| # RUN: llc --verify-machineinstrs -mtriple=aarch64 -mcpu=neoverse-n1 -o - %s -run-pass pipeliner -aarch64-enable-pipeliner -debug-only=pipeliner 2>&1 | FileCheck %s |
| # REQUIRES: asserts |
| |
| # Check that instructions referencing NZCV are not pipelined |
| |
| # CHECK: SU([[SU0:[0-9]+]]): nofpexcept FCMPSri {{.*}}, implicit-def $nzcv, implicit $fpcr |
| # CHECK: SU([[SU1:[0-9]+]]): {{.*}} = FCSELSrrr {{.*}}, {{.*}}, 1, implicit $nzcv |
| # CHECK: Do not pipeline SU([[SU0:[0-9]+]]) |
| # CHECK: Do not pipeline SU([[SU1:[0-9]+]]) |
| |
| --- | |
| define dso_local void @KERNEL(ptr noalias nocapture noundef writeonly %a, ptr nocapture noundef readonly %b, i32 noundef %n) local_unnamed_addr #0 { |
| entry: |
| %cmp19 = icmp sgt i32 %n, 0 |
| br i1 %cmp19, label %for.body.preheader, label %for.cond.cleanup |
| |
| for.body.preheader: ; preds = %entry |
| %wide.trip.count = zext nneg i32 %n to i64 |
| br label %for.body |
| |
| for.cond.cleanup: ; preds = %for.body, %entry |
| ret void |
| |
| for.body: ; preds = %for.body.preheader, %for.body |
| %lsr.iv24 = phi i64 [ %wide.trip.count, %for.body.preheader ], [ %lsr.iv.next, %for.body ] |
| %lsr.iv22 = phi ptr [ %b, %for.body.preheader ], [ %scevgep23, %for.body ] |
| %lsr.iv = phi ptr [ %a, %for.body.preheader ], [ %scevgep, %for.body ] |
| %0 = load float, ptr %lsr.iv22, align 4 |
| %tobool = fcmp une float %0, 0.000000e+00 |
| %. = select i1 %tobool, float 1.000000e+00, float 2.000000e+00 |
| %add = fadd float %0, %. |
| store float %add, ptr %lsr.iv, align 4 |
| %scevgep = getelementptr i8, ptr %lsr.iv, i64 4 |
| %scevgep23 = getelementptr i8, ptr %lsr.iv22, i64 4 |
| %lsr.iv.next = add nsw i64 %lsr.iv24, -1 |
| %exitcond.not = icmp eq i64 %lsr.iv.next, 0 |
| br i1 %exitcond.not, label %for.cond.cleanup, label %for.body |
| } |
| |
| ... |
| --- |
| name: KERNEL |
| tracksRegLiveness: true |
| liveins: |
| - { reg: '$x0', virtual-reg: '%7' } |
| - { reg: '$x1', virtual-reg: '%8' } |
| - { reg: '$w2', virtual-reg: '%9' } |
| body: | |
| bb.0.entry: |
| successors: %bb.1(0x50000000), %bb.2(0x30000000) |
| liveins: $x0, $x1, $w2 |
| |
| %9:gpr32common = COPY $w2 |
| %8:gpr64 = COPY $x1 |
| %7:gpr64 = COPY $x0 |
| dead $wzr = SUBSWri %9, 1, 0, implicit-def $nzcv |
| Bcc 11, %bb.2, implicit $nzcv |
| B %bb.1 |
| |
| bb.1.for.body.preheader: |
| %11:gpr32 = ORRWrs $wzr, %9, 0 |
| %0:gpr64all = SUBREG_TO_REG 0, killed %11, %subreg.sub_32 |
| %14:fpr32 = FMOVSi 0 |
| %15:fpr32 = FMOVSi 112 |
| B %bb.3 |
| |
| bb.2.for.cond.cleanup: |
| RET_ReallyLR |
| |
| bb.3.for.body: |
| successors: %bb.2(0x04000000), %bb.3(0x7c000000) |
| |
| %1:gpr64sp = PHI %0, %bb.1, %6, %bb.3 |
| %2:gpr64sp = PHI %8, %bb.1, %5, %bb.3 |
| %3:gpr64sp = PHI %7, %bb.1, %4, %bb.3 |
| early-clobber %12:gpr64sp, %13:fpr32 = LDRSpost %2, 4 :: (load (s32) from %ir.lsr.iv22) |
| nofpexcept FCMPSri %13, implicit-def $nzcv, implicit $fpcr |
| %16:fpr32 = FCSELSrrr %15, %14, 1, implicit $nzcv |
| %17:fpr32 = nofpexcept FADDSrr %13, killed %16, implicit $fpcr |
| early-clobber %18:gpr64sp = STRSpost killed %17, %3, 4 :: (store (s32) into %ir.lsr.iv) |
| %4:gpr64all = COPY %18 |
| %5:gpr64all = COPY %12 |
| %19:gpr64 = nsw SUBSXri %1, 1, 0, implicit-def $nzcv |
| %6:gpr64all = COPY %19 |
| Bcc 0, %bb.2, implicit $nzcv |
| B %bb.3 |
| |
| ... |