File: arm64-misched-forwarding-A53.ll

package info (click to toggle)
llvm-toolchain-20 1%3A20.1.6-1~exp1
  • links: PTS, VCS
  • area: main
  • in suites: experimental
  • size: 2,111,304 kB
  • sloc: cpp: 7,438,677; ansic: 1,393,822; asm: 1,012,926; python: 241,650; f90: 86,635; objc: 75,479; lisp: 42,144; pascal: 17,286; sh: 10,027; ml: 5,082; perl: 4,730; awk: 3,523; makefile: 3,349; javascript: 2,251; xml: 892; fortran: 672
file content (22 lines) | stat: -rw-r--r-- 836 bytes parent folder | download | duplicates (28) 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
 
; REQUIRES: asserts
; RUN: llc < %s -mtriple=arm64-linux-gnu -mcpu=cortex-a53 -pre-RA-sched=source -enable-misched -verify-misched -debug-only=machine-scheduler -o - 2>&1 > /dev/null | FileCheck %s
;
; For Cortex-A53, shiftable operands that are not actually shifted
; are not needed for an additional two cycles.
;
; CHECK: ********** MI Scheduling **********
; CHECK: shiftable
; CHECK: SU(2):   %2:gpr64common = SUBXri %1:gpr64common, 20, 0
; CHECK:   Successors:
; CHECK-NEXT:    SU(4): Data Latency=1 Reg=%2
; CHECK-NEXT:    SU(3): Data Latency=2 Reg=%2
; CHECK: ********** INTERVALS **********
define i64 @shiftable(i64 %A, i64 %B) {
        %tmp0 = sub i64 %B, 20
        %tmp1 = shl i64 %tmp0, 5;
        %tmp2 = add i64 %A, %tmp1;
        %tmp3 = add i64 %A, %tmp0
        %tmp4 = mul i64 %tmp2, %tmp3

        ret i64 %tmp4
}