e094abde42
The previous expansion of [US]CMP was done using two selects and two compares. It produced decent code, but on many platforms it is better to implement [US]CMP nodes by performing the following operation: ``` [us]cmp(x, y) = (x [us]> y) - (x [us]< y) ``` This patch adds this new expansion, as well as a hook in TargetLowering to allow some targets to still use the select-based approach. AArch64 and SystemZ are currently the only targets to prefer the former approach, but other targets may also start to use it if it provides for better codegen.
144 lines
4.0 KiB
LLVM
144 lines
4.0 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: llc -mtriple=armv7-unknown-eabi %s -o - | FileCheck %s
|
|
|
|
define i8 @ucmp_8_8(i8 zeroext %x, i8 zeroext %y) nounwind {
|
|
; CHECK-LABEL: ucmp_8_8:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: cmp r0, r1
|
|
; CHECK-NEXT: mov r0, #0
|
|
; CHECK-NEXT: mov r2, #0
|
|
; CHECK-NEXT: movwlo r0, #1
|
|
; CHECK-NEXT: movwhi r2, #1
|
|
; CHECK-NEXT: sub r0, r2, r0
|
|
; CHECK-NEXT: bx lr
|
|
%1 = call i8 @llvm.ucmp(i8 %x, i8 %y)
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @ucmp_8_16(i16 zeroext %x, i16 zeroext %y) nounwind {
|
|
; CHECK-LABEL: ucmp_8_16:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: cmp r0, r1
|
|
; CHECK-NEXT: mov r0, #0
|
|
; CHECK-NEXT: mov r2, #0
|
|
; CHECK-NEXT: movwlo r0, #1
|
|
; CHECK-NEXT: movwhi r2, #1
|
|
; CHECK-NEXT: sub r0, r2, r0
|
|
; CHECK-NEXT: bx lr
|
|
%1 = call i8 @llvm.ucmp(i16 %x, i16 %y)
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @ucmp_8_32(i32 %x, i32 %y) nounwind {
|
|
; CHECK-LABEL: ucmp_8_32:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: cmp r0, r1
|
|
; CHECK-NEXT: mov r0, #0
|
|
; CHECK-NEXT: mov r2, #0
|
|
; CHECK-NEXT: movwlo r0, #1
|
|
; CHECK-NEXT: movwhi r2, #1
|
|
; CHECK-NEXT: sub r0, r2, r0
|
|
; CHECK-NEXT: bx lr
|
|
%1 = call i8 @llvm.ucmp(i32 %x, i32 %y)
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @ucmp_8_64(i64 %x, i64 %y) nounwind {
|
|
; CHECK-LABEL: ucmp_8_64:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: .save {r11, lr}
|
|
; CHECK-NEXT: push {r11, lr}
|
|
; CHECK-NEXT: subs lr, r0, r2
|
|
; CHECK-NEXT: mov r12, #0
|
|
; CHECK-NEXT: sbcs lr, r1, r3
|
|
; CHECK-NEXT: mov lr, #0
|
|
; CHECK-NEXT: movwlo lr, #1
|
|
; CHECK-NEXT: subs r0, r2, r0
|
|
; CHECK-NEXT: sbcs r0, r3, r1
|
|
; CHECK-NEXT: movwlo r12, #1
|
|
; CHECK-NEXT: sub r0, r12, lr
|
|
; CHECK-NEXT: pop {r11, pc}
|
|
%1 = call i8 @llvm.ucmp(i64 %x, i64 %y)
|
|
ret i8 %1
|
|
}
|
|
|
|
define i8 @ucmp_8_128(i128 %x, i128 %y) nounwind {
|
|
; CHECK-LABEL: ucmp_8_128:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: .save {r4, r5, r6, r7, r11, lr}
|
|
; CHECK-NEXT: push {r4, r5, r6, r7, r11, lr}
|
|
; CHECK-NEXT: ldr r4, [sp, #24]
|
|
; CHECK-NEXT: mov r5, #0
|
|
; CHECK-NEXT: ldr r6, [sp, #28]
|
|
; CHECK-NEXT: subs r7, r0, r4
|
|
; CHECK-NEXT: ldr r12, [sp, #32]
|
|
; CHECK-NEXT: sbcs r7, r1, r6
|
|
; CHECK-NEXT: ldr lr, [sp, #36]
|
|
; CHECK-NEXT: sbcs r7, r2, r12
|
|
; CHECK-NEXT: sbcs r7, r3, lr
|
|
; CHECK-NEXT: mov r7, #0
|
|
; CHECK-NEXT: movwlo r7, #1
|
|
; CHECK-NEXT: subs r0, r4, r0
|
|
; CHECK-NEXT: sbcs r0, r6, r1
|
|
; CHECK-NEXT: sbcs r0, r12, r2
|
|
; CHECK-NEXT: sbcs r0, lr, r3
|
|
; CHECK-NEXT: movwlo r5, #1
|
|
; CHECK-NEXT: sub r0, r5, r7
|
|
; CHECK-NEXT: pop {r4, r5, r6, r7, r11, pc}
|
|
%1 = call i8 @llvm.ucmp(i128 %x, i128 %y)
|
|
ret i8 %1
|
|
}
|
|
|
|
define i32 @ucmp_32_32(i32 %x, i32 %y) nounwind {
|
|
; CHECK-LABEL: ucmp_32_32:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: cmp r0, r1
|
|
; CHECK-NEXT: mov r0, #0
|
|
; CHECK-NEXT: mov r2, #0
|
|
; CHECK-NEXT: movwlo r0, #1
|
|
; CHECK-NEXT: movwhi r2, #1
|
|
; CHECK-NEXT: sub r0, r2, r0
|
|
; CHECK-NEXT: bx lr
|
|
%1 = call i32 @llvm.ucmp(i32 %x, i32 %y)
|
|
ret i32 %1
|
|
}
|
|
|
|
define i32 @ucmp_32_64(i64 %x, i64 %y) nounwind {
|
|
; CHECK-LABEL: ucmp_32_64:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: .save {r11, lr}
|
|
; CHECK-NEXT: push {r11, lr}
|
|
; CHECK-NEXT: subs lr, r0, r2
|
|
; CHECK-NEXT: mov r12, #0
|
|
; CHECK-NEXT: sbcs lr, r1, r3
|
|
; CHECK-NEXT: mov lr, #0
|
|
; CHECK-NEXT: movwlo lr, #1
|
|
; CHECK-NEXT: subs r0, r2, r0
|
|
; CHECK-NEXT: sbcs r0, r3, r1
|
|
; CHECK-NEXT: movwlo r12, #1
|
|
; CHECK-NEXT: sub r0, r12, lr
|
|
; CHECK-NEXT: pop {r11, pc}
|
|
%1 = call i32 @llvm.ucmp(i64 %x, i64 %y)
|
|
ret i32 %1
|
|
}
|
|
|
|
define i64 @ucmp_64_64(i64 %x, i64 %y) nounwind {
|
|
; CHECK-LABEL: ucmp_64_64:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: .save {r11, lr}
|
|
; CHECK-NEXT: push {r11, lr}
|
|
; CHECK-NEXT: subs lr, r0, r2
|
|
; CHECK-NEXT: mov r12, #0
|
|
; CHECK-NEXT: sbcs lr, r1, r3
|
|
; CHECK-NEXT: mov lr, #0
|
|
; CHECK-NEXT: movwlo lr, #1
|
|
; CHECK-NEXT: subs r0, r2, r0
|
|
; CHECK-NEXT: sbcs r0, r3, r1
|
|
; CHECK-NEXT: movwlo r12, #1
|
|
; CHECK-NEXT: sub r0, r12, lr
|
|
; CHECK-NEXT: asr r1, r0, #31
|
|
; CHECK-NEXT: pop {r11, pc}
|
|
%1 = call i64 @llvm.ucmp(i64 %x, i64 %y)
|
|
ret i64 %1
|
|
}
|