1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-02-01 05:01:59 +01:00
llvm-mirror/test/CodeGen/X86/avx512-i1test.ll
Guozhi Wei 49bf5cf617 [MBP] Don't move bottom block before header if it can't reduce taken branches
If bottom of block BB has only one successor OldTop, in most cases it is profitable to move it before OldTop, except the following case:

-->OldTop<-
|    .    |
|    .    |
|    .    |
---Pred   |
     |    |
    BB-----

Move BB before OldTop can't reduce the number of taken branches, this patch detects this case and prevent the moving.

Differential Revision: https://reviews.llvm.org/D57067

llvm-svn: 352236
2019-01-25 19:45:13 +00:00

95 lines
3.1 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mattr=+avx512f | FileCheck %s --check-prefix=CHECK --check-prefix=KNL
; RUN: llc < %s -mattr=+avx512f -mattr=+avx512vl -mattr=+avx512bw -mattr=+avx512dq | FileCheck %s --check-prefix=CHECK --check-prefix=SKX
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
define void @func() {
; CHECK-LABEL: func:
; CHECK: # %bb.0: # %bb1
; CHECK-NEXT: xorl %eax, %eax
; CHECK-NEXT: testb %al, %al
; CHECK-NEXT: je .LBB0_1
; CHECK-NEXT: # %bb.4: # %L_30
; CHECK-NEXT: retq
; CHECK-NEXT: .LBB0_1: # %bb56
; CHECK-NEXT: xorl %eax, %eax
; CHECK-NEXT: .p2align 4, 0x90
; CHECK-NEXT: .LBB0_2: # %bb33
; CHECK-NEXT: # =>This Inner Loop Header: Depth=1
; CHECK-NEXT: testb %al, %al
; CHECK-NEXT: jne .LBB0_2
; CHECK-NEXT: # %bb.3: # %bb35
; CHECK-NEXT: # in Loop: Header=BB0_2 Depth=1
; CHECK-NEXT: testb %al, %al
; CHECK-NEXT: jmp .LBB0_2
bb1:
br i1 undef, label %L_10, label %L_10
L_10: ; preds = %bb1, %bb1
br i1 undef, label %L_30, label %bb56
bb56: ; preds = %L_10
br label %bb33
bb33: ; preds = %bb51, %bb56
%r111 = load i64, i64* undef, align 8
br i1 undef, label %bb51, label %bb35
bb35: ; preds = %bb33
br i1 undef, label %L_19, label %bb37
bb37: ; preds = %bb35
%r128 = and i64 %r111, 576460752303423488
%phitmp = icmp eq i64 %r128, 0
br label %L_19
L_19: ; preds = %bb37, %bb35
%"$V_S25.0" = phi i1 [ %phitmp, %bb37 ], [ true, %bb35 ]
br i1 undef, label %bb51, label %bb42
bb42: ; preds = %L_19
%r136 = select i1 %"$V_S25.0", i32* undef, i32* undef
br label %bb51
bb51: ; preds = %bb42, %L_19, %bb33
br i1 false, label %L_30, label %bb33
L_30: ; preds = %bb51, %L_10
ret void
}
; The following test generates suboptimal code on AVX-512
; PR 28175
define i64 @func2(i1 zeroext %i, i32 %j) {
; CHECK-LABEL: func2:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: testl %esi, %esi
; CHECK-NEXT: je .LBB1_1
; CHECK-NEXT: # %bb.2: # %if.then
; CHECK-NEXT: jmp bar # TAILCALL
; CHECK-NEXT: .LBB1_1: # %if.end
; CHECK-NEXT: movzbl %dil, %eax
; CHECK-NEXT: orq $-2, %rax
; CHECK-NEXT: retq
entry:
%tobool = icmp eq i32 %j, 0
br i1 %tobool, label %if.end, label %if.then
if.then: ; preds = %entry
%call = tail call i64 @bar()
br label %return
if.end: ; preds = %entry
%conv = zext i1 %i to i64
%or = or i64 %conv, -2
br label %return
return: ; preds = %if.end, %if.then
%or.sink = phi i64 [ %or, %if.end ], [ %call, %if.then ]
ret i64 %or.sink
}
declare i64 @bar()