| # RUN: llc --verify-machineinstrs -mtriple=aarch64 -o - %s -run-pass pipeliner -aarch64-enable-pipeliner -pipeliner-enable-copytophi=0 -debug-only=pipeliner 2>&1 | FileCheck %s |
| # REQUIRES: asserts |
| |
| # Compare and branch instructions are not supported now. |
| # CHECK: Unable to analyzeLoop, can NOT pipeline Loop |
| |
| # (if supported) Check that instructions for loop control are not pipelined when compare and branch is used. |
| # CHECK(if supported): SU([[SU0:[0-9]+]]): [[V0:%[0-9]+]]:gpr64common = SUBXri [[V1:%[0-9]+]]:gpr64common, 1, 0 |
| # CHECK(if supported): Do not pipeline SU([[SU0:[0-9]+]]) |
| |
| --- | |
| define dso_local void @func(ptr noalias nocapture noundef writeonly %a, ptr nocapture noundef readonly %b, i32 noundef %n) local_unnamed_addr #0 { |
| entry: |
| %or.cond = icmp ult i32 %n, 2 |
| br i1 %or.cond, label %for.end, label %for.body.preheader |
| |
| for.body.preheader: ; preds = %entry |
| %i.07 = add i32 %n, -1 |
| %0 = sext i32 %i.07 to i64 |
| br label %for.body |
| |
| for.body: ; preds = %for.body.preheader, %for.body |
| %indvars.iv = phi i64 [ %0, %for.body.preheader ], [ %indvars.iv.next, %for.body ] |
| %1 = shl nsw i64 %indvars.iv, 2 |
| %scevgep = getelementptr i8, ptr %b, i64 %1 |
| %2 = load float, ptr %scevgep, align 4 |
| %add = fadd float %2, 1.000000e+00 |
| %3 = shl nsw i64 %indvars.iv, 2 |
| %scevgep11 = getelementptr i8, ptr %a, i64 %3 |
| store float %add, ptr %scevgep11, align 4 |
| %indvars.iv.next = add nsw i64 %indvars.iv, -1 |
| %4 = add i64 %indvars.iv, -1 |
| %5 = and i64 %4, 4294967295 |
| %tobool.not = icmp eq i64 %5, 0 |
| br i1 %tobool.not, label %for.end, label %for.body |
| |
| for.end: ; preds = %for.body, %entry |
| ret void |
| } |
| |
| ... |
| --- |
| name: func |
| tracksRegLiveness: true |
| liveins: |
| - { reg: '$x0', virtual-reg: '%3' } |
| - { reg: '$x1', virtual-reg: '%4' } |
| - { reg: '$w2', virtual-reg: '%5' } |
| body: | |
| bb.0.entry: |
| liveins: $x0, $x1, $w2 |
| |
| %5:gpr32common = COPY $w2 |
| %4:gpr64common = COPY $x1 |
| %3:gpr64common = COPY $x0 |
| dead $wzr = SUBSWri %5, 2, 0, implicit-def $nzcv |
| Bcc 3, %bb.3, implicit $nzcv |
| B %bb.1 |
| |
| bb.1.for.body.preheader: |
| %7:gpr32common = SUBWri %5, 1, 0 |
| %9:gpr64all = IMPLICIT_DEF |
| %8:gpr64 = SUBREG_TO_REG 0, killed %7, %subreg.sub_32 |
| %10:gpr64 = SBFMXri killed %8, 0, 31 |
| %0:gpr64all = COPY %10 |
| %12:fpr32 = FMOVSi 112 |
| |
| bb.2.for.body: |
| successors: %bb.3(0x04000000), %bb.2(0x7c000000) |
| |
| %1:gpr64common = PHI %0, %bb.1, %2, %bb.2 |
| %11:fpr32 = LDRSroX %4, %1, 0, 1 :: (load (s32) from %ir.scevgep) |
| %13:fpr32 = nofpexcept FADDSrr killed %11, %12, implicit $fpcr |
| STRSroX killed %13, %3, %1, 0, 1 :: (store (s32) into %ir.scevgep11) |
| %14:gpr64common = SUBXri %1, 1, 0 |
| %2:gpr64all = COPY %14 |
| %15:gpr32 = COPY %14.sub_32 |
| CBNZW killed %15, %bb.2 |
| B %bb.3 |
| |
| bb.3.for.end: |
| RET_ReallyLR |
| |
| ... |