From e8e7b2cb46bb16a8939567d770c3f69df35e7bdc Mon Sep 17 00:00:00 2001 From: David Green Date: Thu, 28 May 2020 17:49:01 +0100 Subject: [PATCH] [ARM] More tests for MVE LSR and float issues. NFC --- .../CodeGen/Thumb2/mve-float32regloops.ll | 630 ++++++++++++++++++ 1 file changed, 630 insertions(+) diff --git a/llvm/test/CodeGen/Thumb2/mve-float32regloops.ll b/llvm/test/CodeGen/Thumb2/mve-float32regloops.ll index 111a5871a17b8..45cb9fc5b4bcd 100644 --- a/llvm/test/CodeGen/Thumb2/mve-float32regloops.ll +++ b/llvm/test/CodeGen/Thumb2/mve-float32regloops.ll @@ -1395,6 +1395,636 @@ if.end: ; preds = %while.end, %if.then ret void } +%struct.arm_biquad_cascade_stereo_df2T_instance_f32 = type { i8, float*, float* } +define arm_aapcs_vfpcc void @arm_biquad_cascade_stereo_df2T_f32(%struct.arm_biquad_cascade_stereo_df2T_instance_f32* nocapture readonly %0, float* %1, float* %2, i32 %3) { +; CHECK-LABEL: arm_biquad_cascade_stereo_df2T_f32: +; CHECK: @ %bb.0: +; CHECK-NEXT: .save {r4, r5, r6, r7, lr} +; CHECK-NEXT: push {r4, r5, r6, r7, lr} +; CHECK-NEXT: .pad #4 +; CHECK-NEXT: sub sp, #4 +; CHECK-NEXT: .vsave {d8, d9, d10, d11} +; CHECK-NEXT: vpush {d8, d9, d10, d11} +; CHECK-NEXT: .pad #24 +; CHECK-NEXT: sub sp, #24 +; CHECK-NEXT: ldrb.w lr, [r0] +; CHECK-NEXT: movs r4, #0 +; CHECK-NEXT: ldrd r12, r0, [r0, #4] +; CHECK-NEXT: cmp r3, #0 +; CHECK-NEXT: strd r4, r4, [sp, #16] +; CHECK-NEXT: beq .LBB17_5 +; CHECK-NEXT: @ %bb.1: +; CHECK-NEXT: movs r5, #2 +; CHECK-NEXT: viwdup.u32 q0, r4, r5, #1 +; CHECK-NEXT: mov r4, sp +; CHECK-NEXT: .LBB17_2: @ =>This Loop Header: Depth=1 +; CHECK-NEXT: @ Child Loop BB17_3 Depth 2 +; CHECK-NEXT: mov r7, lr +; CHECK-NEXT: ldr.w lr, [r0, #12] +; CHECK-NEXT: ldrd r5, r6, [r0] +; CHECK-NEXT: vldrw.u32 q1, [r12] +; CHECK-NEXT: vldr s12, [r0, #8] +; CHECK-NEXT: vdup.32 q2, lr +; CHECK-NEXT: vldr s14, [r0, #16] +; CHECK-NEXT: vstrw.32 q1, [r4] +; CHECK-NEXT: vdup.32 q1, r6 +; CHECK-NEXT: mov r6, r2 +; CHECK-NEXT: vmov.f32 s6, s12 +; CHECK-NEXT: dls lr, r3 +; CHECK-NEXT: vmov.f32 s10, s14 +; CHECK-NEXT: vmov.f32 s7, s12 +; CHECK-NEXT: vmov.f32 s11, s14 +; CHECK-NEXT: .LBB17_3: @ Parent Loop BB17_2 Depth=1 +; CHECK-NEXT: @ => This Inner Loop Header: Depth=2 +; CHECK-NEXT: vldrw.u32 q4, [r1, q0, uxtw #2] +; CHECK-NEXT: vldrw.u32 q5, [r4, q0, uxtw #2] +; CHECK-NEXT: adds r1, #8 +; CHECK-NEXT: vfma.f32 q5, q4, r5 +; CHECK-NEXT: vstmia r6, {s20, s21} +; CHECK-NEXT: adds r6, #8 +; CHECK-NEXT: vldrw.u32 q3, [sp, #8] +; CHECK-NEXT: vfma.f32 q3, q5, q2 +; CHECK-NEXT: vfma.f32 q3, q4, q1 +; CHECK-NEXT: vstrw.32 q3, [r4] +; CHECK-NEXT: le lr, .LBB17_3 +; CHECK-NEXT: @ %bb.4: @ in Loop: Header=BB17_2 Depth=1 +; CHECK-NEXT: mov lr, r7 +; CHECK-NEXT: adds r0, #20 +; CHECK-NEXT: subs.w lr, r7, #1 +; CHECK-NEXT: vstrb.8 q3, [r12], #16 +; CHECK-NEXT: mov r1, r2 +; CHECK-NEXT: bne .LBB17_2 +; CHECK-NEXT: b .LBB17_7 +; CHECK-NEXT: .LBB17_5: @ %.preheader +; CHECK-NEXT: mov r0, sp +; CHECK-NEXT: dls lr, lr +; CHECK-NEXT: .LBB17_6: @ =>This Inner Loop Header: Depth=1 +; CHECK-NEXT: vldrw.u32 q0, [r12], #16 +; CHECK-NEXT: vstrw.32 q0, [r0] +; CHECK-NEXT: le lr, .LBB17_6 +; CHECK-NEXT: .LBB17_7: +; CHECK-NEXT: add sp, #24 +; CHECK-NEXT: vpop {d8, d9, d10, d11} +; CHECK-NEXT: add sp, #4 +; CHECK-NEXT: pop {r4, r5, r6, r7, pc} + %5 = alloca [6 x float], align 4 + %6 = getelementptr inbounds %struct.arm_biquad_cascade_stereo_df2T_instance_f32, %struct.arm_biquad_cascade_stereo_df2T_instance_f32* %0, i32 0, i32 1 + %7 = load float*, float** %6, align 4 + %8 = getelementptr inbounds %struct.arm_biquad_cascade_stereo_df2T_instance_f32, %struct.arm_biquad_cascade_stereo_df2T_instance_f32* %0, i32 0, i32 2 + %9 = load float*, float** %8, align 4 + %10 = getelementptr inbounds %struct.arm_biquad_cascade_stereo_df2T_instance_f32, %struct.arm_biquad_cascade_stereo_df2T_instance_f32* %0, i32 0, i32 0 + %11 = load i8, i8* %10, align 4 + %12 = zext i8 %11 to i32 + %13 = bitcast [6 x float]* %5 to i8* + call void @llvm.lifetime.start.p0i8(i64 24, i8* nonnull %13) #5 + %14 = tail call { <4 x i32>, i32 } @llvm.arm.mve.viwdup.v4i32(i32 0, i32 2, i32 1) + %15 = extractvalue { <4 x i32>, i32 } %14, 0 + %16 = getelementptr inbounds [6 x float], [6 x float]* %5, i32 0, i32 4 + store float 0.000000e+00, float* %16, align 4 + %17 = getelementptr inbounds [6 x float], [6 x float]* %5, i32 0, i32 5 + store float 0.000000e+00, float* %17, align 4 + %18 = bitcast [6 x float]* %5 to <4 x float>* + %19 = icmp eq i32 %3, 0 + %20 = bitcast [6 x float]* %5 to i32* + %21 = getelementptr inbounds [6 x float], [6 x float]* %5, i32 0, i32 2 + %22 = bitcast float* %21 to <4 x float>* + br i1 %19, label %23, label %31 + +23: ; preds = %4, %23 + %24 = phi i32 [ %29, %23 ], [ %12, %4 ] + %25 = phi float* [ %28, %23 ], [ %7, %4 ] + %26 = bitcast float* %25 to <4 x float>* + %27 = load <4 x float>, <4 x float>* %26, align 8 + store <4 x float> %27, <4 x float>* %18, align 4 + %28 = getelementptr inbounds float, float* %25, i32 4 + %29 = add i32 %24, -1 + %30 = icmp eq i32 %29, 0 + br i1 %30, label %82, label %23 + +31: ; preds = %4, %77 + %32 = phi i32 [ %80, %77 ], [ %12, %4 ] + %33 = phi float* [ %78, %77 ], [ %9, %4 ] + %34 = phi float* [ %79, %77 ], [ %7, %4 ] + %35 = phi float* [ %2, %77 ], [ %1, %4 ] + %36 = getelementptr inbounds float, float* %33, i32 1 + %37 = load float, float* %33, align 4 + %38 = getelementptr inbounds float, float* %33, i32 2 + %39 = load float, float* %36, align 4 + %40 = getelementptr inbounds float, float* %33, i32 3 + %41 = load float, float* %38, align 4 + %42 = getelementptr inbounds float, float* %33, i32 4 + %43 = load float, float* %40, align 4 + %44 = load float, float* %42, align 4 + %45 = insertelement <4 x float> undef, float %43, i32 0 + %46 = shufflevector <4 x float> %45, <4 x float> undef, <4 x i32> + %47 = insertelement <4 x float> %46, float %44, i32 2 + %48 = insertelement <4 x float> %47, float %44, i32 3 + %49 = insertelement <4 x float> undef, float %39, i32 0 + %50 = shufflevector <4 x float> %49, <4 x float> undef, <4 x i32> + %51 = insertelement <4 x float> %50, float %41, i32 2 + %52 = insertelement <4 x float> %51, float %41, i32 3 + %53 = bitcast float* %34 to <4 x float>* + %54 = load <4 x float>, <4 x float>* %53, align 8 + store <4 x float> %54, <4 x float>* %18, align 4 + %55 = insertelement <4 x float> undef, float %37, i32 0 + %56 = shufflevector <4 x float> %55, <4 x float> undef, <4 x i32> zeroinitializer + br label %57 + +57: ; preds = %31, %57 + %58 = phi float* [ %35, %31 ], [ %74, %57 ] + %59 = phi float* [ %2, %31 ], [ %70, %57 ] + %60 = phi i32 [ %3, %31 ], [ %75, %57 ] + %61 = call <4 x i32> @llvm.arm.mve.vldr.gather.offset.v4i32.p0i32.v4i32(i32* nonnull %20, <4 x i32> %15, i32 32, i32 2, i32 1) + %62 = bitcast <4 x i32> %61 to <4 x float> + %63 = bitcast float* %58 to i32* + %64 = call <4 x i32> @llvm.arm.mve.vldr.gather.offset.v4i32.p0i32.v4i32(i32* %63, <4 x i32> %15, i32 32, i32 2, i32 1) + %65 = bitcast <4 x i32> %64 to <4 x float> + %66 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %65, <4 x float> %56, <4 x float> %62) + %67 = extractelement <4 x float> %66, i32 0 + %68 = getelementptr inbounds float, float* %59, i32 1 + store float %67, float* %59, align 4 + %69 = extractelement <4 x float> %66, i32 1 + %70 = getelementptr inbounds float, float* %59, i32 2 + store float %69, float* %68, align 4 + %71 = load <4 x float>, <4 x float>* %22, align 4 + %72 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %66, <4 x float> %48, <4 x float> %71) + %73 = call fast <4 x float> @llvm.fma.v4f32(<4 x float> %65, <4 x float> %52, <4 x float> %72) + store <4 x float> %73, <4 x float>* %18, align 4 + %74 = getelementptr inbounds float, float* %58, i32 2 + %75 = add i32 %60, -1 + %76 = icmp eq i32 %75, 0 + br i1 %76, label %77, label %57 + +77: ; preds = %57 + %78 = getelementptr inbounds float, float* %33, i32 5 + store <4 x float> %73, <4 x float>* %53, align 4 + %79 = getelementptr inbounds float, float* %34, i32 4 + %80 = add i32 %32, -1 + %81 = icmp eq i32 %80, 0 + br i1 %81, label %82, label %31 + +82: ; preds = %77, %23 + call void @llvm.lifetime.end.p0i8(i64 24, i8* nonnull %13) #5 + ret void +} + +define arm_aapcs_vfpcc void @fms(float* nocapture readonly %pSrc1, float* nocapture readonly %pSrc2, float* nocapture readonly %pSrc3, float* nocapture %pDst, i32 %N, i32 %M) { +; CHECK-LABEL: fms: +; CHECK: @ %bb.0: @ %entry +; CHECK-NEXT: .save {r4, r5, r7, lr} +; CHECK-NEXT: push {r4, r5, r7, lr} +; CHECK-NEXT: ldr.w lr, [sp, #16] +; CHECK-NEXT: movs r4, #0 +; CHECK-NEXT: cmp.w r4, lr, lsr #2 +; CHECK-NEXT: beq .LBB18_5 +; CHECK-NEXT: @ %bb.1: @ %do.body.preheader +; CHECK-NEXT: ldr.w r12, [sp, #20] +; CHECK-NEXT: lsr.w r5, lr, #2 +; CHECK-NEXT: .LBB18_2: @ %do.body +; CHECK-NEXT: @ =>This Loop Header: Depth=1 +; CHECK-NEXT: @ Child Loop BB18_3 Depth 2 +; CHECK-NEXT: ldr r4, [r2] +; CHECK-NEXT: dls lr, r5 +; CHECK-NEXT: vdup.32 q0, r4 +; CHECK-NEXT: .LBB18_3: @ %while.body +; CHECK-NEXT: @ Parent Loop BB18_2 Depth=1 +; CHECK-NEXT: @ => This Inner Loop Header: Depth=2 +; CHECK-NEXT: vldrw.u32 q1, [r1], #16 +; CHECK-NEXT: vldrw.u32 q2, [r0], #16 +; CHECK-NEXT: vfms.f32 q2, q0, q1 +; CHECK-NEXT: vstrb.8 q2, [r3], #16 +; CHECK-NEXT: le lr, .LBB18_3 +; CHECK-NEXT: @ %bb.4: @ %while.end +; CHECK-NEXT: @ in Loop: Header=BB18_2 Depth=1 +; CHECK-NEXT: subs.w r12, r12, #1 +; CHECK-NEXT: add.w r2, r2, #4 +; CHECK-NEXT: bne .LBB18_2 +; CHECK-NEXT: .LBB18_5: @ %do.end +; CHECK-NEXT: pop {r4, r5, r7, pc} +entry: + %shr = lshr i32 %N, 2 + %cmp15 = icmp eq i32 %shr, 0 + br i1 %cmp15, label %do.end, label %do.body + +do.body: ; preds = %entry, %while.end + %pDst.addr.0 = phi float* [ %add.ptr2, %while.end ], [ %pDst, %entry ] + %M.addr.0 = phi i32 [ %dec3, %while.end ], [ %M, %entry ] + %pSrc3.addr.0 = phi float* [ %incdec.ptr, %while.end ], [ %pSrc3, %entry ] + %pSrc2.addr.0 = phi float* [ %add.ptr1, %while.end ], [ %pSrc2, %entry ] + %pSrc1.addr.0 = phi float* [ %add.ptr, %while.end ], [ %pSrc1, %entry ] + %0 = load float, float* %pSrc3.addr.0, align 4 + %.splatinsert = insertelement <4 x float> undef, float %0, i32 0 + %.splat = shufflevector <4 x float> %.splatinsert, <4 x float> undef, <4 x i32> zeroinitializer + br label %while.body + +while.body: ; preds = %do.body, %while.body + %pSrc1.addr.119 = phi float* [ %pSrc1.addr.0, %do.body ], [ %add.ptr, %while.body ] + %pSrc2.addr.118 = phi float* [ %pSrc2.addr.0, %do.body ], [ %add.ptr1, %while.body ] + %blkCnt.017 = phi i32 [ %shr, %do.body ], [ %dec, %while.body ] + %pDst.addr.116 = phi float* [ %pDst.addr.0, %do.body ], [ %add.ptr2, %while.body ] + %1 = bitcast float* %pSrc1.addr.119 to <4 x float>* + %2 = load <4 x float>, <4 x float>* %1, align 4 + %3 = bitcast float* %pSrc2.addr.118 to <4 x float>* + %4 = load <4 x float>, <4 x float>* %3, align 4 + %5 = fneg fast <4 x float> %4 + %6 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %.splat, <4 x float> %5, <4 x float> %2) + %7 = bitcast float* %pDst.addr.116 to <4 x float>* + store <4 x float> %6, <4 x float>* %7, align 4 + %add.ptr = getelementptr inbounds float, float* %pSrc1.addr.119, i32 4 + %add.ptr1 = getelementptr inbounds float, float* %pSrc2.addr.118, i32 4 + %add.ptr2 = getelementptr inbounds float, float* %pDst.addr.116, i32 4 + %dec = add nsw i32 %blkCnt.017, -1 + %cmp = icmp eq i32 %dec, 0 + br i1 %cmp, label %while.end, label %while.body + +while.end: ; preds = %while.body + %incdec.ptr = getelementptr inbounds float, float* %pSrc3.addr.0, i32 1 + %dec3 = add i32 %M.addr.0, -1 + %cmp4 = icmp eq i32 %dec3, 0 + br i1 %cmp4, label %do.end, label %do.body + +do.end: ; preds = %while.end, %entry + ret void +} + + +%struct.arm_biquad_casd_df1_inst_f32 = type { i32, float*, float* } +define arm_aapcs_vfpcc void @arm_biquad_cascade_df1_f32(%struct.arm_biquad_casd_df1_inst_f32* nocapture readonly %S, float* nocapture readonly %pSrc, float* nocapture %pDst, i32 %blockSize) { +; CHECK-LABEL: arm_biquad_cascade_df1_f32: +; CHECK: @ %bb.0: @ %entry +; CHECK-NEXT: .save {r4, r5, r6, r7, r8, r9, r10, r11, lr} +; CHECK-NEXT: push.w {r4, r5, r6, r7, r8, r9, r10, r11, lr} +; CHECK-NEXT: .pad #4 +; CHECK-NEXT: sub sp, #4 +; CHECK-NEXT: .vsave {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: vpush {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: .pad #88 +; CHECK-NEXT: sub sp, #88 +; CHECK-NEXT: ldrd r12, r10, [r0] +; CHECK-NEXT: @ implicit-def: $s2 +; CHECK-NEXT: and r7, r3, #3 +; CHECK-NEXT: ldr.w r11, [r0, #8] +; CHECK-NEXT: lsrs r0, r3, #2 +; CHECK-NEXT: str r0, [sp, #60] @ 4-byte Spill +; CHECK-NEXT: str r7, [sp, #12] @ 4-byte Spill +; CHECK-NEXT: str r2, [sp, #56] @ 4-byte Spill +; CHECK-NEXT: b .LBB19_3 +; CHECK-NEXT: .LBB19_1: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: vmov.f32 s14, s7 +; CHECK-NEXT: ldr r2, [sp, #56] @ 4-byte Reload +; CHECK-NEXT: vmov.f32 s0, s10 +; CHECK-NEXT: vmov.f32 s7, s6 +; CHECK-NEXT: .LBB19_2: @ %if.end69 +; CHECK-NEXT: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: vstr s8, [r10] +; CHECK-NEXT: subs.w r12, r12, #1 +; CHECK-NEXT: vstr s0, [r10, #4] +; CHECK-NEXT: add.w r11, r11, #128 +; CHECK-NEXT: vstr s14, [r10, #8] +; CHECK-NEXT: mov r1, r2 +; CHECK-NEXT: vstr s7, [r10, #12] +; CHECK-NEXT: add.w r10, r10, #16 +; CHECK-NEXT: beq.w .LBB19_13 +; CHECK-NEXT: .LBB19_3: @ %do.body +; CHECK-NEXT: @ =>This Loop Header: Depth=1 +; CHECK-NEXT: @ Child Loop BB19_5 Depth 2 +; CHECK-NEXT: vldr s7, [r10, #8] +; CHECK-NEXT: mov r5, r2 +; CHECK-NEXT: ldr r0, [sp, #60] @ 4-byte Reload +; CHECK-NEXT: vldr s8, [r10] +; CHECK-NEXT: vldr s10, [r10, #4] +; CHECK-NEXT: vldr s6, [r10, #12] +; CHECK-NEXT: wls lr, r0, .LBB19_6 +; CHECK-NEXT: @ %bb.4: @ %while.body.lr.ph +; CHECK-NEXT: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: ldrd r5, lr, [sp, #56] @ 8-byte Folded Reload +; CHECK-NEXT: .LBB19_5: @ %while.body +; CHECK-NEXT: @ Parent Loop BB19_3 Depth=1 +; CHECK-NEXT: @ => This Inner Loop Header: Depth=2 +; CHECK-NEXT: vmov r4, s8 +; CHECK-NEXT: vldr s8, [r1, #12] +; CHECK-NEXT: vldrw.u32 q0, [r11, #112] +; CHECK-NEXT: vmov r0, s10 +; CHECK-NEXT: vldr s10, [r1, #8] +; CHECK-NEXT: vmov r7, s7 +; CHECK-NEXT: vmov r9, s6 +; CHECK-NEXT: vldrw.u32 q1, [r11] +; CHECK-NEXT: vstrw.32 q0, [sp, #64] @ 16-byte Spill +; CHECK-NEXT: vmov r8, s8 +; CHECK-NEXT: vldrw.u32 q0, [r11, #16] +; CHECK-NEXT: ldr r6, [r1, #4] +; CHECK-NEXT: vldrw.u32 q7, [r11, #32] +; CHECK-NEXT: vmul.f32 q1, q1, r8 +; CHECK-NEXT: vmov r3, s10 +; CHECK-NEXT: vldrw.u32 q3, [r11, #48] +; CHECK-NEXT: vfma.f32 q1, q0, r3 +; CHECK-NEXT: ldr r3, [r1] +; CHECK-NEXT: vfma.f32 q1, q7, r6 +; CHECK-NEXT: vldrw.u32 q6, [r11, #64] +; CHECK-NEXT: vfma.f32 q1, q3, r3 +; CHECK-NEXT: vldrw.u32 q5, [r11, #80] +; CHECK-NEXT: vfma.f32 q1, q6, r4 +; CHECK-NEXT: vldrw.u32 q4, [r11, #96] +; CHECK-NEXT: vfma.f32 q1, q5, r0 +; CHECK-NEXT: vldrw.u32 q0, [sp, #64] @ 16-byte Reload +; CHECK-NEXT: vfma.f32 q1, q4, r7 +; CHECK-NEXT: adds r1, #16 +; CHECK-NEXT: vfma.f32 q1, q0, r9 +; CHECK-NEXT: vmov.f32 s2, s8 +; CHECK-NEXT: vstrb.8 q1, [r5], #16 +; CHECK-NEXT: le lr, .LBB19_5 +; CHECK-NEXT: .LBB19_6: @ %while.end +; CHECK-NEXT: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: ldr r7, [sp, #12] @ 4-byte Reload +; CHECK-NEXT: cmp r7, #0 +; CHECK-NEXT: beq .LBB19_1 +; CHECK-NEXT: @ %bb.7: @ %if.then +; CHECK-NEXT: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: vldr s24, [r1] +; CHECK-NEXT: vmov r0, s8 +; CHECK-NEXT: vldr s0, [r1, #4] +; CHECK-NEXT: vldrw.u32 q3, [r11] +; CHECK-NEXT: vldr s3, [r1, #12] +; CHECK-NEXT: vldrw.u32 q4, [r11, #32] +; CHECK-NEXT: vldr s1, [r1, #8] +; CHECK-NEXT: vmov r1, s10 +; CHECK-NEXT: vldrw.u32 q2, [r11, #96] +; CHECK-NEXT: vmov r6, s3 +; CHECK-NEXT: vmul.f32 q3, q3, r6 +; CHECK-NEXT: vmov r6, s1 +; CHECK-NEXT: vstrw.32 q2, [sp, #32] @ 16-byte Spill +; CHECK-NEXT: vldrw.u32 q2, [r11, #112] +; CHECK-NEXT: vldrw.u32 q5, [r11, #48] +; CHECK-NEXT: vmov r4, s0 +; CHECK-NEXT: vstrw.32 q2, [sp, #64] @ 16-byte Spill +; CHECK-NEXT: vldrw.u32 q2, [r11, #80] +; CHECK-NEXT: vldrw.u32 q7, [r11, #64] +; CHECK-NEXT: vmov r3, s24 +; CHECK-NEXT: vstrw.32 q2, [sp, #16] @ 16-byte Spill +; CHECK-NEXT: vldrw.u32 q2, [r11, #16] +; CHECK-NEXT: vmov r2, s7 +; CHECK-NEXT: cmp r7, #1 +; CHECK-NEXT: vfma.f32 q3, q2, r6 +; CHECK-NEXT: vldrw.u32 q2, [sp, #16] @ 16-byte Reload +; CHECK-NEXT: vfma.f32 q3, q4, r4 +; CHECK-NEXT: vmov lr, s6 +; CHECK-NEXT: vfma.f32 q3, q5, r3 +; CHECK-NEXT: vfma.f32 q3, q7, r0 +; CHECK-NEXT: vfma.f32 q3, q2, r1 +; CHECK-NEXT: vldrw.u32 q2, [sp, #32] @ 16-byte Reload +; CHECK-NEXT: vfma.f32 q3, q2, r2 +; CHECK-NEXT: vldrw.u32 q2, [sp, #64] @ 16-byte Reload +; CHECK-NEXT: vfma.f32 q3, q2, lr +; CHECK-NEXT: bne .LBB19_9 +; CHECK-NEXT: @ %bb.8: @ %if.then58 +; CHECK-NEXT: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: vstr s12, [r5] +; CHECK-NEXT: vmov.f32 s8, s24 +; CHECK-NEXT: vmov.f32 s0, s2 +; CHECK-NEXT: vmov.f32 s14, s12 +; CHECK-NEXT: b .LBB19_11 +; CHECK-NEXT: .LBB19_9: @ %if.else +; CHECK-NEXT: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: cmp r7, #2 +; CHECK-NEXT: vstmia r5, {s12, s13} +; CHECK-NEXT: bne .LBB19_12 +; CHECK-NEXT: @ %bb.10: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: vmov.f32 s8, s0 +; CHECK-NEXT: vmov.f32 s14, s13 +; CHECK-NEXT: vmov.f32 s0, s24 +; CHECK-NEXT: vmov.f32 s7, s12 +; CHECK-NEXT: .LBB19_11: @ %if.end69 +; CHECK-NEXT: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: vmov.f32 s2, s3 +; CHECK-NEXT: ldr r2, [sp, #56] @ 4-byte Reload +; CHECK-NEXT: b .LBB19_2 +; CHECK-NEXT: .LBB19_12: @ %if.else64 +; CHECK-NEXT: @ in Loop: Header=BB19_3 Depth=1 +; CHECK-NEXT: vmov.f32 s7, s13 +; CHECK-NEXT: ldr r2, [sp, #56] @ 4-byte Reload +; CHECK-NEXT: vmov.f32 s2, s3 +; CHECK-NEXT: vstr s14, [r5, #8] +; CHECK-NEXT: vmov.f32 s8, s1 +; CHECK-NEXT: b .LBB19_2 +; CHECK-NEXT: .LBB19_13: @ %do.end +; CHECK-NEXT: add sp, #88 +; CHECK-NEXT: vpop {d8, d9, d10, d11, d12, d13, d14, d15} +; CHECK-NEXT: add sp, #4 +; CHECK-NEXT: pop.w {r4, r5, r6, r7, r8, r9, r10, r11, pc} +entry: + %pState1 = getelementptr inbounds %struct.arm_biquad_casd_df1_inst_f32, %struct.arm_biquad_casd_df1_inst_f32* %S, i32 0, i32 1 + %0 = load float*, float** %pState1, align 4 + %pCoeffs2 = getelementptr inbounds %struct.arm_biquad_casd_df1_inst_f32, %struct.arm_biquad_casd_df1_inst_f32* %S, i32 0, i32 2 + %1 = load float*, float** %pCoeffs2, align 4 + %numStages = getelementptr inbounds %struct.arm_biquad_casd_df1_inst_f32, %struct.arm_biquad_casd_df1_inst_f32* %S, i32 0, i32 0 + %2 = load i32, i32* %numStages, align 4 + %shr = lshr i32 %blockSize, 2 + %cmp201 = icmp eq i32 %shr, 0 + %and = and i32 %blockSize, 3 + %tobool = icmp eq i32 %and, 0 + %cmp57 = icmp eq i32 %and, 1 + %cmp60 = icmp eq i32 %and, 2 + br label %do.body + +do.body: ; preds = %if.end69, %entry + %pState.0 = phi float* [ %0, %entry ], [ %incdec.ptr73, %if.end69 ] + %pCoeffs.0 = phi float* [ %1, %entry ], [ %add.ptr74, %if.end69 ] + %pIn.0 = phi float* [ %pSrc, %entry ], [ %pDst, %if.end69 ] + %X3.0 = phi float [ undef, %entry ], [ %X3.2, %if.end69 ] + %stage.0 = phi i32 [ %2, %entry ], [ %dec75, %if.end69 ] + %3 = load float, float* %pState.0, align 4 + %arrayidx3 = getelementptr inbounds float, float* %pState.0, i32 1 + %4 = load float, float* %arrayidx3, align 4 + %arrayidx4 = getelementptr inbounds float, float* %pState.0, i32 2 + %5 = load float, float* %arrayidx4, align 4 + %arrayidx5 = getelementptr inbounds float, float* %pState.0, i32 3 + %6 = load float, float* %arrayidx5, align 4 + br i1 %cmp201, label %while.end, label %while.body.lr.ph + +while.body.lr.ph: ; preds = %do.body + %7 = bitcast float* %pCoeffs.0 to <4 x float>* + %arrayidx9 = getelementptr inbounds float, float* %pCoeffs.0, i32 4 + %8 = bitcast float* %arrayidx9 to <4 x float>* + %arrayidx12 = getelementptr inbounds float, float* %pCoeffs.0, i32 8 + %9 = bitcast float* %arrayidx12 to <4 x float>* + %arrayidx15 = getelementptr inbounds float, float* %pCoeffs.0, i32 12 + %10 = bitcast float* %arrayidx15 to <4 x float>* + %arrayidx18 = getelementptr inbounds float, float* %pCoeffs.0, i32 16 + %11 = bitcast float* %arrayidx18 to <4 x float>* + %arrayidx21 = getelementptr inbounds float, float* %pCoeffs.0, i32 20 + %12 = bitcast float* %arrayidx21 to <4 x float>* + %arrayidx24 = getelementptr inbounds float, float* %pCoeffs.0, i32 24 + %13 = bitcast float* %arrayidx24 to <4 x float>* + %arrayidx27 = getelementptr inbounds float, float* %pCoeffs.0, i32 28 + %14 = bitcast float* %arrayidx27 to <4 x float>* + br label %while.body + +while.body: ; preds = %while.body.lr.ph, %while.body + %sample.0208 = phi i32 [ %shr, %while.body.lr.ph ], [ %dec, %while.body ] + %pIn.1207 = phi float* [ %pIn.0, %while.body.lr.ph ], [ %incdec.ptr8, %while.body ] + %pOut.1206 = phi float* [ %pDst, %while.body.lr.ph ], [ %add.ptr, %while.body ] + %Yn2.0205 = phi float [ %6, %while.body.lr.ph ], [ %37, %while.body ] + %Yn1.0204 = phi float [ %5, %while.body.lr.ph ], [ %36, %while.body ] + %Xn2.0203 = phi float [ %4, %while.body.lr.ph ], [ %17, %while.body ] + %Xn1.0202 = phi float [ %3, %while.body.lr.ph ], [ %18, %while.body ] + %incdec.ptr = getelementptr inbounds float, float* %pIn.1207, i32 1 + %15 = load float, float* %pIn.1207, align 4 + %incdec.ptr6 = getelementptr inbounds float, float* %pIn.1207, i32 2 + %16 = load float, float* %incdec.ptr, align 4 + %incdec.ptr7 = getelementptr inbounds float, float* %pIn.1207, i32 3 + %17 = load float, float* %incdec.ptr6, align 4 + %incdec.ptr8 = getelementptr inbounds float, float* %pIn.1207, i32 4 + %18 = load float, float* %incdec.ptr7, align 4 + %19 = load <4 x float>, <4 x float>* %7, align 4 + %.splatinsert = insertelement <4 x float> undef, float %18, i32 0 + %.splat = shufflevector <4 x float> %.splatinsert, <4 x float> undef, <4 x i32> zeroinitializer + %20 = fmul fast <4 x float> %.splat, %19 + %21 = load <4 x float>, <4 x float>* %8, align 4 + %.splatinsert10 = insertelement <4 x float> undef, float %17, i32 0 + %.splat11 = shufflevector <4 x float> %.splatinsert10, <4 x float> undef, <4 x i32> zeroinitializer + %22 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %21, <4 x float> %.splat11, <4 x float> %20) + %23 = load <4 x float>, <4 x float>* %9, align 4 + %.splatinsert13 = insertelement <4 x float> undef, float %16, i32 0 + %.splat14 = shufflevector <4 x float> %.splatinsert13, <4 x float> undef, <4 x i32> zeroinitializer + %24 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %23, <4 x float> %.splat14, <4 x float> %22) + %25 = load <4 x float>, <4 x float>* %10, align 4 + %.splatinsert16 = insertelement <4 x float> undef, float %15, i32 0 + %.splat17 = shufflevector <4 x float> %.splatinsert16, <4 x float> undef, <4 x i32> zeroinitializer + %26 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %25, <4 x float> %.splat17, <4 x float> %24) + %27 = load <4 x float>, <4 x float>* %11, align 4 + %.splatinsert19 = insertelement <4 x float> undef, float %Xn1.0202, i32 0 + %.splat20 = shufflevector <4 x float> %.splatinsert19, <4 x float> undef, <4 x i32> zeroinitializer + %28 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %27, <4 x float> %.splat20, <4 x float> %26) + %29 = load <4 x float>, <4 x float>* %12, align 4 + %.splatinsert22 = insertelement <4 x float> undef, float %Xn2.0203, i32 0 + %.splat23 = shufflevector <4 x float> %.splatinsert22, <4 x float> undef, <4 x i32> zeroinitializer + %30 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %29, <4 x float> %.splat23, <4 x float> %28) + %31 = load <4 x float>, <4 x float>* %13, align 4 + %.splatinsert25 = insertelement <4 x float> undef, float %Yn1.0204, i32 0 + %.splat26 = shufflevector <4 x float> %.splatinsert25, <4 x float> undef, <4 x i32> zeroinitializer + %32 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %31, <4 x float> %.splat26, <4 x float> %30) + %33 = load <4 x float>, <4 x float>* %14, align 4 + %.splatinsert28 = insertelement <4 x float> undef, float %Yn2.0205, i32 0 + %.splat29 = shufflevector <4 x float> %.splatinsert28, <4 x float> undef, <4 x i32> zeroinitializer + %34 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %33, <4 x float> %.splat29, <4 x float> %32) + %35 = bitcast float* %pOut.1206 to <4 x float>* + store <4 x float> %34, <4 x float>* %35, align 4 + %add.ptr = getelementptr inbounds float, float* %pOut.1206, i32 4 + %36 = extractelement <4 x float> %34, i32 3 + %37 = extractelement <4 x float> %34, i32 2 + %dec = add nsw i32 %sample.0208, -1 + %cmp = icmp eq i32 %dec, 0 + br i1 %cmp, label %while.end, label %while.body + +while.end: ; preds = %while.body, %do.body + %Xn1.0.lcssa = phi float [ %3, %do.body ], [ %18, %while.body ] + %Xn2.0.lcssa = phi float [ %4, %do.body ], [ %17, %while.body ] + %Yn1.0.lcssa = phi float [ %5, %do.body ], [ %36, %while.body ] + %Yn2.0.lcssa = phi float [ %6, %do.body ], [ %37, %while.body ] + %pOut.1.lcssa = phi float* [ %pDst, %do.body ], [ %add.ptr, %while.body ] + %pIn.1.lcssa = phi float* [ %pIn.0, %do.body ], [ %incdec.ptr8, %while.body ] + %X3.1.lcssa = phi float [ %X3.0, %do.body ], [ %18, %while.body ] + br i1 %tobool, label %if.end69, label %if.then + +if.then: ; preds = %while.end + %incdec.ptr30 = getelementptr inbounds float, float* %pIn.1.lcssa, i32 1 + %38 = load float, float* %pIn.1.lcssa, align 4 + %incdec.ptr31 = getelementptr inbounds float, float* %pIn.1.lcssa, i32 2 + %39 = load float, float* %incdec.ptr30, align 4 + %incdec.ptr32 = getelementptr inbounds float, float* %pIn.1.lcssa, i32 3 + %40 = load float, float* %incdec.ptr31, align 4 + %41 = load float, float* %incdec.ptr32, align 4 + %42 = bitcast float* %pCoeffs.0 to <4 x float>* + %43 = load <4 x float>, <4 x float>* %42, align 4 + %.splatinsert34 = insertelement <4 x float> undef, float %41, i32 0 + %.splat35 = shufflevector <4 x float> %.splatinsert34, <4 x float> undef, <4 x i32> zeroinitializer + %44 = fmul fast <4 x float> %.splat35, %43 + %arrayidx36 = getelementptr inbounds float, float* %pCoeffs.0, i32 4 + %45 = bitcast float* %arrayidx36 to <4 x float>* + %46 = load <4 x float>, <4 x float>* %45, align 4 + %.splatinsert37 = insertelement <4 x float> undef, float %40, i32 0 + %.splat38 = shufflevector <4 x float> %.splatinsert37, <4 x float> undef, <4 x i32> zeroinitializer + %47 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %46, <4 x float> %.splat38, <4 x float> %44) + %arrayidx39 = getelementptr inbounds float, float* %pCoeffs.0, i32 8 + %48 = bitcast float* %arrayidx39 to <4 x float>* + %49 = load <4 x float>, <4 x float>* %48, align 4 + %.splatinsert40 = insertelement <4 x float> undef, float %39, i32 0 + %.splat41 = shufflevector <4 x float> %.splatinsert40, <4 x float> undef, <4 x i32> zeroinitializer + %50 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %49, <4 x float> %.splat41, <4 x float> %47) + %arrayidx42 = getelementptr inbounds float, float* %pCoeffs.0, i32 12 + %51 = bitcast float* %arrayidx42 to <4 x float>* + %52 = load <4 x float>, <4 x float>* %51, align 4 + %.splatinsert43 = insertelement <4 x float> undef, float %38, i32 0 + %.splat44 = shufflevector <4 x float> %.splatinsert43, <4 x float> undef, <4 x i32> zeroinitializer + %53 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %52, <4 x float> %.splat44, <4 x float> %50) + %arrayidx45 = getelementptr inbounds float, float* %pCoeffs.0, i32 16 + %54 = bitcast float* %arrayidx45 to <4 x float>* + %55 = load <4 x float>, <4 x float>* %54, align 4 + %.splatinsert46 = insertelement <4 x float> undef, float %Xn1.0.lcssa, i32 0 + %.splat47 = shufflevector <4 x float> %.splatinsert46, <4 x float> undef, <4 x i32> zeroinitializer + %56 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %55, <4 x float> %.splat47, <4 x float> %53) + %arrayidx48 = getelementptr inbounds float, float* %pCoeffs.0, i32 20 + %57 = bitcast float* %arrayidx48 to <4 x float>* + %58 = load <4 x float>, <4 x float>* %57, align 4 + %.splatinsert49 = insertelement <4 x float> undef, float %Xn2.0.lcssa, i32 0 + %.splat50 = shufflevector <4 x float> %.splatinsert49, <4 x float> undef, <4 x i32> zeroinitializer + %59 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %58, <4 x float> %.splat50, <4 x float> %56) + %arrayidx51 = getelementptr inbounds float, float* %pCoeffs.0, i32 24 + %60 = bitcast float* %arrayidx51 to <4 x float>* + %61 = load <4 x float>, <4 x float>* %60, align 4 + %.splatinsert52 = insertelement <4 x float> undef, float %Yn1.0.lcssa, i32 0 + %.splat53 = shufflevector <4 x float> %.splatinsert52, <4 x float> undef, <4 x i32> zeroinitializer + %62 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %61, <4 x float> %.splat53, <4 x float> %59) + %arrayidx54 = getelementptr inbounds float, float* %pCoeffs.0, i32 28 + %63 = bitcast float* %arrayidx54 to <4 x float>* + %64 = load <4 x float>, <4 x float>* %63, align 4 + %.splatinsert55 = insertelement <4 x float> undef, float %Yn2.0.lcssa, i32 0 + %.splat56 = shufflevector <4 x float> %.splatinsert55, <4 x float> undef, <4 x i32> zeroinitializer + %65 = tail call fast <4 x float> @llvm.fma.v4f32(<4 x float> %64, <4 x float> %.splat56, <4 x float> %62) + %66 = extractelement <4 x float> %65, i32 0 + br i1 %cmp57, label %if.then58, label %if.else + +if.then58: ; preds = %if.then + store float %66, float* %pOut.1.lcssa, align 4 + br label %if.end69 + +if.else: ; preds = %if.then + %incdec.ptr62 = getelementptr inbounds float, float* %pOut.1.lcssa, i32 1 + store float %66, float* %pOut.1.lcssa, align 4 + %67 = extractelement <4 x float> %65, i32 1 + store float %67, float* %incdec.ptr62, align 4 + br i1 %cmp60, label %if.end69, label %if.else64 + +if.else64: ; preds = %if.else + %incdec.ptr63 = getelementptr inbounds float, float* %pOut.1.lcssa, i32 2 + %68 = extractelement <4 x float> %65, i32 2 + store float %68, float* %incdec.ptr63, align 4 + br label %if.end69 + +if.end69: ; preds = %if.else, %while.end, %if.then58, %if.else64 + %Xn1.1 = phi float [ %38, %if.then58 ], [ %40, %if.else64 ], [ %Xn1.0.lcssa, %while.end ], [ %39, %if.else ] + %Xn2.1 = phi float [ %X3.1.lcssa, %if.then58 ], [ %39, %if.else64 ], [ %Xn2.0.lcssa, %while.end ], [ %38, %if.else ] + %Yn1.1 = phi float [ %66, %if.then58 ], [ %68, %if.else64 ], [ %Yn1.0.lcssa, %while.end ], [ %67, %if.else ] + %Yn2.1 = phi float [ %Yn1.0.lcssa, %if.then58 ], [ %67, %if.else64 ], [ %Yn2.0.lcssa, %while.end ], [ %66, %if.else ] + %X3.2 = phi float [ %41, %if.then58 ], [ %41, %if.else64 ], [ %X3.1.lcssa, %while.end ], [ %41, %if.else ] + store float %Xn1.1, float* %pState.0, align 4 + store float %Xn2.1, float* %arrayidx3, align 4 + store float %Yn1.1, float* %arrayidx4, align 4 + %incdec.ptr73 = getelementptr inbounds float, float* %pState.0, i32 4 + store float %Yn2.1, float* %arrayidx5, align 4 + %add.ptr74 = getelementptr inbounds float, float* %pCoeffs.0, i32 32 + %dec75 = add i32 %stage.0, -1 + %cmp76 = icmp eq i32 %dec75, 0 + br i1 %cmp76, label %do.end, label %do.body + +do.end: ; preds = %if.end69 + ret void +} + +declare void @llvm.lifetime.end.p0i8(i64 immarg, i8* nocapture) +declare void @llvm.lifetime.start.p0i8(i64 immarg, i8* nocapture) +declare { <4 x i32>, i32 } @llvm.arm.mve.viwdup.v4i32(i32, i32, i32) +declare <4 x i32> @llvm.arm.mve.vldr.gather.offset.v4i32.p0i32.v4i32(i32*, <4 x i32>, i32, i32, i32) declare void @llvm.assume(i1) declare <4 x i1> @llvm.arm.mve.vctp32(i32) declare <4 x float> @llvm.fma.v4f32(<4 x float>, <4 x float>, <4 x float>)