i8> %x, i32 %a) {
+; CHECK-LABEL: add_v8i8_v8i32_acc_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.u16 r1, q0[6]
+; CHECK-NEXT: vmov.u16 r2, q0[4]
+; CHECK-NEXT: vmov q1[2], q1[0], r2, r1
+; CHECK-NEXT: vmov.u16 r1, q0[7]
+; CHECK-NEXT: vmov.u16 r2, q0[5]
+; CHECK-NEXT: vmov q1[3], q1[1], r2, r1
+; CHECK-NEXT: vmov.u16 r1, q0[2]
+; CHECK-NEXT: vmov.u16 r2, q0[0]
+; CHECK-NEXT: vmovlb.s8 q1, q1
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r1
+; CHECK-NEXT: vmov.u16 r1, q0[3]
+; CHECK-NEXT: vmov.u16 r2, q0[1]
+; CHECK-NEXT: vmovlb.s16 q1, q1
+; CHECK-NEXT: vmov q2[3], q2[1], r2, r1
+; CHECK-NEXT: vmovlb.s8 q0, q2
+; CHECK-NEXT: vmovlb.s16 q0, q0
+; CHECK-NEXT: vadd.i32 q0, q0, q1
+; CHECK-NEXT: vaddva.u32 r0, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = sext <8 x i8> %x to <8 x i32>
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
+ %r = add i32 %z, %a
+ ret i32 %r
+}
+
define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_acc_zext(<4 x i8> %x, i32 %a) {
; CHECK-LABEL: add_v4i8_v4i32_acc_zext:
; CHECK: @ %bb.0: @ %entry
@@ -1299,6 +1640,181 @@ entry:
ret i64 %r
}
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_acc_zext(<8 x i8> %x, i64 %a) {
+; CHECK-LABEL: add_v8i8_v8i64_acc_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r4, lr}
+; CHECK-NEXT: push {r4, lr}
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vmov.i64 q1, #0xffff
+; CHECK-NEXT: vmov.u16 r2, q0[1]
+; CHECK-NEXT: vmov.u16 r3, q0[0]
+; CHECK-NEXT: vmov q2[2], q2[0], r3, r2
+; CHECK-NEXT: vand q2, q2, q1
+; CHECK-NEXT: vmov r2, s10
+; CHECK-NEXT: vmov r3, s8
+; CHECK-NEXT: add.w r12, r3, r2
+; CHECK-NEXT: vmov.u16 r3, q0[3]
+; CHECK-NEXT: vmov.u16 r2, q0[2]
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r3
+; CHECK-NEXT: vand q3, q3, q1
+; CHECK-NEXT: vmov r2, s12
+; CHECK-NEXT: vmov r3, s14
+; CHECK-NEXT: add r2, r12
+; CHECK-NEXT: add.w r12, r2, r3
+; CHECK-NEXT: vmov.u16 r3, q0[5]
+; CHECK-NEXT: vmov.u16 r2, q0[4]
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r3
+; CHECK-NEXT: vand q3, q3, q1
+; CHECK-NEXT: vmov r2, s12
+; CHECK-NEXT: vmov r3, s14
+; CHECK-NEXT: vmov lr, s15
+; CHECK-NEXT: add r12, r2
+; CHECK-NEXT: vmov r2, s11
+; CHECK-NEXT: adds.w r4, r12, r3
+; CHECK-NEXT: vmov.u16 r3, q0[6]
+; CHECK-NEXT: adc.w r12, r2, lr
+; CHECK-NEXT: vmov.u16 r2, q0[7]
+; CHECK-NEXT: vmov q0[2], q0[0], r3, r2
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r2, s1
+; CHECK-NEXT: adds r3, r3, r4
+; CHECK-NEXT: vmov r4, s3
+; CHECK-NEXT: adc.w r12, r12, r2
+; CHECK-NEXT: vmov r2, s2
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: adc.w r3, r12, r4
+; CHECK-NEXT: adds r0, r0, r2
+; CHECK-NEXT: adcs r1, r3
+; CHECK-NEXT: pop {r4, pc}
+entry:
+ %xx = zext <8 x i8> %x to <8 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
+ %r = add i64 %z, %a
+ ret i64 %r
+}
+
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_acc_sext(<8 x i8> %x, i64 %a) {
+; CHECK-LABEL: add_v8i8_v8i64_acc_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r4, lr}
+; CHECK-NEXT: push {r4, lr}
+; CHECK-NEXT: vmov.u16 r2, q0[1]
+; CHECK-NEXT: vmov.u16 r3, q0[0]
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: vmov q1[2], q1[0], r3, r2
+; CHECK-NEXT: asr.w r12, r2, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r3, r12
+; CHECK-NEXT: vmov lr, s6
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r12, s5
+; CHECK-NEXT: adds.w lr, lr, r3
+; CHECK-NEXT: vmov.u16 r3, q0[2]
+; CHECK-NEXT: adc.w r12, r12, r2, asr #31
+; CHECK-NEXT: vmov.u16 r2, q0[3]
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: vmov q1[2], q1[0], r3, r2
+; CHECK-NEXT: asrs r4, r2, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r3, r4
+; CHECK-NEXT: vmov r4, s4
+; CHECK-NEXT: vmov r3, s5
+; CHECK-NEXT: adds.w r4, r4, lr
+; CHECK-NEXT: adc.w r12, r12, r3
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: adds.w lr, r4, r3
+; CHECK-NEXT: vmov.u16 r4, q0[5]
+; CHECK-NEXT: adc.w r12, r12, r2, asr #31
+; CHECK-NEXT: vmov.u16 r2, q0[4]
+; CHECK-NEXT: sxtb r4, r4
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: vmov q1[2], q1[0], r2, r4
+; CHECK-NEXT: asrs r3, r4, #31
+; CHECK-NEXT: asrs r2, r2, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r2, r3
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r2, s5
+; CHECK-NEXT: adds.w r3, r3, lr
+; CHECK-NEXT: adc.w r12, r12, r2
+; CHECK-NEXT: vmov r2, s6
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: adc.w r3, r12, r4, asr #31
+; CHECK-NEXT: vmov.u16 r4, q0[6]
+; CHECK-NEXT: sxtb r4, r4
+; CHECK-NEXT: adds r2, r2, r4
+; CHECK-NEXT: adc.w r3, r3, r4, asr #31
+; CHECK-NEXT: vmov.u16 r4, q0[7]
+; CHECK-NEXT: sxtb r4, r4
+; CHECK-NEXT: adds r2, r2, r4
+; CHECK-NEXT: adc.w r3, r3, r4, asr #31
+; CHECK-NEXT: adds r0, r0, r2
+; CHECK-NEXT: adcs r1, r3
+; CHECK-NEXT: pop {r4, pc}
+entry:
+ %xx = sext <8 x i8> %x to <8 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
+ %r = add i64 %z, %a
+ ret i64 %r
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_acc_zext(<4 x i8> %x, i64 %a) {
+; CHECK-LABEL: add_v4i8_v4i64_acc_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.i32 q1, #0xff
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vaddlva.u32 r0, r1, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <4 x i8> %x to <4 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
+ %r = add i64 %z, %a
+ ret i64 %r
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_acc_sext(<4 x i8> %x, i64 %a) {
+; CHECK-LABEL: add_v4i8_v4i64_acc_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r7, lr}
+; CHECK-NEXT: push {r7, lr}
+; CHECK-NEXT: vmov.f32 s4, s0
+; CHECK-NEXT: vmov.f32 s6, s1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r2, s6
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: vmov q1[2], q1[0], r3, r2
+; CHECK-NEXT: asr.w r12, r2, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q1[3], q1[1], r3, r12
+; CHECK-NEXT: vmov lr, s6
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r12, s5
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: adds.w r3, r3, lr
+; CHECK-NEXT: adc.w r12, r12, r2, asr #31
+; CHECK-NEXT: vmov r2, s4
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: adds r3, r3, r2
+; CHECK-NEXT: adc.w r12, r12, r2, asr #31
+; CHECK-NEXT: vmov r2, s6
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: adds r3, r3, r2
+; CHECK-NEXT: adc.w r2, r12, r2, asr #31
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: pop {r7, pc}
+entry:
+ %xx = sext <4 x i8> %x to <4 x i64>
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
+ %r = add i64 %z, %a
+ ret i64 %r
+}
+
define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_acc_zext(<2 x i8> %x, i64 %a) {
; CHECK-LABEL: add_v2i8_v2i64_acc_zext:
; CHECK: @ %bb.0: @ %entry
diff --git a/llvm/test/CodeGen/Thumb2/mve-vecreduce-addpred.ll b/llvm/test/CodeGen/Thumb2/mve-vecreduce-addpred.ll
index 0403bb4781be..a5ad4eff87e0 100644
--- a/llvm/test/CodeGen/Thumb2/mve-vecreduce-addpred.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-vecreduce-addpred.ll
@@ -418,6 +418,122 @@ entry:
ret i64 %z
}
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_zext(<4 x i16> %x, <4 x i16> %b) {
+; CHECK-LABEL: add_v4i16_v4i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmovlb.u16 q1, q1
+; CHECK-NEXT: vmovlb.u16 q0, q0
+; CHECK-NEXT: vcmp.i32 eq, q1, zr
+; CHECK-NEXT: vmov.f32 s12, s0
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: vmov.f32 s14, s1
+; CHECK-NEXT: vmov.i64 q2, #0xffffffff
+; CHECK-NEXT: vand q3, q3, q2
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q1[2], q1[0], r2, r1
+; CHECK-NEXT: vmov q1[3], q1[1], r2, r1
+; CHECK-NEXT: vand q1, q3, q1
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: vmov r1, s4
+; CHECK-NEXT: vmov r12, s7
+; CHECK-NEXT: vmov r2, s5
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: vand q0, q1, q2
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: ubfx r3, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsb.w r3, r3, #0
+; CHECK-NEXT: rsb.w r0, r0, #0
+; CHECK-NEXT: adc.w r2, r2, r12
+; CHECK-NEXT: vmov q1[2], q1[0], r0, r3
+; CHECK-NEXT: vmov q1[3], q1[1], r0, r3
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adcs r2, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <4 x i16> %b, zeroinitializer
+ %xx = zext <4 x i16> %x to <4 x i64>
+ %s = select <4 x i1> %c, <4 x i64> %xx, <4 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %s)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_sext(<4 x i16> %x, <4 x i16> %b) {
+; CHECK-LABEL: add_v4i16_v4i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r7, lr}
+; CHECK-NEXT: push {r7, lr}
+; CHECK-NEXT: vmov.f32 s8, s0
+; CHECK-NEXT: vmovlb.u16 q1, q1
+; CHECK-NEXT: vmov.f32 s10, s1
+; CHECK-NEXT: vcmp.i32 eq, q1, zr
+; CHECK-NEXT: vmov r1, s0
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: sxth r1, r1
+; CHECK-NEXT: sxth r0, r0
+; CHECK-NEXT: vmov q2[2], q2[0], r1, r0
+; CHECK-NEXT: asrs r0, r0, #31
+; CHECK-NEXT: asrs r1, r1, #31
+; CHECK-NEXT: vmov q2[3], q2[1], r1, r0
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q1[2], q1[0], r2, r1
+; CHECK-NEXT: vmov q1[3], q1[1], r2, r1
+; CHECK-NEXT: vand q1, q2, q1
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: vmov r1, s4
+; CHECK-NEXT: vmov r12, s7
+; CHECK-NEXT: vmov r2, s5
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: adds.w lr, r1, r3
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: vmov r1, s4
+; CHECK-NEXT: adc.w r2, r2, r12
+; CHECK-NEXT: sxth r3, r3
+; CHECK-NEXT: sxth r1, r1
+; CHECK-NEXT: vmov q0[2], q0[0], r1, r3
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: asrs r1, r1, #31
+; CHECK-NEXT: vmov q0[3], q0[1], r1, r3
+; CHECK-NEXT: ubfx r1, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: vmov q1[2], q1[0], r0, r1
+; CHECK-NEXT: vmov q1[3], q1[1], r0, r1
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r1, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adds.w r1, r1, lr
+; CHECK-NEXT: adcs r2, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: pop {r7, pc}
+entry:
+ %c = icmp eq <4 x i16> %b, zeroinitializer
+ %xx = sext <4 x i16> %x to <4 x i64>
+ %s = select <4 x i1> %c, <4 x i64> %xx, <4 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %s)
+ ret i64 %z
+}
+
define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_zext(<2 x i16> %x, <2 x i16> %b) {
; CHECK-LABEL: add_v2i16_v2i64_zext:
; CHECK: @ %bb.0: @ %entry
@@ -521,6 +637,111 @@ entry:
ret i32 %z
}
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_zext(<8 x i8> %x, <8 x i8> %b) {
+; CHECK-LABEL: add_v8i8_v8i32_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .vsave {d8, d9}
+; CHECK-NEXT: vpush {d8, d9}
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vmovlb.u8 q1, q1
+; CHECK-NEXT: vcmp.i16 eq, q1, zr
+; CHECK-NEXT: vmov.u16 r0, q0[2]
+; CHECK-NEXT: vmov.u16 r1, q0[0]
+; CHECK-NEXT: vmov.i8 q1, #0x0
+; CHECK-NEXT: vmov.i8 q2, #0xff
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[3]
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vpsel q1, q2, q1
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[2]
+; CHECK-NEXT: vmov.u16 r1, q1[0]
+; CHECK-NEXT: vmov q2[2], q2[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[3]
+; CHECK-NEXT: vmov.u16 r1, q1[1]
+; CHECK-NEXT: vmov.i32 q4, #0xffff
+; CHECK-NEXT: vmov q2[3], q2[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[6]
+; CHECK-NEXT: vcmp.i32 ne, q2, zr
+; CHECK-NEXT: vmov.i32 q2, #0x0
+; CHECK-NEXT: vmov.u16 r1, q0[4]
+; CHECK-NEXT: vpst
+; CHECK-NEXT: vandt q2, q3, q4
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[7]
+; CHECK-NEXT: vmov.u16 r1, q0[5]
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[6]
+; CHECK-NEXT: vmov.u16 r1, q1[4]
+; CHECK-NEXT: vmovlb.u16 q0, q3
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[7]
+; CHECK-NEXT: vmov.u16 r1, q1[5]
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vpt.i32 ne, q3, zr
+; CHECK-NEXT: vaddt.i32 q2, q2, q0
+; CHECK-NEXT: vaddv.u32 r0, q2
+; CHECK-NEXT: vpop {d8, d9}
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <8 x i8> %b, zeroinitializer
+ %xx = zext <8 x i8> %x to <8 x i32>
+ %s = select <8 x i1> %c, <8 x i32> %xx, <8 x i32> zeroinitializer
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %s)
+ ret i32 %z
+}
+
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_sext(<8 x i8> %x, <8 x i8> %b) {
+; CHECK-LABEL: add_v8i8_v8i32_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmovlb.u8 q1, q1
+; CHECK-NEXT: vmov.u16 r0, q0[2]
+; CHECK-NEXT: vmov.u16 r1, q0[0]
+; CHECK-NEXT: vcmp.i16 eq, q1, zr
+; CHECK-NEXT: vmov.i8 q1, #0x0
+; CHECK-NEXT: vmov.i8 q3, #0xff
+; CHECK-NEXT: vmov q2[2], q2[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[3]
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vpsel q1, q3, q1
+; CHECK-NEXT: vmov q2[3], q2[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[2]
+; CHECK-NEXT: vmov.u16 r1, q1[0]
+; CHECK-NEXT: vmovlb.s8 q2, q2
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[3]
+; CHECK-NEXT: vmov.u16 r1, q1[1]
+; CHECK-NEXT: vmovlb.s16 q2, q2
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[6]
+; CHECK-NEXT: vcmp.i32 ne, q3, zr
+; CHECK-NEXT: vmov.i32 q3, #0x0
+; CHECK-NEXT: vmov.u16 r1, q0[4]
+; CHECK-NEXT: vpsel q2, q2, q3
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[7]
+; CHECK-NEXT: vmov.u16 r1, q0[5]
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[6]
+; CHECK-NEXT: vmov.u16 r1, q1[4]
+; CHECK-NEXT: vmovlb.s8 q0, q3
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[7]
+; CHECK-NEXT: vmov.u16 r1, q1[5]
+; CHECK-NEXT: vmovlb.s16 q0, q0
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vpt.i32 ne, q3, zr
+; CHECK-NEXT: vaddt.i32 q2, q2, q0
+; CHECK-NEXT: vaddv.u32 r0, q2
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <8 x i8> %b, zeroinitializer
+ %xx = sext <8 x i8> %x to <8 x i32>
+ %s = select <8 x i1> %c, <8 x i32> %xx, <8 x i32> zeroinitializer
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %s)
+ ret i32 %z
+}
+
define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_zext(<4 x i8> %x, <4 x i8> %b) {
; CHECK-LABEL: add_v4i8_v4i32_zext:
; CHECK: @ %bb.0: @ %entry
@@ -1141,6 +1362,360 @@ entry:
ret i64 %z
}
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_zext(<8 x i8> %x, <8 x i8> %b) {
+; CHECK-LABEL: add_v8i8_v8i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .vsave {d8, d9}
+; CHECK-NEXT: vpush {d8, d9}
+; CHECK-NEXT: vmovlb.u8 q1, q1
+; CHECK-NEXT: vmov.i8 q2, #0xff
+; CHECK-NEXT: vcmp.i16 eq, q1, zr
+; CHECK-NEXT: vmov.i8 q1, #0x0
+; CHECK-NEXT: vpsel q2, q2, q1
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vmov.u16 r0, q2[2]
+; CHECK-NEXT: vmov.u16 r1, q2[0]
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q2[3]
+; CHECK-NEXT: vmov.u16 r1, q2[1]
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r0
+; CHECK-NEXT: vcmp.i32 ne, q1, zr
+; CHECK-NEXT: vmov.i64 q1, #0xffff
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r1
+; CHECK-NEXT: vmov q3[3], q3[1], r2, r1
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vmov.u16 r2, q0[0]
+; CHECK-NEXT: vmov q4[2], q4[0], r2, r1
+; CHECK-NEXT: vand q4, q4, q1
+; CHECK-NEXT: vand q3, q4, q3
+; CHECK-NEXT: vmov r1, s15
+; CHECK-NEXT: vmov r2, s13
+; CHECK-NEXT: vmov r3, s12
+; CHECK-NEXT: orrs r1, r2
+; CHECK-NEXT: vmov r2, s14
+; CHECK-NEXT: add r2, r3
+; CHECK-NEXT: ubfx r3, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsbs r3, r3, #0
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: vmov q3[2], q3[0], r0, r3
+; CHECK-NEXT: vmov q3[3], q3[1], r0, r3
+; CHECK-NEXT: vmov.u16 r0, q0[3]
+; CHECK-NEXT: vmov.u16 r3, q0[2]
+; CHECK-NEXT: vmov q4[2], q4[0], r3, r0
+; CHECK-NEXT: vand q4, q4, q1
+; CHECK-NEXT: vand q3, q4, q3
+; CHECK-NEXT: vmov r3, s12
+; CHECK-NEXT: vmov r0, s13
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov r3, s14
+; CHECK-NEXT: adcs r0, r1
+; CHECK-NEXT: vmov r1, s15
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov.u16 r3, q2[4]
+; CHECK-NEXT: adc.w r12, r0, r1
+; CHECK-NEXT: vmov.u16 r1, q2[6]
+; CHECK-NEXT: vmov q3[2], q3[0], r3, r1
+; CHECK-NEXT: vmov.u16 r1, q2[7]
+; CHECK-NEXT: vmov.u16 r3, q2[5]
+; CHECK-NEXT: vmov q3[3], q3[1], r3, r1
+; CHECK-NEXT: vcmp.i32 ne, q3, zr
+; CHECK-NEXT: vmrs r1, p0
+; CHECK-NEXT: and r0, r1, #1
+; CHECK-NEXT: ubfx r3, r1, #4, #1
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: rsbs r3, r3, #0
+; CHECK-NEXT: vmov q2[2], q2[0], r0, r3
+; CHECK-NEXT: vmov q2[3], q2[1], r0, r3
+; CHECK-NEXT: vmov.u16 r0, q0[5]
+; CHECK-NEXT: vmov.u16 r3, q0[4]
+; CHECK-NEXT: vmov q3[2], q3[0], r3, r0
+; CHECK-NEXT: vand q3, q3, q1
+; CHECK-NEXT: vand q2, q3, q2
+; CHECK-NEXT: vmov r3, s8
+; CHECK-NEXT: vmov r0, s9
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov r3, s11
+; CHECK-NEXT: adc.w r12, r12, r0
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: adds r0, r0, r2
+; CHECK-NEXT: adc.w r2, r12, r3
+; CHECK-NEXT: ubfx r3, r1, #12, #1
+; CHECK-NEXT: ubfx r1, r1, #8, #1
+; CHECK-NEXT: rsbs r3, r3, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q2[2], q2[0], r1, r3
+; CHECK-NEXT: vmov q2[3], q2[1], r1, r3
+; CHECK-NEXT: vmov.u16 r1, q0[7]
+; CHECK-NEXT: vmov.u16 r3, q0[6]
+; CHECK-NEXT: vmov q0[2], q0[0], r3, r1
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vand q0, q0, q2
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r1, s1
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: vmov r3, s2
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: vmov r2, s3
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: vpop {d8, d9}
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <8 x i8> %b, zeroinitializer
+ %xx = zext <8 x i8> %x to <8 x i64>
+ %s = select <8 x i1> %c, <8 x i64> %xx, <8 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %s)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_sext(<8 x i8> %x, <8 x i8> %b) {
+; CHECK-LABEL: add_v8i8_v8i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmovlb.u8 q1, q1
+; CHECK-NEXT: vmov.i8 q2, #0xff
+; CHECK-NEXT: vcmp.i16 eq, q1, zr
+; CHECK-NEXT: vmov.i8 q1, #0x0
+; CHECK-NEXT: vpsel q1, q2, q1
+; CHECK-NEXT: vmov.u16 r0, q1[2]
+; CHECK-NEXT: vmov.u16 r1, q1[0]
+; CHECK-NEXT: vmov q2[2], q2[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[3]
+; CHECK-NEXT: vmov.u16 r1, q1[1]
+; CHECK-NEXT: vmov q2[3], q2[1], r1, r0
+; CHECK-NEXT: vcmp.i32 ne, q2, zr
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r1
+; CHECK-NEXT: vmov q2[3], q2[1], r2, r1
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vmov.u16 r2, q0[0]
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r1
+; CHECK-NEXT: asrs r1, r1, #31
+; CHECK-NEXT: asrs r2, r2, #31
+; CHECK-NEXT: vmov q3[3], q3[1], r2, r1
+; CHECK-NEXT: vand q2, q3, q2
+; CHECK-NEXT: vmov r3, s10
+; CHECK-NEXT: vmov r1, s8
+; CHECK-NEXT: vmov r12, s11
+; CHECK-NEXT: vmov r2, s9
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: ubfx r3, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsb.w r3, r3, #0
+; CHECK-NEXT: rsb.w r0, r0, #0
+; CHECK-NEXT: adc.w r2, r2, r12
+; CHECK-NEXT: vmov q2[2], q2[0], r0, r3
+; CHECK-NEXT: vmov q2[3], q2[1], r0, r3
+; CHECK-NEXT: vmov.u16 r0, q0[3]
+; CHECK-NEXT: vmov.u16 r3, q0[2]
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: vmov q3[2], q3[0], r3, r0
+; CHECK-NEXT: asrs r0, r0, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q3[3], q3[1], r3, r0
+; CHECK-NEXT: vand q2, q3, q2
+; CHECK-NEXT: vmov r3, s8
+; CHECK-NEXT: vmov r0, s9
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: vmov r3, s11
+; CHECK-NEXT: adcs r2, r0
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: adds.w r12, r1, r0
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: vmov.u16 r2, q1[6]
+; CHECK-NEXT: vmov.u16 r3, q1[4]
+; CHECK-NEXT: vmov q2[2], q2[0], r3, r2
+; CHECK-NEXT: vmov.u16 r2, q1[7]
+; CHECK-NEXT: vmov.u16 r3, q1[5]
+; CHECK-NEXT: vmov q2[3], q2[1], r3, r2
+; CHECK-NEXT: vcmp.i32 ne, q2, zr
+; CHECK-NEXT: vmrs r2, p0
+; CHECK-NEXT: and r0, r2, #1
+; CHECK-NEXT: ubfx r3, r2, #4, #1
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: rsbs r3, r3, #0
+; CHECK-NEXT: vmov q1[2], q1[0], r0, r3
+; CHECK-NEXT: vmov q1[3], q1[1], r0, r3
+; CHECK-NEXT: vmov.u16 r0, q0[5]
+; CHECK-NEXT: vmov.u16 r3, q0[4]
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: vmov q2[2], q2[0], r3, r0
+; CHECK-NEXT: asrs r0, r0, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q2[3], q2[1], r3, r0
+; CHECK-NEXT: vand q1, q2, q1
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: vmov r0, s5
+; CHECK-NEXT: adds.w r3, r3, r12
+; CHECK-NEXT: adc.w r12, r1, r0
+; CHECK-NEXT: vmov r0, s6
+; CHECK-NEXT: vmov r1, s7
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: ubfx r3, r2, #12, #1
+; CHECK-NEXT: ubfx r2, r2, #8, #1
+; CHECK-NEXT: rsb.w r3, r3, #0
+; CHECK-NEXT: rsb.w r2, r2, #0
+; CHECK-NEXT: adc.w r1, r1, r12
+; CHECK-NEXT: vmov q1[2], q1[0], r2, r3
+; CHECK-NEXT: vmov q1[3], q1[1], r2, r3
+; CHECK-NEXT: vmov.u16 r2, q0[7]
+; CHECK-NEXT: vmov.u16 r3, q0[6]
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: vmov q0[2], q0[0], r3, r2
+; CHECK-NEXT: asrs r2, r2, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q0[3], q0[1], r3, r2
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r2, s1
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: vmov r3, s2
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: vmov r2, s3
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <8 x i8> %b, zeroinitializer
+ %xx = sext <8 x i8> %x to <8 x i64>
+ %s = select <8 x i1> %c, <8 x i64> %xx, <8 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %s)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_zext(<4 x i8> %x, <4 x i8> %b) {
+; CHECK-LABEL: add_v4i8_v4i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .vsave {d8, d9}
+; CHECK-NEXT: vpush {d8, d9}
+; CHECK-NEXT: vmov.i32 q3, #0xff
+; CHECK-NEXT: vmov.i64 q2, #0xffffffff
+; CHECK-NEXT: vand q1, q1, q3
+; CHECK-NEXT: vand q0, q0, q3
+; CHECK-NEXT: vcmp.i32 eq, q1, zr
+; CHECK-NEXT: vmov.f32 s16, s0
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: vmov.f32 s18, s1
+; CHECK-NEXT: vand q4, q4, q2
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q1[2], q1[0], r2, r1
+; CHECK-NEXT: vmov q1[3], q1[1], r2, r1
+; CHECK-NEXT: vand q1, q4, q1
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: vmov r1, s4
+; CHECK-NEXT: vmov r12, s7
+; CHECK-NEXT: vmov r2, s5
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: vand q0, q1, q2
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: ubfx r3, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsb.w r3, r3, #0
+; CHECK-NEXT: rsb.w r0, r0, #0
+; CHECK-NEXT: adc.w r2, r2, r12
+; CHECK-NEXT: vmov q1[2], q1[0], r0, r3
+; CHECK-NEXT: vmov q1[3], q1[1], r0, r3
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adcs r2, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: vpop {d8, d9}
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <4 x i8> %b, zeroinitializer
+ %xx = zext <4 x i8> %x to <4 x i64>
+ %s = select <4 x i1> %c, <4 x i64> %xx, <4 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %s)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_sext(<4 x i8> %x, <4 x i8> %b) {
+; CHECK-LABEL: add_v4i8_v4i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.i32 q2, #0xff
+; CHECK-NEXT: vand q1, q1, q2
+; CHECK-NEXT: vmov.f32 s8, s0
+; CHECK-NEXT: vcmp.i32 eq, q1, zr
+; CHECK-NEXT: vmov.f32 s10, s1
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q1[2], q1[0], r2, r1
+; CHECK-NEXT: vmov q1[3], q1[1], r2, r1
+; CHECK-NEXT: vmov r1, s10
+; CHECK-NEXT: vmov r2, s0
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r1
+; CHECK-NEXT: asrs r1, r1, #31
+; CHECK-NEXT: asrs r2, r2, #31
+; CHECK-NEXT: vmov q2[3], q2[1], r2, r1
+; CHECK-NEXT: vand q1, q2, q1
+; CHECK-NEXT: vmov.f32 s8, s2
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: vmov r1, s4
+; CHECK-NEXT: vmov.f32 s10, s3
+; CHECK-NEXT: vmov r12, s7
+; CHECK-NEXT: vmov r2, s5
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: ubfx r3, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsb.w r3, r3, #0
+; CHECK-NEXT: rsb.w r0, r0, #0
+; CHECK-NEXT: vmov q1[2], q1[0], r0, r3
+; CHECK-NEXT: adc.w r2, r2, r12
+; CHECK-NEXT: vmov q1[3], q1[1], r0, r3
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: vmov r3, s8
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: vmov q0[2], q0[0], r3, r0
+; CHECK-NEXT: asrs r0, r0, #31
+; CHECK-NEXT: asrs r3, r3, #31
+; CHECK-NEXT: vmov q0[3], q0[1], r3, r0
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adcs r2, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <4 x i8> %b, zeroinitializer
+ %xx = sext <4 x i8> %x to <4 x i64>
+ %s = select <4 x i1> %c, <4 x i64> %xx, <4 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %s)
+ ret i64 %z
+}
+
define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_zext(<2 x i8> %x, <2 x i8> %b) {
; CHECK-LABEL: add_v2i8_v2i64_zext:
; CHECK: @ %bb.0: @ %entry
diff --git a/llvm/test/CodeGen/Thumb2/mve-vecreduce-mla.ll b/llvm/test/CodeGen/Thumb2/mve-vecreduce-mla.ll
index efc55e209756..ba2651e43ea5 100644
--- a/llvm/test/CodeGen/Thumb2/mve-vecreduce-mla.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-vecreduce-mla.ll
@@ -170,6 +170,94 @@ entry:
ret i64 %z
}
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_zext(<4 x i16> %x, <4 x i16> %y) {
+; CHECK-LABEL: add_v4i16_v4i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .vsave {d8, d9}
+; CHECK-NEXT: vpush {d8, d9}
+; CHECK-NEXT: vmovlb.u16 q1, q1
+; CHECK-NEXT: vmovlb.u16 q0, q0
+; CHECK-NEXT: vmov.f32 s8, s4
+; CHECK-NEXT: vmov.f32 s12, s0
+; CHECK-NEXT: vmov.f32 s10, s5
+; CHECK-NEXT: vmov.f32 s14, s1
+; CHECK-NEXT: vmullb.u32 q4, q3, q2
+; CHECK-NEXT: vmov.f32 s8, s6
+; CHECK-NEXT: vmov r2, s18
+; CHECK-NEXT: vmov r3, s16
+; CHECK-NEXT: vmov r0, s19
+; CHECK-NEXT: vmov r1, s17
+; CHECK-NEXT: vmov.f32 s10, s7
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: vmullb.u32 q0, q1, q2
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: adcs r0, r1
+; CHECK-NEXT: vmov r1, s1
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adcs r1, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r2
+; CHECK-NEXT: adcs r1, r3
+; CHECK-NEXT: vpop {d8, d9}
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <4 x i16> %x to <4 x i64>
+ %yy = zext <4 x i16> %y to <4 x i64>
+ %m = mul <4 x i64> %xx, %yy
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_sext(<4 x i16> %x, <4 x i16> %y) {
+; CHECK-LABEL: add_v4i16_v4i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.f32 s8, s4
+; CHECK-NEXT: vmov.f32 s10, s5
+; CHECK-NEXT: vmov r2, s4
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: vmov.f32 s8, s0
+; CHECK-NEXT: vmov.f32 s10, s1
+; CHECK-NEXT: vmov r1, s10
+; CHECK-NEXT: sxth r2, r2
+; CHECK-NEXT: sxth r3, r3
+; CHECK-NEXT: smull r2, r3, r3, r2
+; CHECK-NEXT: sxth r0, r0
+; CHECK-NEXT: sxth r1, r1
+; CHECK-NEXT: smull r0, r1, r1, r0
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r0
+; CHECK-NEXT: vmov q2[3], q2[1], r3, r1
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: vmov r3, s8
+; CHECK-NEXT: vmov r2, s9
+; CHECK-NEXT: vmov.f32 s8, s6
+; CHECK-NEXT: vmov.f32 s10, s7
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: vmov r2, s8
+; CHECK-NEXT: sxth r3, r3
+; CHECK-NEXT: sxth r2, r2
+; CHECK-NEXT: smlal r0, r1, r3, r2
+; CHECK-NEXT: vmov r2, s10
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: sxth r2, r2
+; CHECK-NEXT: sxth r3, r3
+; CHECK-NEXT: smlal r0, r1, r3, r2
+; CHECK-NEXT: bx lr
+entry:
+ %xx = sext <4 x i16> %x to <4 x i64>
+ %yy = sext <4 x i16> %y to <4 x i64>
+ %m = mul <4 x i64> %xx, %yy
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
+ ret i64 %z
+}
+
define arm_aapcs_vfpcc i64 @add_v8i16_v8i32_v8i64_zext(<8 x i16> %x, <8 x i16> %y) {
; CHECK-LABEL: add_v8i16_v8i32_v8i64_zext:
; CHECK: @ %bb.0: @ %entry
@@ -280,6 +368,96 @@ entry:
ret i32 %z
}
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_zext(<8 x i8> %x, <8 x i8> %y) {
+; CHECK-LABEL: add_v8i8_v8i32_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmovlb.u8 q1, q1
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vmov.u16 r0, q1[6]
+; CHECK-NEXT: vmov.u16 r1, q1[4]
+; CHECK-NEXT: vmov q2[2], q2[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[7]
+; CHECK-NEXT: vmov.u16 r1, q1[5]
+; CHECK-NEXT: vmov q2[3], q2[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[6]
+; CHECK-NEXT: vmov.u16 r1, q0[4]
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[7]
+; CHECK-NEXT: vmov.u16 r1, q0[5]
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[2]
+; CHECK-NEXT: vmov.u16 r1, q1[0]
+; CHECK-NEXT: vmullb.u16 q2, q3, q2
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[3]
+; CHECK-NEXT: vmov.u16 r1, q1[1]
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[2]
+; CHECK-NEXT: vmov.u16 r1, q0[0]
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[3]
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r0
+; CHECK-NEXT: vmullb.u16 q0, q1, q3
+; CHECK-NEXT: vadd.i32 q0, q0, q2
+; CHECK-NEXT: vaddv.u32 r0, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <8 x i8> %x to <8 x i32>
+ %yy = zext <8 x i8> %y to <8 x i32>
+ %m = mul <8 x i32> %xx, %yy
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
+ ret i32 %z
+}
+
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_sext(<8 x i8> %x, <8 x i8> %y) {
+; CHECK-LABEL: add_v8i8_v8i32_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.u16 r0, q1[6]
+; CHECK-NEXT: vmov.u16 r1, q1[4]
+; CHECK-NEXT: vmov q2[2], q2[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[7]
+; CHECK-NEXT: vmov.u16 r1, q1[5]
+; CHECK-NEXT: vmov q2[3], q2[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[6]
+; CHECK-NEXT: vmov.u16 r1, q0[4]
+; CHECK-NEXT: vmovlb.s8 q2, q2
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[7]
+; CHECK-NEXT: vmov.u16 r1, q0[5]
+; CHECK-NEXT: vmovlb.s16 q2, q2
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[2]
+; CHECK-NEXT: vmovlb.s8 q3, q3
+; CHECK-NEXT: vmov.u16 r1, q1[0]
+; CHECK-NEXT: vmovlb.s16 q3, q3
+; CHECK-NEXT: vmul.i32 q2, q3, q2
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[3]
+; CHECK-NEXT: vmov.u16 r1, q1[1]
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[2]
+; CHECK-NEXT: vmov.u16 r1, q0[0]
+; CHECK-NEXT: vmovlb.s8 q1, q3
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[3]
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vmovlb.s16 q1, q1
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmovlb.s8 q0, q3
+; CHECK-NEXT: vmovlb.s16 q0, q0
+; CHECK-NEXT: vmul.i32 q0, q0, q1
+; CHECK-NEXT: vadd.i32 q0, q0, q2
+; CHECK-NEXT: vaddv.u32 r0, q0
+; CHECK-NEXT: bx lr
+entry:
+ %xx = sext <8 x i8> %x to <8 x i32>
+ %yy = sext <8 x i8> %y to <8 x i32>
+ %m = mul <8 x i32> %xx, %yy
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %m)
+ ret i32 %z
+}
+
define arm_aapcs_vfpcc i32 @add_v16i8_v16i16_v16i32_zext(<16 x i8> %x, <16 x i8> %y) {
; CHECK-LABEL: add_v16i8_v16i16_v16i32_zext:
; CHECK: @ %bb.0: @ %entry
@@ -767,6 +945,275 @@ entry:
ret i64 %z
}
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_zext(<8 x i8> %x, <8 x i8> %y) {
+; CHECK-LABEL: add_v8i8_v8i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r7, lr}
+; CHECK-NEXT: push {r7, lr}
+; CHECK-NEXT: .vsave {d8, d9}
+; CHECK-NEXT: vpush {d8, d9}
+; CHECK-NEXT: vmovlb.u8 q2, q1
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vmov.u16 r0, q2[1]
+; CHECK-NEXT: vmov.u16 r1, q2[0]
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vmov.u16 r2, q0[0]
+; CHECK-NEXT: vmov.i64 q1, #0xffff
+; CHECK-NEXT: vmov q4[2], q4[0], r2, r1
+; CHECK-NEXT: vand q3, q3, q1
+; CHECK-NEXT: vand q4, q4, q1
+; CHECK-NEXT: vmov r0, s14
+; CHECK-NEXT: vmov r1, s18
+; CHECK-NEXT: vmov r2, s12
+; CHECK-NEXT: vmov r3, s16
+; CHECK-NEXT: umull r0, r1, r1, r0
+; CHECK-NEXT: umull r2, r3, r3, r2
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r0
+; CHECK-NEXT: vmov q3[3], q3[1], r3, r1
+; CHECK-NEXT: vmov r2, s14
+; CHECK-NEXT: vmov r3, s12
+; CHECK-NEXT: vmov r0, s13
+; CHECK-NEXT: adds.w lr, r3, r2
+; CHECK-NEXT: vmov.u16 r3, q2[2]
+; CHECK-NEXT: adc.w r12, r0, r1
+; CHECK-NEXT: vmov.u16 r1, q2[3]
+; CHECK-NEXT: vmov q3[2], q3[0], r3, r1
+; CHECK-NEXT: vmov.u16 r3, q0[3]
+; CHECK-NEXT: vmov.u16 r0, q0[2]
+; CHECK-NEXT: vand q3, q3, q1
+; CHECK-NEXT: vmov q4[2], q4[0], r0, r3
+; CHECK-NEXT: vmov r1, s14
+; CHECK-NEXT: vand q4, q4, q1
+; CHECK-NEXT: vmov r3, s12
+; CHECK-NEXT: vmov r0, s18
+; CHECK-NEXT: vmov r2, s16
+; CHECK-NEXT: umull r0, r1, r0, r1
+; CHECK-NEXT: umull r2, r3, r2, r3
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r0
+; CHECK-NEXT: vmov q3[3], q3[1], r3, r1
+; CHECK-NEXT: vmov r2, s12
+; CHECK-NEXT: vmov r0, s13
+; CHECK-NEXT: vmov r3, s14
+; CHECK-NEXT: adds.w r2, r2, lr
+; CHECK-NEXT: adc.w r0, r0, r12
+; CHECK-NEXT: adds.w lr, r2, r3
+; CHECK-NEXT: vmov.u16 r3, q2[4]
+; CHECK-NEXT: adc.w r12, r0, r1
+; CHECK-NEXT: vmov.u16 r1, q2[5]
+; CHECK-NEXT: vmov q3[2], q3[0], r3, r1
+; CHECK-NEXT: vmov.u16 r3, q0[5]
+; CHECK-NEXT: vmov.u16 r0, q0[4]
+; CHECK-NEXT: vand q3, q3, q1
+; CHECK-NEXT: vmov q4[2], q4[0], r0, r3
+; CHECK-NEXT: vmov r1, s14
+; CHECK-NEXT: vand q4, q4, q1
+; CHECK-NEXT: vmov r3, s12
+; CHECK-NEXT: vmov r0, s18
+; CHECK-NEXT: vmov r2, s16
+; CHECK-NEXT: umull r0, r1, r0, r1
+; CHECK-NEXT: umull r2, r3, r2, r3
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r0
+; CHECK-NEXT: vmov q3[3], q3[1], r3, r1
+; CHECK-NEXT: vmov r2, s12
+; CHECK-NEXT: vmov r0, s13
+; CHECK-NEXT: adds.w r2, r2, lr
+; CHECK-NEXT: adc.w r3, r12, r0
+; CHECK-NEXT: vmov r0, s14
+; CHECK-NEXT: adds r0, r0, r2
+; CHECK-NEXT: vmov.u16 r2, q2[7]
+; CHECK-NEXT: adcs r1, r3
+; CHECK-NEXT: vmov.u16 r3, q2[6]
+; CHECK-NEXT: vmov q2[2], q2[0], r3, r2
+; CHECK-NEXT: vmov.u16 r3, q0[7]
+; CHECK-NEXT: vmov.u16 r2, q0[6]
+; CHECK-NEXT: vand q2, q2, q1
+; CHECK-NEXT: vmov q0[2], q0[0], r2, r3
+; CHECK-NEXT: vmov r12, s8
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r2, s0
+; CHECK-NEXT: vmov r3, s2
+; CHECK-NEXT: umlal r0, r1, r2, r12
+; CHECK-NEXT: vmov r2, s10
+; CHECK-NEXT: umlal r0, r1, r3, r2
+; CHECK-NEXT: vpop {d8, d9}
+; CHECK-NEXT: pop {r7, pc}
+entry:
+ %xx = zext <8 x i8> %x to <8 x i64>
+ %yy = zext <8 x i8> %y to <8 x i64>
+ %m = mul <8 x i64> %xx, %yy
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_sext(<8 x i8> %x, <8 x i8> %y) {
+; CHECK-LABEL: add_v8i8_v8i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r7, lr}
+; CHECK-NEXT: push {r7, lr}
+; CHECK-NEXT: vmov.u16 r0, q1[1]
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vmov.u16 r2, q1[0]
+; CHECK-NEXT: vmov.u16 r3, q0[0]
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: smull r0, r1, r1, r0
+; CHECK-NEXT: smull r2, r3, r3, r2
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r0
+; CHECK-NEXT: vmov q2[3], q2[1], r3, r1
+; CHECK-NEXT: vmov r2, s10
+; CHECK-NEXT: vmov r3, s8
+; CHECK-NEXT: vmov r0, s9
+; CHECK-NEXT: adds.w lr, r3, r2
+; CHECK-NEXT: vmov.u16 r3, q0[3]
+; CHECK-NEXT: adc.w r12, r0, r1
+; CHECK-NEXT: vmov.u16 r1, q1[3]
+; CHECK-NEXT: vmov.u16 r0, q1[2]
+; CHECK-NEXT: vmov.u16 r2, q0[2]
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: smull r1, r3, r3, r1
+; CHECK-NEXT: smull r0, r2, r2, r0
+; CHECK-NEXT: vmov q2[2], q2[0], r0, r1
+; CHECK-NEXT: vmov q2[3], q2[1], r2, r3
+; CHECK-NEXT: vmov r1, s8
+; CHECK-NEXT: vmov r0, s9
+; CHECK-NEXT: vmov r2, s10
+; CHECK-NEXT: adds.w r1, r1, lr
+; CHECK-NEXT: adc.w r0, r0, r12
+; CHECK-NEXT: adds.w lr, r1, r2
+; CHECK-NEXT: vmov.u16 r2, q1[5]
+; CHECK-NEXT: adc.w r12, r0, r3
+; CHECK-NEXT: vmov.u16 r3, q0[5]
+; CHECK-NEXT: vmov.u16 r0, q1[4]
+; CHECK-NEXT: vmov.u16 r1, q0[4]
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: smull r2, r3, r3, r2
+; CHECK-NEXT: smull r0, r1, r1, r0
+; CHECK-NEXT: vmov q2[2], q2[0], r0, r2
+; CHECK-NEXT: vmov q2[3], q2[1], r1, r3
+; CHECK-NEXT: vmov r1, s8
+; CHECK-NEXT: vmov r0, s9
+; CHECK-NEXT: adds.w r1, r1, lr
+; CHECK-NEXT: adc.w r2, r12, r0
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: vmov.u16 r2, q1[6]
+; CHECK-NEXT: vmov.u16 r3, q0[6]
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: smlal r0, r1, r3, r2
+; CHECK-NEXT: vmov.u16 r2, q1[7]
+; CHECK-NEXT: vmov.u16 r3, q0[7]
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: smlal r0, r1, r3, r2
+; CHECK-NEXT: pop {r7, pc}
+entry:
+ %xx = sext <8 x i8> %x to <8 x i64>
+ %yy = sext <8 x i8> %y to <8 x i64>
+ %m = mul <8 x i64> %xx, %yy
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %m)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_zext(<4 x i8> %x, <4 x i8> %y) {
+; CHECK-LABEL: add_v4i8_v4i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .vsave {d8, d9}
+; CHECK-NEXT: vpush {d8, d9}
+; CHECK-NEXT: vmov.i32 q2, #0xff
+; CHECK-NEXT: vand q1, q1, q2
+; CHECK-NEXT: vand q0, q0, q2
+; CHECK-NEXT: vmov.f32 s12, s4
+; CHECK-NEXT: vmov.f32 s8, s0
+; CHECK-NEXT: vmov.f32 s14, s5
+; CHECK-NEXT: vmov.f32 s10, s1
+; CHECK-NEXT: vmullb.u32 q4, q2, q3
+; CHECK-NEXT: vmov.f32 s8, s6
+; CHECK-NEXT: vmov r2, s18
+; CHECK-NEXT: vmov r3, s16
+; CHECK-NEXT: vmov r0, s19
+; CHECK-NEXT: vmov r1, s17
+; CHECK-NEXT: vmov.f32 s10, s7
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: vmullb.u32 q0, q1, q2
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: adcs r0, r1
+; CHECK-NEXT: vmov r1, s1
+; CHECK-NEXT: adds r2, r2, r3
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adcs r1, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r2
+; CHECK-NEXT: adcs r1, r3
+; CHECK-NEXT: vpop {d8, d9}
+; CHECK-NEXT: bx lr
+entry:
+ %xx = zext <4 x i8> %x to <4 x i64>
+ %yy = zext <4 x i8> %y to <4 x i64>
+ %m = mul <4 x i64> %xx, %yy
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_sext(<4 x i8> %x, <4 x i8> %y) {
+; CHECK-LABEL: add_v4i8_v4i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: vmov.f32 s8, s4
+; CHECK-NEXT: vmov.f32 s10, s5
+; CHECK-NEXT: vmov r2, s4
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: vmov.f32 s8, s0
+; CHECK-NEXT: vmov.f32 s10, s1
+; CHECK-NEXT: vmov r1, s10
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: smull r2, r3, r3, r2
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: smull r0, r1, r1, r0
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r0
+; CHECK-NEXT: vmov q2[3], q2[1], r3, r1
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: vmov r3, s8
+; CHECK-NEXT: vmov r2, s9
+; CHECK-NEXT: vmov.f32 s8, s6
+; CHECK-NEXT: vmov.f32 s10, s7
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: adds r0, r0, r3
+; CHECK-NEXT: vmov r3, s4
+; CHECK-NEXT: adcs r1, r2
+; CHECK-NEXT: vmov r2, s8
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: smlal r0, r1, r3, r2
+; CHECK-NEXT: vmov r2, s10
+; CHECK-NEXT: vmov r3, s6
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: smlal r0, r1, r3, r2
+; CHECK-NEXT: bx lr
+entry:
+ %xx = sext <4 x i8> %x to <4 x i64>
+ %yy = sext <4 x i8> %y to <4 x i64>
+ %m = mul <4 x i64> %xx, %yy
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %m)
+ ret i64 %z
+}
+
define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_zext(<2 x i8> %x, <2 x i8> %y) {
; CHECK-LABEL: add_v2i8_v2i64_zext:
; CHECK: @ %bb.0: @ %entry
diff --git a/llvm/test/CodeGen/Thumb2/mve-vecreduce-mlapred.ll b/llvm/test/CodeGen/Thumb2/mve-vecreduce-mlapred.ll
index dbdb9fc925d7..e4252019c3d8 100644
--- a/llvm/test/CodeGen/Thumb2/mve-vecreduce-mlapred.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-vecreduce-mlapred.ll
@@ -282,6 +282,145 @@ entry:
ret i64 %z
}
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_zext(<4 x i16> %x, <4 x i16> %y, <4 x i16> %b) {
+; CHECK-LABEL: add_v4i16_v4i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .vsave {d8, d9, d10, d11}
+; CHECK-NEXT: vpush {d8, d9, d10, d11}
+; CHECK-NEXT: vmovlb.u16 q2, q2
+; CHECK-NEXT: vmovlb.u16 q1, q1
+; CHECK-NEXT: vcmp.i32 eq, q2, zr
+; CHECK-NEXT: vmovlb.u16 q0, q0
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: vmov.f32 s12, s4
+; CHECK-NEXT: vmov.f32 s16, s0
+; CHECK-NEXT: vmov.f32 s14, s5
+; CHECK-NEXT: vmov.f32 s18, s1
+; CHECK-NEXT: vmullb.u32 q5, q4, q3
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r1
+; CHECK-NEXT: vmov q2[3], q2[1], r2, r1
+; CHECK-NEXT: vand q2, q5, q2
+; CHECK-NEXT: vmov r3, s10
+; CHECK-NEXT: vmov r1, s8
+; CHECK-NEXT: vmov r12, s11
+; CHECK-NEXT: vmov r2, s9
+; CHECK-NEXT: vmov.f32 s8, s6
+; CHECK-NEXT: vmov.f32 s10, s7
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: vmullb.u32 q0, q1, q2
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: ubfx r3, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsb.w r3, r3, #0
+; CHECK-NEXT: rsb.w r0, r0, #0
+; CHECK-NEXT: adc.w r2, r2, r12
+; CHECK-NEXT: vmov q1[2], q1[0], r0, r3
+; CHECK-NEXT: vmov q1[3], q1[1], r0, r3
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adcs r2, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: vpop {d8, d9, d10, d11}
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <4 x i16> %b, zeroinitializer
+ %xx = zext <4 x i16> %x to <4 x i64>
+ %yy = zext <4 x i16> %y to <4 x i64>
+ %m = mul <4 x i64> %xx, %yy
+ %s = select <4 x i1> %c, <4 x i64> %m, <4 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %s)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_sext(<4 x i16> %x, <4 x i16> %y, <4 x i16> %b) {
+; CHECK-LABEL: add_v4i16_v4i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r4, lr}
+; CHECK-NEXT: push {r4, lr}
+; CHECK-NEXT: vmov.f32 s12, s4
+; CHECK-NEXT: vmovlb.u16 q2, q2
+; CHECK-NEXT: vmov.f32 s14, s5
+; CHECK-NEXT: vcmp.i32 eq, q2, zr
+; CHECK-NEXT: vmov r2, s4
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r0, s14
+; CHECK-NEXT: vmov.f32 s12, s0
+; CHECK-NEXT: vmov.f32 s14, s1
+; CHECK-NEXT: vmov r1, s14
+; CHECK-NEXT: sxth r2, r2
+; CHECK-NEXT: sxth r3, r3
+; CHECK-NEXT: smull r2, r3, r3, r2
+; CHECK-NEXT: sxth r0, r0
+; CHECK-NEXT: sxth r1, r1
+; CHECK-NEXT: smull r0, r1, r1, r0
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r0
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: vmov q3[3], q3[1], r3, r1
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r1
+; CHECK-NEXT: vmov q2[3], q2[1], r2, r1
+; CHECK-NEXT: vand q2, q3, q2
+; CHECK-NEXT: vmov r1, s10
+; CHECK-NEXT: vmov r2, s8
+; CHECK-NEXT: vmov r12, s11
+; CHECK-NEXT: vmov r3, s9
+; CHECK-NEXT: vmov.f32 s8, s6
+; CHECK-NEXT: vmov.f32 s10, s7
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: vmov r4, s4
+; CHECK-NEXT: adds.w lr, r2, r1
+; CHECK-NEXT: vmov r1, s6
+; CHECK-NEXT: vmov r2, s8
+; CHECK-NEXT: adc.w r12, r12, r3
+; CHECK-NEXT: vmov r3, s10
+; CHECK-NEXT: sxth r4, r4
+; CHECK-NEXT: sxth r1, r1
+; CHECK-NEXT: sxth r2, r2
+; CHECK-NEXT: sxth r3, r3
+; CHECK-NEXT: smull r2, r4, r4, r2
+; CHECK-NEXT: smull r1, r3, r1, r3
+; CHECK-NEXT: vmov q0[2], q0[0], r2, r1
+; CHECK-NEXT: ubfx r1, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: vmov q0[3], q0[1], r4, r3
+; CHECK-NEXT: vmov q1[2], q1[0], r0, r1
+; CHECK-NEXT: vmov q1[3], q1[1], r0, r1
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r1, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adds.w r1, r1, lr
+; CHECK-NEXT: adc.w r2, r12, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: pop {r4, pc}
+entry:
+ %c = icmp eq <4 x i16> %b, zeroinitializer
+ %xx = sext <4 x i16> %x to <4 x i64>
+ %yy = sext <4 x i16> %y to <4 x i64>
+ %m = mul <4 x i64> %xx, %yy
+ %s = select <4 x i1> %c, <4 x i64> %m, <4 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %s)
+ ret i64 %z
+}
+
define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_zext(<2 x i16> %x, <2 x i16> %y, <2 x i16> %b) {
; CHECK-LABEL: add_v2i16_v2i64_zext:
; CHECK: @ %bb.0: @ %entry
@@ -456,6 +595,150 @@ entry:
ret i32 %z
}
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_zext(<8 x i8> %x, <8 x i8> %y, <8 x i8> %b) {
+; CHECK-LABEL: add_v8i8_v8i32_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .vsave {d8, d9, d10, d11}
+; CHECK-NEXT: vpush {d8, d9, d10, d11}
+; CHECK-NEXT: vmovlb.u8 q1, q1
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vmov.u16 r0, q1[2]
+; CHECK-NEXT: vmov.u16 r1, q1[0]
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[3]
+; CHECK-NEXT: vmov.u16 r1, q1[1]
+; CHECK-NEXT: vmovlb.u8 q2, q2
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[2]
+; CHECK-NEXT: vmov.u16 r1, q0[0]
+; CHECK-NEXT: vmovlb.u16 q4, q3
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[3]
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vcmp.i16 eq, q2, zr
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.i8 q2, #0x0
+; CHECK-NEXT: vmovlb.u16 q5, q3
+; CHECK-NEXT: vmov.i8 q3, #0xff
+; CHECK-NEXT: vpsel q2, q3, q2
+; CHECK-NEXT: vmov.u16 r0, q2[2]
+; CHECK-NEXT: vmov.u16 r1, q2[0]
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q2[3]
+; CHECK-NEXT: vmov.u16 r1, q2[1]
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[6]
+; CHECK-NEXT: vcmp.i32 ne, q3, zr
+; CHECK-NEXT: vmov.i32 q3, #0x0
+; CHECK-NEXT: vmov.u16 r1, q1[4]
+; CHECK-NEXT: vpst
+; CHECK-NEXT: vmult.i32 q3, q5, q4
+; CHECK-NEXT: vmov q4[2], q4[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[7]
+; CHECK-NEXT: vmov.u16 r1, q1[5]
+; CHECK-NEXT: vmov q4[3], q4[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[6]
+; CHECK-NEXT: vmov.u16 r1, q0[4]
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[7]
+; CHECK-NEXT: vmov.u16 r1, q0[5]
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q2[6]
+; CHECK-NEXT: vmov.u16 r1, q2[4]
+; CHECK-NEXT: vmullb.u16 q0, q1, q4
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q2[7]
+; CHECK-NEXT: vmov.u16 r1, q2[5]
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r0
+; CHECK-NEXT: vpt.i32 ne, q1, zr
+; CHECK-NEXT: vaddt.i32 q3, q3, q0
+; CHECK-NEXT: vaddv.u32 r0, q3
+; CHECK-NEXT: vpop {d8, d9, d10, d11}
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <8 x i8> %b, zeroinitializer
+ %xx = zext <8 x i8> %x to <8 x i32>
+ %yy = zext <8 x i8> %y to <8 x i32>
+ %m = mul <8 x i32> %xx, %yy
+ %s = select <8 x i1> %c, <8 x i32> %m, <8 x i32> zeroinitializer
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %s)
+ ret i32 %z
+}
+
+define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_sext(<8 x i8> %x, <8 x i8> %y, <8 x i8> %b) {
+; CHECK-LABEL: add_v8i8_v8i32_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .vsave {d8, d9, d10, d11}
+; CHECK-NEXT: vpush {d8, d9, d10, d11}
+; CHECK-NEXT: vmov.u16 r0, q1[2]
+; CHECK-NEXT: vmov.u16 r1, q1[0]
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[3]
+; CHECK-NEXT: vmov.u16 r1, q1[1]
+; CHECK-NEXT: vmovlb.u8 q2, q2
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[2]
+; CHECK-NEXT: vmovlb.s8 q3, q3
+; CHECK-NEXT: vmov.u16 r1, q0[0]
+; CHECK-NEXT: vmovlb.s16 q4, q3
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[3]
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vcmp.i16 eq, q2, zr
+; CHECK-NEXT: vmovlb.s8 q3, q3
+; CHECK-NEXT: vmov.i8 q2, #0x0
+; CHECK-NEXT: vmovlb.s16 q5, q3
+; CHECK-NEXT: vmov.i8 q3, #0xff
+; CHECK-NEXT: vpsel q2, q3, q2
+; CHECK-NEXT: vmov.u16 r0, q2[2]
+; CHECK-NEXT: vmov.u16 r1, q2[0]
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q2[3]
+; CHECK-NEXT: vmov.u16 r1, q2[1]
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[6]
+; CHECK-NEXT: vcmp.i32 ne, q3, zr
+; CHECK-NEXT: vmov.i32 q3, #0x0
+; CHECK-NEXT: vmov.u16 r1, q1[4]
+; CHECK-NEXT: vpst
+; CHECK-NEXT: vmult.i32 q3, q5, q4
+; CHECK-NEXT: vmov q4[2], q4[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q1[7]
+; CHECK-NEXT: vmov.u16 r1, q1[5]
+; CHECK-NEXT: vmov q4[3], q4[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[6]
+; CHECK-NEXT: vmov.u16 r1, q0[4]
+; CHECK-NEXT: vmovlb.s8 q1, q4
+; CHECK-NEXT: vmov q4[2], q4[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q0[7]
+; CHECK-NEXT: vmov.u16 r1, q0[5]
+; CHECK-NEXT: vmovlb.s16 q1, q1
+; CHECK-NEXT: vmov q4[3], q4[1], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q2[6]
+; CHECK-NEXT: vmovlb.s8 q0, q4
+; CHECK-NEXT: vmov.u16 r1, q2[4]
+; CHECK-NEXT: vmovlb.s16 q0, q0
+; CHECK-NEXT: vmul.i32 q0, q0, q1
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q2[7]
+; CHECK-NEXT: vmov.u16 r1, q2[5]
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r0
+; CHECK-NEXT: vpt.i32 ne, q1, zr
+; CHECK-NEXT: vaddt.i32 q3, q3, q0
+; CHECK-NEXT: vaddv.u32 r0, q3
+; CHECK-NEXT: vpop {d8, d9, d10, d11}
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <8 x i8> %b, zeroinitializer
+ %xx = sext <8 x i8> %x to <8 x i32>
+ %yy = sext <8 x i8> %y to <8 x i32>
+ %m = mul <8 x i32> %xx, %yy
+ %s = select <8 x i1> %c, <8 x i32> %m, <8 x i32> zeroinitializer
+ %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %s)
+ ret i32 %z
+}
+
define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_zext(<4 x i8> %x, <4 x i8> %y, <4 x i8> %b) {
; CHECK-LABEL: add_v4i8_v4i32_zext:
; CHECK: @ %bb.0: @ %entry
@@ -1240,6 +1523,461 @@ entry:
ret i64 %z
}
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_zext(<8 x i8> %x, <8 x i8> %y, <8 x i8> %b) {
+; CHECK-LABEL: add_v8i8_v8i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r4, lr}
+; CHECK-NEXT: push {r4, lr}
+; CHECK-NEXT: .vsave {d8, d9, d10, d11}
+; CHECK-NEXT: vpush {d8, d9, d10, d11}
+; CHECK-NEXT: vmovlb.u8 q3, q1
+; CHECK-NEXT: vmovlb.u8 q0, q0
+; CHECK-NEXT: vmov.u16 r0, q3[1]
+; CHECK-NEXT: vmov.u16 r1, q3[0]
+; CHECK-NEXT: vmov q4[2], q4[0], r1, r0
+; CHECK-NEXT: vmov.u16 r1, q0[1]
+; CHECK-NEXT: vmov.u16 r2, q0[0]
+; CHECK-NEXT: vmov.i64 q1, #0xffff
+; CHECK-NEXT: vmov q5[2], q5[0], r2, r1
+; CHECK-NEXT: vand q4, q4, q1
+; CHECK-NEXT: vand q5, q5, q1
+; CHECK-NEXT: vmov r0, s18
+; CHECK-NEXT: vmov r1, s22
+; CHECK-NEXT: vmovlb.u8 q2, q2
+; CHECK-NEXT: vmov r3, s20
+; CHECK-NEXT: vcmp.i16 eq, q2, zr
+; CHECK-NEXT: vmov r2, s16
+; CHECK-NEXT: vmov.i8 q2, #0x0
+; CHECK-NEXT: vmov.i8 q5, #0xff
+; CHECK-NEXT: vpsel q2, q5, q2
+; CHECK-NEXT: umull r0, r1, r1, r0
+; CHECK-NEXT: umull r2, r3, r3, r2
+; CHECK-NEXT: vmov q4[2], q4[0], r2, r0
+; CHECK-NEXT: vmov.u16 r0, q2[2]
+; CHECK-NEXT: vmov q4[3], q4[1], r3, r1
+; CHECK-NEXT: vmov.u16 r1, q2[0]
+; CHECK-NEXT: vmov q5[2], q5[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q2[3]
+; CHECK-NEXT: vmov.u16 r1, q2[1]
+; CHECK-NEXT: vmov q5[3], q5[1], r1, r0
+; CHECK-NEXT: vcmp.i32 ne, q5, zr
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q5[2], q5[0], r2, r1
+; CHECK-NEXT: vmov q5[3], q5[1], r2, r1
+; CHECK-NEXT: vand q4, q4, q5
+; CHECK-NEXT: vmov r1, s18
+; CHECK-NEXT: vmov r2, s16
+; CHECK-NEXT: vmov r12, s19
+; CHECK-NEXT: vmov r3, s17
+; CHECK-NEXT: adds.w lr, r2, r1
+; CHECK-NEXT: vmov.u16 r1, q3[2]
+; CHECK-NEXT: vmov.u16 r2, q0[2]
+; CHECK-NEXT: adc.w r12, r12, r3
+; CHECK-NEXT: vmov.u16 r3, q3[3]
+; CHECK-NEXT: vmov q4[2], q4[0], r1, r3
+; CHECK-NEXT: vmov.u16 r3, q0[3]
+; CHECK-NEXT: vmov q5[2], q5[0], r2, r3
+; CHECK-NEXT: vand q4, q4, q1
+; CHECK-NEXT: vand q5, q5, q1
+; CHECK-NEXT: vmov r1, s18
+; CHECK-NEXT: vmov r2, s22
+; CHECK-NEXT: vmov r3, s16
+; CHECK-NEXT: vmov r4, s20
+; CHECK-NEXT: umull r1, r2, r2, r1
+; CHECK-NEXT: umull r3, r4, r4, r3
+; CHECK-NEXT: vmov q4[2], q4[0], r3, r1
+; CHECK-NEXT: ubfx r1, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: vmov q4[3], q4[1], r4, r2
+; CHECK-NEXT: vmov q5[2], q5[0], r0, r1
+; CHECK-NEXT: vmov.u16 r4, q0[4]
+; CHECK-NEXT: vmov q5[3], q5[1], r0, r1
+; CHECK-NEXT: vand q4, q4, q5
+; CHECK-NEXT: vmov r1, s16
+; CHECK-NEXT: vmov r0, s17
+; CHECK-NEXT: vmov r3, s19
+; CHECK-NEXT: adds.w r1, r1, lr
+; CHECK-NEXT: adc.w r2, r12, r0
+; CHECK-NEXT: vmov r0, s18
+; CHECK-NEXT: adds.w r12, r1, r0
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: vmov.u16 r2, q3[5]
+; CHECK-NEXT: vmov.u16 r3, q3[4]
+; CHECK-NEXT: vmov q4[2], q4[0], r3, r2
+; CHECK-NEXT: vmov.u16 r3, q0[5]
+; CHECK-NEXT: vmov q5[2], q5[0], r4, r3
+; CHECK-NEXT: vand q4, q4, q1
+; CHECK-NEXT: vand q5, q5, q1
+; CHECK-NEXT: vmov r2, s18
+; CHECK-NEXT: vmov r3, s22
+; CHECK-NEXT: vmov r4, s16
+; CHECK-NEXT: vmov r0, s20
+; CHECK-NEXT: umull r2, r3, r3, r2
+; CHECK-NEXT: umull r0, r4, r0, r4
+; CHECK-NEXT: vmov q4[2], q4[0], r0, r2
+; CHECK-NEXT: vmov.u16 r0, q2[6]
+; CHECK-NEXT: vmov.u16 r2, q2[4]
+; CHECK-NEXT: vmov q4[3], q4[1], r4, r3
+; CHECK-NEXT: vmov q5[2], q5[0], r2, r0
+; CHECK-NEXT: vmov.u16 r0, q2[7]
+; CHECK-NEXT: vmov.u16 r2, q2[5]
+; CHECK-NEXT: vmov q5[3], q5[1], r2, r0
+; CHECK-NEXT: vcmp.i32 ne, q5, zr
+; CHECK-NEXT: vmrs r2, p0
+; CHECK-NEXT: and r3, r2, #1
+; CHECK-NEXT: ubfx r0, r2, #4, #1
+; CHECK-NEXT: rsbs r3, r3, #0
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: vmov q2[2], q2[0], r3, r0
+; CHECK-NEXT: vmov q2[3], q2[1], r3, r0
+; CHECK-NEXT: vand q2, q4, q2
+; CHECK-NEXT: vmov r3, s8
+; CHECK-NEXT: vmov r0, s9
+; CHECK-NEXT: vmov r4, s11
+; CHECK-NEXT: adds.w r3, r3, r12
+; CHECK-NEXT: adcs r1, r0
+; CHECK-NEXT: vmov r0, s10
+; CHECK-NEXT: adds.w r12, r3, r0
+; CHECK-NEXT: vmov.u16 r3, q3[7]
+; CHECK-NEXT: adc.w lr, r1, r4
+; CHECK-NEXT: vmov.u16 r4, q3[6]
+; CHECK-NEXT: vmov q2[2], q2[0], r4, r3
+; CHECK-NEXT: vmov.u16 r4, q0[7]
+; CHECK-NEXT: vmov.u16 r0, q0[6]
+; CHECK-NEXT: vand q2, q2, q1
+; CHECK-NEXT: vmov q0[2], q0[0], r0, r4
+; CHECK-NEXT: vmov r3, s10
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r4, s8
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: vmov r1, s0
+; CHECK-NEXT: umull r0, r3, r0, r3
+; CHECK-NEXT: umull r1, r4, r1, r4
+; CHECK-NEXT: vmov q0[2], q0[0], r1, r0
+; CHECK-NEXT: ubfx r0, r2, #12, #1
+; CHECK-NEXT: ubfx r1, r2, #8, #1
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q0[3], q0[1], r4, r3
+; CHECK-NEXT: vmov q1[2], q1[0], r1, r0
+; CHECK-NEXT: vmov q1[3], q1[1], r1, r0
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r1, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adds.w r1, r1, r12
+; CHECK-NEXT: adc.w r2, lr, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: vpop {d8, d9, d10, d11}
+; CHECK-NEXT: pop {r4, pc}
+entry:
+ %c = icmp eq <8 x i8> %b, zeroinitializer
+ %xx = zext <8 x i8> %x to <8 x i64>
+ %yy = zext <8 x i8> %y to <8 x i64>
+ %m = mul <8 x i64> %xx, %yy
+ %s = select <8 x i1> %c, <8 x i64> %m, <8 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %s)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_sext(<8 x i8> %x, <8 x i8> %y, <8 x i8> %b) {
+; CHECK-LABEL: add_v8i8_v8i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r4, lr}
+; CHECK-NEXT: push {r4, lr}
+; CHECK-NEXT: .vsave {d8, d9}
+; CHECK-NEXT: vpush {d8, d9}
+; CHECK-NEXT: vmovlb.u8 q2, q2
+; CHECK-NEXT: vmov.i8 q3, #0xff
+; CHECK-NEXT: vcmp.i16 eq, q2, zr
+; CHECK-NEXT: vmov.i8 q2, #0x0
+; CHECK-NEXT: vpsel q2, q3, q2
+; CHECK-NEXT: vmov.u16 r3, q1[0]
+; CHECK-NEXT: vmov.u16 r0, q2[2]
+; CHECK-NEXT: vmov.u16 r1, q2[0]
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov.u16 r0, q2[3]
+; CHECK-NEXT: vmov.u16 r1, q2[1]
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: vmov q3[3], q3[1], r1, r0
+; CHECK-NEXT: vmov.u16 r4, q0[4]
+; CHECK-NEXT: vcmp.i32 ne, q3, zr
+; CHECK-NEXT: sxtb r4, r4
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r1
+; CHECK-NEXT: vmov q3[3], q3[1], r2, r1
+; CHECK-NEXT: vmov.u16 r1, q1[1]
+; CHECK-NEXT: vmov.u16 r2, q0[1]
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: smull r1, r12, r2, r1
+; CHECK-NEXT: vmov.u16 r2, q0[0]
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: smull r2, r3, r2, r3
+; CHECK-NEXT: vmov q4[2], q4[0], r2, r1
+; CHECK-NEXT: vmov q4[3], q4[1], r3, r12
+; CHECK-NEXT: vand q3, q4, q3
+; CHECK-NEXT: vmov r3, s14
+; CHECK-NEXT: vmov r1, s12
+; CHECK-NEXT: vmov r12, s15
+; CHECK-NEXT: vmov r2, s13
+; CHECK-NEXT: adds.w lr, r1, r3
+; CHECK-NEXT: ubfx r3, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsb.w r3, r3, #0
+; CHECK-NEXT: rsb.w r0, r0, #0
+; CHECK-NEXT: vmov.u16 r1, q1[2]
+; CHECK-NEXT: vmov q3[2], q3[0], r0, r3
+; CHECK-NEXT: adc.w r12, r12, r2
+; CHECK-NEXT: vmov q3[3], q3[1], r0, r3
+; CHECK-NEXT: vmov.u16 r0, q1[3]
+; CHECK-NEXT: vmov.u16 r3, q0[3]
+; CHECK-NEXT: vmov.u16 r2, q0[2]
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: smull r0, r3, r3, r0
+; CHECK-NEXT: smull r1, r2, r2, r1
+; CHECK-NEXT: vmov q4[2], q4[0], r1, r0
+; CHECK-NEXT: vmov q4[3], q4[1], r2, r3
+; CHECK-NEXT: vand q3, q4, q3
+; CHECK-NEXT: vmov r1, s12
+; CHECK-NEXT: vmov r0, s13
+; CHECK-NEXT: vmov r3, s15
+; CHECK-NEXT: adds.w r1, r1, lr
+; CHECK-NEXT: adc.w r2, r12, r0
+; CHECK-NEXT: vmov r0, s14
+; CHECK-NEXT: adds.w r12, r1, r0
+; CHECK-NEXT: vmov.u16 r1, q1[4]
+; CHECK-NEXT: adc.w lr, r2, r3
+; CHECK-NEXT: vmov.u16 r2, q2[6]
+; CHECK-NEXT: vmov.u16 r3, q2[4]
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: vmov q3[2], q3[0], r3, r2
+; CHECK-NEXT: vmov.u16 r2, q2[7]
+; CHECK-NEXT: vmov.u16 r3, q2[5]
+; CHECK-NEXT: smull r1, r4, r4, r1
+; CHECK-NEXT: vmov q3[3], q3[1], r3, r2
+; CHECK-NEXT: vcmp.i32 ne, q3, zr
+; CHECK-NEXT: vmrs r2, p0
+; CHECK-NEXT: and r0, r2, #1
+; CHECK-NEXT: ubfx r3, r2, #4, #1
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: rsbs r3, r3, #0
+; CHECK-NEXT: vmov q2[2], q2[0], r0, r3
+; CHECK-NEXT: vmov q2[3], q2[1], r0, r3
+; CHECK-NEXT: vmov.u16 r0, q1[5]
+; CHECK-NEXT: vmov.u16 r3, q0[5]
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: smull r0, r3, r3, r0
+; CHECK-NEXT: vmov q3[2], q3[0], r1, r0
+; CHECK-NEXT: vmov q3[3], q3[1], r4, r3
+; CHECK-NEXT: vand q2, q3, q2
+; CHECK-NEXT: vmov r1, s8
+; CHECK-NEXT: vmov r0, s9
+; CHECK-NEXT: vmov r4, s10
+; CHECK-NEXT: vmov r3, s11
+; CHECK-NEXT: adds.w r1, r1, r12
+; CHECK-NEXT: adc.w r0, r0, lr
+; CHECK-NEXT: adds r1, r1, r4
+; CHECK-NEXT: vmov.u16 r4, q1[6]
+; CHECK-NEXT: adc.w r12, r0, r3
+; CHECK-NEXT: ubfx r3, r2, #12, #1
+; CHECK-NEXT: ubfx r2, r2, #8, #1
+; CHECK-NEXT: rsbs r3, r3, #0
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: vmov.u16 r0, q0[6]
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r3
+; CHECK-NEXT: sxtb r4, r4
+; CHECK-NEXT: vmov q2[3], q2[1], r2, r3
+; CHECK-NEXT: vmov.u16 r2, q1[7]
+; CHECK-NEXT: vmov.u16 r3, q0[7]
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: smull r2, r3, r3, r2
+; CHECK-NEXT: smull r0, r4, r0, r4
+; CHECK-NEXT: vmov q0[2], q0[0], r0, r2
+; CHECK-NEXT: vmov q0[3], q0[1], r4, r3
+; CHECK-NEXT: vand q0, q0, q2
+; CHECK-NEXT: vmov r2, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adds r1, r1, r2
+; CHECK-NEXT: adc.w r2, r12, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: vpop {d8, d9}
+; CHECK-NEXT: pop {r4, pc}
+entry:
+ %c = icmp eq <8 x i8> %b, zeroinitializer
+ %xx = sext <8 x i8> %x to <8 x i64>
+ %yy = sext <8 x i8> %y to <8 x i64>
+ %m = mul <8 x i64> %xx, %yy
+ %s = select <8 x i1> %c, <8 x i64> %m, <8 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %s)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_zext(<4 x i8> %x, <4 x i8> %y, <4 x i8> %b) {
+; CHECK-LABEL: add_v4i8_v4i64_zext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13}
+; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13}
+; CHECK-NEXT: vmov.i32 q3, #0xff
+; CHECK-NEXT: vand q2, q2, q3
+; CHECK-NEXT: vand q1, q1, q3
+; CHECK-NEXT: vcmp.i32 eq, q2, zr
+; CHECK-NEXT: vand q0, q0, q3
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: vmov.f32 s16, s4
+; CHECK-NEXT: vmov.f32 s20, s0
+; CHECK-NEXT: vmov.f32 s18, s5
+; CHECK-NEXT: vmov.f32 s22, s1
+; CHECK-NEXT: vmullb.u32 q6, q5, q4
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r1
+; CHECK-NEXT: vmov q2[3], q2[1], r2, r1
+; CHECK-NEXT: vand q2, q6, q2
+; CHECK-NEXT: vmov r3, s10
+; CHECK-NEXT: vmov r1, s8
+; CHECK-NEXT: vmov r12, s11
+; CHECK-NEXT: vmov r2, s9
+; CHECK-NEXT: vmov.f32 s8, s6
+; CHECK-NEXT: vmov.f32 s10, s7
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: vmullb.u32 q0, q1, q2
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: ubfx r3, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsb.w r3, r3, #0
+; CHECK-NEXT: rsb.w r0, r0, #0
+; CHECK-NEXT: adc.w r2, r2, r12
+; CHECK-NEXT: vmov q1[2], q1[0], r0, r3
+; CHECK-NEXT: vmov q1[3], q1[1], r0, r3
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: adds r1, r1, r3
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adcs r2, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13}
+; CHECK-NEXT: bx lr
+entry:
+ %c = icmp eq <4 x i8> %b, zeroinitializer
+ %xx = zext <4 x i8> %x to <4 x i64>
+ %yy = zext <4 x i8> %y to <4 x i64>
+ %m = mul <4 x i64> %xx, %yy
+ %s = select <4 x i1> %c, <4 x i64> %m, <4 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %s)
+ ret i64 %z
+}
+
+define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_sext(<4 x i8> %x, <4 x i8> %y, <4 x i8> %b) {
+; CHECK-LABEL: add_v4i8_v4i64_sext:
+; CHECK: @ %bb.0: @ %entry
+; CHECK-NEXT: .save {r4, lr}
+; CHECK-NEXT: push {r4, lr}
+; CHECK-NEXT: .vsave {d8, d9}
+; CHECK-NEXT: vpush {d8, d9}
+; CHECK-NEXT: vmov.f32 s12, s4
+; CHECK-NEXT: vmov.i32 q4, #0xff
+; CHECK-NEXT: vmov.f32 s14, s5
+; CHECK-NEXT: vand q2, q2, q4
+; CHECK-NEXT: vmov r2, s4
+; CHECK-NEXT: vcmp.i32 eq, q2, zr
+; CHECK-NEXT: vmov r3, s0
+; CHECK-NEXT: vmov r0, s14
+; CHECK-NEXT: vmov.f32 s12, s0
+; CHECK-NEXT: vmov.f32 s14, s1
+; CHECK-NEXT: vmov r1, s14
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: smull r2, r3, r3, r2
+; CHECK-NEXT: sxtb r0, r0
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: smull r0, r1, r1, r0
+; CHECK-NEXT: vmov q3[2], q3[0], r2, r0
+; CHECK-NEXT: vmrs r0, p0
+; CHECK-NEXT: vmov q3[3], q3[1], r3, r1
+; CHECK-NEXT: and r2, r0, #1
+; CHECK-NEXT: ubfx r1, r0, #4, #1
+; CHECK-NEXT: rsbs r2, r2, #0
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: vmov q2[2], q2[0], r2, r1
+; CHECK-NEXT: vmov q2[3], q2[1], r2, r1
+; CHECK-NEXT: vand q2, q3, q2
+; CHECK-NEXT: vmov r1, s10
+; CHECK-NEXT: vmov r2, s8
+; CHECK-NEXT: vmov r12, s11
+; CHECK-NEXT: vmov r3, s9
+; CHECK-NEXT: vmov.f32 s8, s6
+; CHECK-NEXT: vmov.f32 s10, s7
+; CHECK-NEXT: vmov.f32 s4, s2
+; CHECK-NEXT: vmov.f32 s6, s3
+; CHECK-NEXT: vmov r4, s4
+; CHECK-NEXT: adds.w lr, r2, r1
+; CHECK-NEXT: vmov r1, s6
+; CHECK-NEXT: vmov r2, s8
+; CHECK-NEXT: adc.w r12, r12, r3
+; CHECK-NEXT: vmov r3, s10
+; CHECK-NEXT: sxtb r4, r4
+; CHECK-NEXT: sxtb r1, r1
+; CHECK-NEXT: sxtb r2, r2
+; CHECK-NEXT: sxtb r3, r3
+; CHECK-NEXT: smull r2, r4, r4, r2
+; CHECK-NEXT: smull r1, r3, r1, r3
+; CHECK-NEXT: vmov q0[2], q0[0], r2, r1
+; CHECK-NEXT: ubfx r1, r0, #12, #1
+; CHECK-NEXT: ubfx r0, r0, #8, #1
+; CHECK-NEXT: rsbs r1, r1, #0
+; CHECK-NEXT: rsbs r0, r0, #0
+; CHECK-NEXT: vmov q0[3], q0[1], r4, r3
+; CHECK-NEXT: vmov q1[2], q1[0], r0, r1
+; CHECK-NEXT: vmov q1[3], q1[1], r0, r1
+; CHECK-NEXT: vand q0, q0, q1
+; CHECK-NEXT: vmov r1, s0
+; CHECK-NEXT: vmov r0, s1
+; CHECK-NEXT: vmov r3, s3
+; CHECK-NEXT: adds.w r1, r1, lr
+; CHECK-NEXT: adc.w r2, r12, r0
+; CHECK-NEXT: vmov r0, s2
+; CHECK-NEXT: adds r0, r0, r1
+; CHECK-NEXT: adc.w r1, r2, r3
+; CHECK-NEXT: vpop {d8, d9}
+; CHECK-NEXT: pop {r4, pc}
+entry:
+ %c = icmp eq <4 x i8> %b, zeroinitializer
+ %xx = sext <4 x i8> %x to <4 x i64>
+ %yy = sext <4 x i8> %y to <4 x i64>
+ %m = mul <4 x i64> %xx, %yy
+ %s = select <4 x i1> %c, <4 x i64> %m, <4 x i64> zeroinitializer
+ %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %s)
+ ret i64 %z
+}
+
define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_zext(<2 x i8> %x, <2 x i8> %y, <2 x i8> %b) {
; CHECK-LABEL: add_v2i8_v2i64_zext:
; CHECK: @ %bb.0: @ %entry
More information about the llvm-branch-commits
mailing list