1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 11:13:28 +01:00
llvm-mirror/test/CodeGen/Thumb2/mve-intrinsics/vmldav.ll
Mikhail Maltsev c04187c6e9 [ARM][MVE] Add vector reduction intrinsics with two vector operands
Summary:
This patch adds intrinsics for the following MVE instructions:
* VABAV
* VMLADAV, VMLSDAV
* VMLALDAV, VMLSLDAV
* VRMLALDAVH, VRMLSLDAVH

Each of the above 4 groups has a corresponding new LLVM IR intrinsic,
since the instructions cannot be easily represented using
general-purpose IR operations.

Reviewers: simon_tatham, ostannard, dmgreen, MarkMurrayARM

Reviewed By: MarkMurrayARM

Subscribers: merge_guards_bot, kristof.beyls, hiraditya, cfe-commits, llvm-commits

Tags: #clang, #llvm

Differential Revision: https://reviews.llvm.org/D71062
2019-12-13 13:17:29 +00:00

735 lines
26 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=thumbv8.1m.main -mattr=+mve.fp -verify-machineinstrs -o - %s | FileCheck %s
declare <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32)
declare <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32)
declare <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32)
declare i32 @llvm.arm.mve.vmldava.v16i8(i32, i32, i32, i32, <16 x i8>, <16 x i8>)
declare i32 @llvm.arm.mve.vmldava.v8i16(i32, i32, i32, i32, <8 x i16>, <8 x i16>)
declare i32 @llvm.arm.mve.vmldava.v4i32(i32, i32, i32, i32, <4 x i32>, <4 x i32>)
declare i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32, i32, i32, i32, <16 x i8>, <16 x i8>, <16 x i1>)
declare i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32, i32, i32, i32, <8 x i16>, <8 x i16>, <8 x i1>)
declare i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32, i32, i32, i32, <4 x i32>, <4 x i32>, <4 x i1>)
define arm_aapcs_vfpcc i32 @test_vmladavaq_s8(i32 %a, <16 x i8> %b, <16 x i8> %c) {
; CHECK-LABEL: test_vmladavaq_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlava.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 0, i32 0, i32 0, i32 %a, <16 x i8> %b, <16 x i8> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_s16(i32 %a, <8 x i16> %b, <8 x i16> %c) {
; CHECK-LABEL: test_vmladavaq_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlava.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 0, i32 0, i32 0, i32 %a, <8 x i16> %b, <8 x i16> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_s32(i32 %a, <4 x i32> %b, <4 x i32> %c) {
; CHECK-LABEL: test_vmladavaq_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlava.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 0, i32 0, i32 0, i32 %a, <4 x i32> %b, <4 x i32> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_u8(i32 %a, <16 x i8> %b, <16 x i8> %c) {
; CHECK-LABEL: test_vmladavaq_u8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlava.u8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 1, i32 0, i32 0, i32 %a, <16 x i8> %b, <16 x i8> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_u16(i32 %a, <8 x i16> %b, <8 x i16> %c) {
; CHECK-LABEL: test_vmladavaq_u16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlava.u16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 1, i32 0, i32 0, i32 %a, <8 x i16> %b, <8 x i16> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_u32(i32 %a, <4 x i32> %b, <4 x i32> %c) {
; CHECK-LABEL: test_vmladavaq_u32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlava.u32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 1, i32 0, i32 0, i32 %a, <4 x i32> %b, <4 x i32> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavaxq_s8(i32 %a, <16 x i8> %b, <16 x i8> %c) {
; CHECK-LABEL: test_vmladavaxq_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmladavax.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 0, i32 0, i32 1, i32 %a, <16 x i8> %b, <16 x i8> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavaxq_s16(i32 %a, <8 x i16> %b, <8 x i16> %c) {
; CHECK-LABEL: test_vmladavaxq_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmladavax.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 0, i32 0, i32 1, i32 %a, <8 x i16> %b, <8 x i16> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavaxq_s32(i32 %a, <4 x i32> %b, <4 x i32> %c) {
; CHECK-LABEL: test_vmladavaxq_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmladavax.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 0, i32 0, i32 1, i32 %a, <4 x i32> %b, <4 x i32> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaq_s8(i32 %a, <16 x i8> %b, <16 x i8> %c) {
; CHECK-LABEL: test_vmlsdavaq_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdava.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 0, i32 1, i32 0, i32 %a, <16 x i8> %b, <16 x i8> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaq_s16(i32 %a, <8 x i16> %b, <8 x i16> %c) {
; CHECK-LABEL: test_vmlsdavaq_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdava.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 0, i32 1, i32 0, i32 %a, <8 x i16> %b, <8 x i16> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaq_s32(i32 %a, <4 x i32> %b, <4 x i32> %c) {
; CHECK-LABEL: test_vmlsdavaq_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdava.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 0, i32 1, i32 0, i32 %a, <4 x i32> %b, <4 x i32> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaxq_s8(i32 %a, <16 x i8> %b, <16 x i8> %c) {
; CHECK-LABEL: test_vmlsdavaxq_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdavax.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 0, i32 1, i32 1, i32 %a, <16 x i8> %b, <16 x i8> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaxq_s16(i32 %a, <8 x i16> %b, <8 x i16> %c) {
; CHECK-LABEL: test_vmlsdavaxq_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdavax.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 0, i32 1, i32 1, i32 %a, <8 x i16> %b, <8 x i16> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaxq_s32(i32 %a, <4 x i32> %b, <4 x i32> %c) {
; CHECK-LABEL: test_vmlsdavaxq_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdavax.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 0, i32 1, i32 1, i32 %a, <4 x i32> %b, <4 x i32> %c)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_p_s8(i32 %a, <16 x i8> %b, <16 x i8> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavaq_p_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavat.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 0, i32 0, i32 0, i32 %a, <16 x i8> %b, <16 x i8> %c, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_p_s16(i32 %a, <8 x i16> %b, <8 x i16> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavaq_p_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavat.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 0, i32 0, i32 0, i32 %a, <8 x i16> %b, <8 x i16> %c, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_p_s32(i32 %a, <4 x i32> %b, <4 x i32> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavaq_p_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavat.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 0, i32 0, i32 0, i32 %a, <4 x i32> %b, <4 x i32> %c, <4 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_p_u8(i32 %a, <16 x i8> %b, <16 x i8> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavaq_p_u8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavat.u8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 1, i32 0, i32 0, i32 %a, <16 x i8> %b, <16 x i8> %c, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_p_u16(i32 %a, <8 x i16> %b, <8 x i16> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavaq_p_u16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavat.u16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 1, i32 0, i32 0, i32 %a, <8 x i16> %b, <8 x i16> %c, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavaq_p_u32(i32 %a, <4 x i32> %b, <4 x i32> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavaq_p_u32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavat.u32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 1, i32 0, i32 0, i32 %a, <4 x i32> %b, <4 x i32> %c, <4 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavaxq_p_s8(i32 %a, <16 x i8> %b, <16 x i8> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavaxq_p_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmladavaxt.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 0, i32 0, i32 1, i32 %a, <16 x i8> %b, <16 x i8> %c, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavaxq_p_s16(i32 %a, <8 x i16> %b, <8 x i16> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavaxq_p_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmladavaxt.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 0, i32 0, i32 1, i32 %a, <8 x i16> %b, <8 x i16> %c, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavaxq_p_s32(i32 %a, <4 x i32> %b, <4 x i32> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavaxq_p_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmladavaxt.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 0, i32 0, i32 1, i32 %a, <4 x i32> %b, <4 x i32> %c, <4 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaq_p_s8(i32 %a, <16 x i8> %b, <16 x i8> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavaq_p_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavat.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 0, i32 1, i32 0, i32 %a, <16 x i8> %b, <16 x i8> %c, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaq_p_s16(i32 %a, <8 x i16> %b, <8 x i16> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavaq_p_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavat.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 0, i32 1, i32 0, i32 %a, <8 x i16> %b, <8 x i16> %c, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaq_p_s32(i32 %a, <4 x i32> %b, <4 x i32> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavaq_p_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavat.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 0, i32 1, i32 0, i32 %a, <4 x i32> %b, <4 x i32> %c, <4 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaxq_p_s8(i32 %a, <16 x i8> %b, <16 x i8> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavaxq_p_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavaxt.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 0, i32 1, i32 1, i32 %a, <16 x i8> %b, <16 x i8> %c, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaxq_p_s16(i32 %a, <8 x i16> %b, <8 x i16> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavaxq_p_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavaxt.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 0, i32 1, i32 1, i32 %a, <8 x i16> %b, <8 x i16> %c, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavaxq_p_s32(i32 %a, <4 x i32> %b, <4 x i32> %c, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavaxq_p_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r1
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavaxt.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 0, i32 1, i32 1, i32 %a, <4 x i32> %b, <4 x i32> %c, <4 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavq_s8(<16 x i8> %a, <16 x i8> %b) {
; CHECK-LABEL: test_vmladavq_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlav.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 0, i32 0, i32 0, i32 0, <16 x i8> %a, <16 x i8> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavq_s16(<8 x i16> %a, <8 x i16> %b) {
; CHECK-LABEL: test_vmladavq_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlav.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 0, i32 0, i32 0, i32 0, <8 x i16> %a, <8 x i16> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavq_s32(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_vmladavq_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlav.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 0, i32 0, i32 0, i32 0, <4 x i32> %a, <4 x i32> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavq_u8(<16 x i8> %a, <16 x i8> %b) {
; CHECK-LABEL: test_vmladavq_u8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlav.u8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 1, i32 0, i32 0, i32 0, <16 x i8> %a, <16 x i8> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavq_u16(<8 x i16> %a, <8 x i16> %b) {
; CHECK-LABEL: test_vmladavq_u16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlav.u16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 1, i32 0, i32 0, i32 0, <8 x i16> %a, <8 x i16> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavq_u32(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_vmladavq_u32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlav.u32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 1, i32 0, i32 0, i32 0, <4 x i32> %a, <4 x i32> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavxq_s8(<16 x i8> %a, <16 x i8> %b) {
; CHECK-LABEL: test_vmladavxq_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmladavx.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 0, i32 0, i32 1, i32 0, <16 x i8> %a, <16 x i8> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavxq_s16(<8 x i16> %a, <8 x i16> %b) {
; CHECK-LABEL: test_vmladavxq_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmladavx.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 0, i32 0, i32 1, i32 0, <8 x i16> %a, <8 x i16> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavxq_s32(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_vmladavxq_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmladavx.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 0, i32 0, i32 1, i32 0, <4 x i32> %a, <4 x i32> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavq_s8(<16 x i8> %a, <16 x i8> %b) {
; CHECK-LABEL: test_vmlsdavq_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdav.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 0, i32 1, i32 0, i32 0, <16 x i8> %a, <16 x i8> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavq_s16(<8 x i16> %a, <8 x i16> %b) {
; CHECK-LABEL: test_vmlsdavq_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdav.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 0, i32 1, i32 0, i32 0, <8 x i16> %a, <8 x i16> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavq_s32(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_vmlsdavq_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdav.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 0, i32 1, i32 0, i32 0, <4 x i32> %a, <4 x i32> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavxq_s8(<16 x i8> %a, <16 x i8> %b) {
; CHECK-LABEL: test_vmlsdavxq_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdavx.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v16i8(i32 0, i32 1, i32 1, i32 0, <16 x i8> %a, <16 x i8> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavxq_s16(<8 x i16> %a, <8 x i16> %b) {
; CHECK-LABEL: test_vmlsdavxq_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdavx.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v8i16(i32 0, i32 1, i32 1, i32 0, <8 x i16> %a, <8 x i16> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmlsdavxq_s32(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_vmlsdavxq_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmlsdavx.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = call i32 @llvm.arm.mve.vmldava.v4i32(i32 0, i32 1, i32 1, i32 0, <4 x i32> %a, <4 x i32> %b)
ret i32 %0
}
define arm_aapcs_vfpcc i32 @test_vmladavq_p_s8(<16 x i8> %a, <16 x i8> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavq_p_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavt.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 0, i32 0, i32 0, i32 0, <16 x i8> %a, <16 x i8> %b, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavq_p_s16(<8 x i16> %a, <8 x i16> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavq_p_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavt.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 0, i32 0, i32 0, i32 0, <8 x i16> %a, <8 x i16> %b, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavq_p_s32(<4 x i32> %a, <4 x i32> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavq_p_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavt.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 0, i32 0, i32 0, i32 0, <4 x i32> %a, <4 x i32> %b, <4 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavq_p_u8(<16 x i8> %a, <16 x i8> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavq_p_u8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavt.u8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 1, i32 0, i32 0, i32 0, <16 x i8> %a, <16 x i8> %b, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavq_p_u16(<8 x i16> %a, <8 x i16> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavq_p_u16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavt.u16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 1, i32 0, i32 0, i32 0, <8 x i16> %a, <8 x i16> %b, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavq_p_u32(<4 x i32> %a, <4 x i32> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavq_p_u32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlavt.u32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 1, i32 0, i32 0, i32 0, <4 x i32> %a, <4 x i32> %b, <4 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavxq_p_s8(<16 x i8> %a, <16 x i8> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavxq_p_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmladavxt.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 0, i32 0, i32 1, i32 0, <16 x i8> %a, <16 x i8> %b, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavxq_p_s16(<8 x i16> %a, <8 x i16> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavxq_p_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmladavxt.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 0, i32 0, i32 1, i32 0, <8 x i16> %a, <8 x i16> %b, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmladavxq_p_s32(<4 x i32> %a, <4 x i32> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmladavxq_p_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmladavxt.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 0, i32 0, i32 1, i32 0, <4 x i32> %a, <4 x i32> %b, <4 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavq_p_s8(<16 x i8> %a, <16 x i8> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavq_p_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavt.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 0, i32 1, i32 0, i32 0, <16 x i8> %a, <16 x i8> %b, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavq_p_s16(<8 x i16> %a, <8 x i16> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavq_p_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavt.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 0, i32 1, i32 0, i32 0, <8 x i16> %a, <8 x i16> %b, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavq_p_s32(<4 x i32> %a, <4 x i32> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavq_p_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavt.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 0, i32 1, i32 0, i32 0, <4 x i32> %a, <4 x i32> %b, <4 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavxq_p_s8(<16 x i8> %a, <16 x i8> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavxq_p_s8:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavxt.s8 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v16i8.v16i1(i32 0, i32 1, i32 1, i32 0, <16 x i8> %a, <16 x i8> %b, <16 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavxq_p_s16(<8 x i16> %a, <8 x i16> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavxq_p_s16:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavxt.s16 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v8i16.v8i1(i32 0, i32 1, i32 1, i32 0, <8 x i16> %a, <8 x i16> %b, <8 x i1> %1)
ret i32 %2
}
define arm_aapcs_vfpcc i32 @test_vmlsdavxq_p_s32(<4 x i32> %a, <4 x i32> %b, i16 zeroext %p) {
; CHECK-LABEL: test_vmlsdavxq_p_s32:
; CHECK: @ %bb.0: @ %entry
; CHECK-NEXT: vmsr p0, r0
; CHECK-NEXT: vpst
; CHECK-NEXT: vmlsdavxt.s32 r0, q0, q1
; CHECK-NEXT: bx lr
entry:
%0 = zext i16 %p to i32
%1 = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 %0)
%2 = call i32 @llvm.arm.mve.vmldava.predicated.v4i32.v4i1(i32 0, i32 1, i32 1, i32 0, <4 x i32> %a, <4 x i32> %b, <4 x i1> %1)
ret i32 %2
}