c3ebbfd736
SelectionDAG will not reassociate adds to the end of a chain if
there are multiple users of later additions. This prevents isel
from folding the immediate into a load/store address.
One easy way to see this is accessing an array in a struct with
two different indices. An ADDI will be used to get to the start
of the array then 2 different SHXADD instructions will be used to
add the scaled indices. Finally the SHXADD will be used by different
load instructions. We can remove the ADDI by folding the offset into
each load.
This patch adds a new pass that analyzes how an ADDI constant
propagates through address arithmetic. If the arithmetic is only
used by a load/store and the offset is small enough, we can adjust
the load/store offset and remove the ADDI.
This pass is placed before MachineCSE to allow cleanups if some
instructions become common after removing offsets from their inputs.
This pass gives ~3% improvement on dynamic instruction count on
541.leela_r and 544.nab_r from SPEC2017 for the train data set. There's
a ~1% improvement on 557.xz_r.
734 lines
23 KiB
LLVM
734 lines
23 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 | FileCheck %s --check-prefixes=CHECK,RV32I
|
|
; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 | FileCheck %s --check-prefixes=CHECK,RV64I
|
|
; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+zba | FileCheck %s --check-prefixes=ZBA,RV32ZBA
|
|
; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+zba | FileCheck %s --check-prefixes=ZBA,RV64ZBA
|
|
|
|
define i64 @test_sh3add(ptr %p, iXLen %x, iXLen %y) {
|
|
; RV32I-LABEL: test_sh3add:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: slli a1, a1, 3
|
|
; RV32I-NEXT: slli a2, a2, 3
|
|
; RV32I-NEXT: add a1, a1, a0
|
|
; RV32I-NEXT: add a0, a0, a2
|
|
; RV32I-NEXT: lw a2, 480(a1)
|
|
; RV32I-NEXT: lw a1, 484(a1)
|
|
; RV32I-NEXT: lw a3, 400(a0)
|
|
; RV32I-NEXT: lw a0, 404(a0)
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a3, a2
|
|
; RV32I-NEXT: sltu a2, a0, a3
|
|
; RV32I-NEXT: add a1, a1, a2
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_sh3add:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: slli a1, a1, 3
|
|
; RV64I-NEXT: slli a2, a2, 3
|
|
; RV64I-NEXT: add a1, a1, a0
|
|
; RV64I-NEXT: add a0, a0, a2
|
|
; RV64I-NEXT: ld a1, 480(a1)
|
|
; RV64I-NEXT: ld a0, 400(a0)
|
|
; RV64I-NEXT: add a0, a0, a1
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_sh3add:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: sh3add a1, a1, a0
|
|
; RV32ZBA-NEXT: sh3add a0, a2, a0
|
|
; RV32ZBA-NEXT: lw a2, 480(a1)
|
|
; RV32ZBA-NEXT: lw a1, 484(a1)
|
|
; RV32ZBA-NEXT: lw a3, 400(a0)
|
|
; RV32ZBA-NEXT: lw a0, 404(a0)
|
|
; RV32ZBA-NEXT: add a1, a0, a1
|
|
; RV32ZBA-NEXT: add a0, a3, a2
|
|
; RV32ZBA-NEXT: sltu a2, a0, a3
|
|
; RV32ZBA-NEXT: add a1, a1, a2
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_sh3add:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: sh3add a1, a1, a0
|
|
; RV64ZBA-NEXT: sh3add a0, a2, a0
|
|
; RV64ZBA-NEXT: ld a1, 480(a1)
|
|
; RV64ZBA-NEXT: ld a0, 400(a0)
|
|
; RV64ZBA-NEXT: add a0, a0, a1
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%b = getelementptr inbounds nuw i8, ptr %p, i64 400
|
|
%add = add iXLen %x, 10
|
|
%arrayidx = getelementptr inbounds nuw [100 x i64], ptr %b, i64 0, iXLen %add
|
|
%0 = load i64, ptr %arrayidx, align 8
|
|
%arrayidx2 = getelementptr inbounds nuw [100 x i64], ptr %b, i64 0, iXLen %y
|
|
%1 = load i64, ptr %arrayidx2, align 8
|
|
%add3 = add nsw i64 %1, %0
|
|
ret i64 %add3
|
|
}
|
|
|
|
define signext i32 @test_sh2add(ptr %p, iXLen %x, iXLen %y) {
|
|
; RV32I-LABEL: test_sh2add:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: slli a1, a1, 2
|
|
; RV32I-NEXT: slli a2, a2, 2
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a2, a0
|
|
; RV32I-NEXT: lw a1, 1200(a1)
|
|
; RV32I-NEXT: lw a0, 1240(a0)
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_sh2add:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: slli a1, a1, 2
|
|
; RV64I-NEXT: slli a2, a2, 2
|
|
; RV64I-NEXT: add a1, a0, a1
|
|
; RV64I-NEXT: add a0, a2, a0
|
|
; RV64I-NEXT: lw a1, 1200(a1)
|
|
; RV64I-NEXT: lw a0, 1240(a0)
|
|
; RV64I-NEXT: addw a0, a0, a1
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_sh2add:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: sh2add a1, a1, a0
|
|
; RV32ZBA-NEXT: sh2add a0, a2, a0
|
|
; RV32ZBA-NEXT: lw a1, 1200(a1)
|
|
; RV32ZBA-NEXT: lw a0, 1240(a0)
|
|
; RV32ZBA-NEXT: add a0, a0, a1
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_sh2add:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: sh2add a1, a1, a0
|
|
; RV64ZBA-NEXT: sh2add a0, a2, a0
|
|
; RV64ZBA-NEXT: lw a1, 1200(a1)
|
|
; RV64ZBA-NEXT: lw a0, 1240(a0)
|
|
; RV64ZBA-NEXT: addw a0, a0, a1
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%c = getelementptr inbounds nuw i8, ptr %p, i64 1200
|
|
%arrayidx = getelementptr inbounds nuw [100 x i32], ptr %c, i64 0, iXLen %x
|
|
%0 = load i32, ptr %arrayidx, align 4
|
|
%add = add iXLen %y, 10
|
|
%arrayidx2 = getelementptr inbounds nuw [100 x i32], ptr %c, i64 0, iXLen %add
|
|
%1 = load i32, ptr %arrayidx2, align 4
|
|
%add3 = add nsw i32 %1, %0
|
|
ret i32 %add3
|
|
}
|
|
|
|
define signext i16 @test_sh1add(ptr %p, iXLen %x, iXLen %y) {
|
|
; RV32I-LABEL: test_sh1add:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: slli a1, a1, 1
|
|
; RV32I-NEXT: slli a2, a2, 1
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a2, a0
|
|
; RV32I-NEXT: lh a1, 1600(a1)
|
|
; RV32I-NEXT: lh a0, 1620(a0)
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: slli a0, a0, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_sh1add:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: slli a1, a1, 1
|
|
; RV64I-NEXT: slli a2, a2, 1
|
|
; RV64I-NEXT: add a1, a0, a1
|
|
; RV64I-NEXT: add a0, a2, a0
|
|
; RV64I-NEXT: lh a1, 1600(a1)
|
|
; RV64I-NEXT: lh a0, 1620(a0)
|
|
; RV64I-NEXT: add a0, a0, a1
|
|
; RV64I-NEXT: slli a0, a0, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_sh1add:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: sh1add a1, a1, a0
|
|
; RV32ZBA-NEXT: sh1add a0, a2, a0
|
|
; RV32ZBA-NEXT: lh a1, 1600(a1)
|
|
; RV32ZBA-NEXT: lh a0, 1620(a0)
|
|
; RV32ZBA-NEXT: add a0, a0, a1
|
|
; RV32ZBA-NEXT: slli a0, a0, 16
|
|
; RV32ZBA-NEXT: srai a0, a0, 16
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_sh1add:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: sh1add a1, a1, a0
|
|
; RV64ZBA-NEXT: sh1add a0, a2, a0
|
|
; RV64ZBA-NEXT: lh a1, 1600(a1)
|
|
; RV64ZBA-NEXT: lh a0, 1620(a0)
|
|
; RV64ZBA-NEXT: add a0, a0, a1
|
|
; RV64ZBA-NEXT: slli a0, a0, 48
|
|
; RV64ZBA-NEXT: srai a0, a0, 48
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%d = getelementptr inbounds nuw i8, ptr %p, i64 1600
|
|
%arrayidx = getelementptr inbounds nuw [100 x i16], ptr %d, i64 0, iXLen %x
|
|
%0 = load i16, ptr %arrayidx, align 2
|
|
%add = add iXLen %y, 10
|
|
%arrayidx2 = getelementptr inbounds nuw [100 x i16], ptr %d, i64 0, iXLen %add
|
|
%1 = load i16, ptr %arrayidx2, align 2
|
|
%add4 = add i16 %1, %0
|
|
ret i16 %add4
|
|
}
|
|
|
|
define zeroext i8 @test_add(ptr %p, iXLen %x, iXLen %y) {
|
|
; CHECK-LABEL: test_add:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: add a1, a0, a1
|
|
; CHECK-NEXT: add a0, a2, a0
|
|
; CHECK-NEXT: lbu a1, 1800(a1)
|
|
; CHECK-NEXT: lbu a0, 1810(a0)
|
|
; CHECK-NEXT: add a0, a0, a1
|
|
; CHECK-NEXT: andi a0, a0, 255
|
|
; CHECK-NEXT: ret
|
|
;
|
|
; ZBA-LABEL: test_add:
|
|
; ZBA: # %bb.0: # %entry
|
|
; ZBA-NEXT: add a1, a0, a1
|
|
; ZBA-NEXT: add a0, a2, a0
|
|
; ZBA-NEXT: lbu a1, 1800(a1)
|
|
; ZBA-NEXT: lbu a0, 1810(a0)
|
|
; ZBA-NEXT: add a0, a0, a1
|
|
; ZBA-NEXT: andi a0, a0, 255
|
|
; ZBA-NEXT: ret
|
|
entry:
|
|
%e = getelementptr inbounds nuw i8, ptr %p, i64 1800
|
|
%arrayidx = getelementptr inbounds nuw [1000 x i8], ptr %e, i64 0, iXLen %x
|
|
%0 = load i8, ptr %arrayidx, align 1
|
|
%add = add iXLen %y, 10
|
|
%arrayidx2 = getelementptr inbounds nuw [1000 x i8], ptr %e, i64 0, iXLen %add
|
|
%1 = load i8, ptr %arrayidx2, align 1
|
|
%add4 = add i8 %1, %0
|
|
ret i8 %add4
|
|
}
|
|
|
|
define i64 @test_sh3add_uw(ptr %p, i32 signext %x, i32 signext %y) {
|
|
; RV32I-LABEL: test_sh3add_uw:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: slli a1, a1, 3
|
|
; RV32I-NEXT: slli a2, a2, 3
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a0, a2
|
|
; RV32I-NEXT: lw a2, 400(a1)
|
|
; RV32I-NEXT: lw a1, 404(a1)
|
|
; RV32I-NEXT: lw a3, 400(a0)
|
|
; RV32I-NEXT: lw a0, 404(a0)
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a3, a2
|
|
; RV32I-NEXT: sltu a2, a0, a3
|
|
; RV32I-NEXT: add a1, a1, a2
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_sh3add_uw:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: slli a1, a1, 32
|
|
; RV64I-NEXT: slli a2, a2, 32
|
|
; RV64I-NEXT: srli a1, a1, 29
|
|
; RV64I-NEXT: srli a2, a2, 29
|
|
; RV64I-NEXT: add a1, a0, a1
|
|
; RV64I-NEXT: add a0, a0, a2
|
|
; RV64I-NEXT: ld a1, 400(a1)
|
|
; RV64I-NEXT: ld a0, 400(a0)
|
|
; RV64I-NEXT: add a0, a0, a1
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_sh3add_uw:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: sh3add a1, a1, a0
|
|
; RV32ZBA-NEXT: sh3add a0, a2, a0
|
|
; RV32ZBA-NEXT: lw a2, 400(a1)
|
|
; RV32ZBA-NEXT: lw a1, 404(a1)
|
|
; RV32ZBA-NEXT: lw a3, 400(a0)
|
|
; RV32ZBA-NEXT: lw a0, 404(a0)
|
|
; RV32ZBA-NEXT: add a1, a0, a1
|
|
; RV32ZBA-NEXT: add a0, a3, a2
|
|
; RV32ZBA-NEXT: sltu a2, a0, a3
|
|
; RV32ZBA-NEXT: add a1, a1, a2
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_sh3add_uw:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: sh3add.uw a1, a1, a0
|
|
; RV64ZBA-NEXT: sh3add.uw a0, a2, a0
|
|
; RV64ZBA-NEXT: ld a1, 400(a1)
|
|
; RV64ZBA-NEXT: ld a0, 400(a0)
|
|
; RV64ZBA-NEXT: add a0, a0, a1
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%b = getelementptr inbounds nuw i8, ptr %p, i64 400
|
|
%idxprom = zext i32 %x to i64
|
|
%arrayidx = getelementptr inbounds nuw [100 x i64], ptr %b, i64 0, i64 %idxprom
|
|
%0 = load i64, ptr %arrayidx, align 8
|
|
%idxprom2 = zext i32 %y to i64
|
|
%arrayidx3 = getelementptr inbounds nuw [100 x i64], ptr %b, i64 0, i64 %idxprom2
|
|
%1 = load i64, ptr %arrayidx3, align 8
|
|
%add4 = add nsw i64 %1, %0
|
|
ret i64 %add4
|
|
}
|
|
|
|
define signext i32 @test_sh2add_uw(ptr %p, i32 signext %x, i32 signext %y) {
|
|
; RV32I-LABEL: test_sh2add_uw:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: slli a1, a1, 2
|
|
; RV32I-NEXT: slli a2, a2, 2
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a0, a2
|
|
; RV32I-NEXT: lw a1, 1200(a1)
|
|
; RV32I-NEXT: lw a0, 1200(a0)
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_sh2add_uw:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: slli a1, a1, 32
|
|
; RV64I-NEXT: slli a2, a2, 32
|
|
; RV64I-NEXT: srli a1, a1, 30
|
|
; RV64I-NEXT: srli a2, a2, 30
|
|
; RV64I-NEXT: add a1, a0, a1
|
|
; RV64I-NEXT: add a0, a0, a2
|
|
; RV64I-NEXT: lw a1, 1200(a1)
|
|
; RV64I-NEXT: lw a0, 1200(a0)
|
|
; RV64I-NEXT: addw a0, a0, a1
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_sh2add_uw:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: sh2add a1, a1, a0
|
|
; RV32ZBA-NEXT: sh2add a0, a2, a0
|
|
; RV32ZBA-NEXT: lw a1, 1200(a1)
|
|
; RV32ZBA-NEXT: lw a0, 1200(a0)
|
|
; RV32ZBA-NEXT: add a0, a0, a1
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_sh2add_uw:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: sh2add.uw a1, a1, a0
|
|
; RV64ZBA-NEXT: sh2add.uw a0, a2, a0
|
|
; RV64ZBA-NEXT: lw a1, 1200(a1)
|
|
; RV64ZBA-NEXT: lw a0, 1200(a0)
|
|
; RV64ZBA-NEXT: addw a0, a0, a1
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%c = getelementptr inbounds nuw i8, ptr %p, i64 1200
|
|
%idxprom = zext i32 %x to i64
|
|
%arrayidx = getelementptr inbounds nuw [100 x i32], ptr %c, i64 0, i64 %idxprom
|
|
%0 = load i32, ptr %arrayidx, align 4
|
|
%idxprom2 = zext i32 %y to i64
|
|
%arrayidx3 = getelementptr inbounds nuw [100 x i32], ptr %c, i64 0, i64 %idxprom2
|
|
%1 = load i32, ptr %arrayidx3, align 4
|
|
%add4 = add nsw i32 %1, %0
|
|
ret i32 %add4
|
|
}
|
|
|
|
define signext i16 @test_sh1add_uw(ptr %p, i32 signext %x, i32 signext %y) {
|
|
; RV32I-LABEL: test_sh1add_uw:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: slli a1, a1, 1
|
|
; RV32I-NEXT: slli a2, a2, 1
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a2, a0
|
|
; RV32I-NEXT: lh a1, 1600(a1)
|
|
; RV32I-NEXT: lh a0, 1620(a0)
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: slli a0, a0, 16
|
|
; RV32I-NEXT: srai a0, a0, 16
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_sh1add_uw:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: slli a1, a1, 32
|
|
; RV64I-NEXT: addi a2, a2, 10
|
|
; RV64I-NEXT: srli a1, a1, 31
|
|
; RV64I-NEXT: slli a2, a2, 32
|
|
; RV64I-NEXT: add a1, a0, a1
|
|
; RV64I-NEXT: srli a2, a2, 31
|
|
; RV64I-NEXT: add a0, a0, a2
|
|
; RV64I-NEXT: lh a1, 1600(a1)
|
|
; RV64I-NEXT: lh a0, 1600(a0)
|
|
; RV64I-NEXT: add a0, a0, a1
|
|
; RV64I-NEXT: slli a0, a0, 48
|
|
; RV64I-NEXT: srai a0, a0, 48
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_sh1add_uw:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: sh1add a1, a1, a0
|
|
; RV32ZBA-NEXT: sh1add a0, a2, a0
|
|
; RV32ZBA-NEXT: lh a1, 1600(a1)
|
|
; RV32ZBA-NEXT: lh a0, 1620(a0)
|
|
; RV32ZBA-NEXT: add a0, a0, a1
|
|
; RV32ZBA-NEXT: slli a0, a0, 16
|
|
; RV32ZBA-NEXT: srai a0, a0, 16
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_sh1add_uw:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: sh1add.uw a1, a1, a0
|
|
; RV64ZBA-NEXT: addi a2, a2, 10
|
|
; RV64ZBA-NEXT: sh1add.uw a0, a2, a0
|
|
; RV64ZBA-NEXT: lh a1, 1600(a1)
|
|
; RV64ZBA-NEXT: lh a0, 1600(a0)
|
|
; RV64ZBA-NEXT: add a0, a0, a1
|
|
; RV64ZBA-NEXT: slli a0, a0, 48
|
|
; RV64ZBA-NEXT: srai a0, a0, 48
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%d = getelementptr inbounds nuw i8, ptr %p, i64 1600
|
|
%idxprom = zext i32 %x to i64
|
|
%arrayidx = getelementptr inbounds nuw [100 x i16], ptr %d, i64 0, i64 %idxprom
|
|
%0 = load i16, ptr %arrayidx, align 2
|
|
%add = add i32 %y, 10
|
|
%idxprom2 = zext i32 %add to i64
|
|
%arrayidx3 = getelementptr inbounds nuw [100 x i16], ptr %d, i64 0, i64 %idxprom2
|
|
%1 = load i16, ptr %arrayidx3, align 2
|
|
%add5 = add i16 %1, %0
|
|
ret i16 %add5
|
|
}
|
|
|
|
define zeroext i8 @test_add_uw(ptr %p, i32 signext %x, i32 signext %y) {
|
|
; RV32I-LABEL: test_add_uw:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a0, a2
|
|
; RV32I-NEXT: lbu a1, 1800(a1)
|
|
; RV32I-NEXT: lbu a0, 1800(a0)
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: andi a0, a0, 255
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_add_uw:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: slli a1, a1, 32
|
|
; RV64I-NEXT: slli a2, a2, 32
|
|
; RV64I-NEXT: srli a1, a1, 32
|
|
; RV64I-NEXT: srli a2, a2, 32
|
|
; RV64I-NEXT: add a1, a0, a1
|
|
; RV64I-NEXT: add a0, a0, a2
|
|
; RV64I-NEXT: lbu a1, 1800(a1)
|
|
; RV64I-NEXT: lbu a0, 1800(a0)
|
|
; RV64I-NEXT: add a0, a0, a1
|
|
; RV64I-NEXT: andi a0, a0, 255
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_add_uw:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: add a1, a0, a1
|
|
; RV32ZBA-NEXT: add a0, a0, a2
|
|
; RV32ZBA-NEXT: lbu a1, 1800(a1)
|
|
; RV32ZBA-NEXT: lbu a0, 1800(a0)
|
|
; RV32ZBA-NEXT: add a0, a0, a1
|
|
; RV32ZBA-NEXT: andi a0, a0, 255
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_add_uw:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: add.uw a1, a1, a0
|
|
; RV64ZBA-NEXT: add.uw a0, a2, a0
|
|
; RV64ZBA-NEXT: lbu a1, 1800(a1)
|
|
; RV64ZBA-NEXT: lbu a0, 1800(a0)
|
|
; RV64ZBA-NEXT: add a0, a0, a1
|
|
; RV64ZBA-NEXT: andi a0, a0, 255
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%e = getelementptr inbounds nuw i8, ptr %p, i64 1800
|
|
%idxprom = zext i32 %x to i64
|
|
%arrayidx = getelementptr inbounds nuw [1000 x i8], ptr %e, i64 0, i64 %idxprom
|
|
%0 = load i8, ptr %arrayidx, align 1
|
|
%idxprom2 = zext i32 %y to i64
|
|
%arrayidx3 = getelementptr inbounds nuw [1000 x i8], ptr %e, i64 0, i64 %idxprom2
|
|
%1 = load i8, ptr %arrayidx3, align 1
|
|
%add5 = add i8 %1, %0
|
|
ret i8 %add5
|
|
}
|
|
|
|
; The addi is part of the index and used with 2 different scales.
|
|
define signext i32 @test_scaled_index_addi(ptr %p, iXLen %x) {
|
|
; RV32I-LABEL: test_scaled_index_addi:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: slli a2, a1, 2
|
|
; RV32I-NEXT: slli a1, a1, 1
|
|
; RV32I-NEXT: add a2, a0, a2
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: lw a1, 1196(a2)
|
|
; RV32I-NEXT: lh a0, 1598(a0)
|
|
; RV32I-NEXT: add a0, a1, a0
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_scaled_index_addi:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: slli a2, a1, 2
|
|
; RV64I-NEXT: slli a1, a1, 1
|
|
; RV64I-NEXT: add a2, a0, a2
|
|
; RV64I-NEXT: add a0, a0, a1
|
|
; RV64I-NEXT: lw a1, 1196(a2)
|
|
; RV64I-NEXT: lh a0, 1598(a0)
|
|
; RV64I-NEXT: addw a0, a1, a0
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_scaled_index_addi:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: sh2add a2, a1, a0
|
|
; RV32ZBA-NEXT: sh1add a0, a1, a0
|
|
; RV32ZBA-NEXT: lw a1, 1196(a2)
|
|
; RV32ZBA-NEXT: lh a0, 1598(a0)
|
|
; RV32ZBA-NEXT: add a0, a1, a0
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_scaled_index_addi:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: sh2add a2, a1, a0
|
|
; RV64ZBA-NEXT: sh1add a0, a1, a0
|
|
; RV64ZBA-NEXT: lw a1, 1196(a2)
|
|
; RV64ZBA-NEXT: lh a0, 1598(a0)
|
|
; RV64ZBA-NEXT: addw a0, a1, a0
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%c = getelementptr inbounds nuw i8, ptr %p, i64 1200
|
|
%sub = add iXLen %x, -1
|
|
%arrayidx = getelementptr inbounds nuw [100 x i32], ptr %c, i64 0, iXLen %sub
|
|
%0 = load i32, ptr %arrayidx, align 4
|
|
%d = getelementptr inbounds nuw i8, ptr %p, i64 1600
|
|
%arrayidx2 = getelementptr inbounds nuw [100 x i16], ptr %d, i64 0, iXLen %sub
|
|
%1 = load i16, ptr %arrayidx2, align 2
|
|
%conv = sext i16 %1 to i32
|
|
%add = add nsw i32 %0, %conv
|
|
ret i32 %add
|
|
}
|
|
|
|
; Offset is a pair of addis. We can fold one of them.
|
|
define signext i32 @test_medium_offset(ptr %p, iXLen %x, iXLen %y) {
|
|
; RV32I-LABEL: test_medium_offset:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: addi a0, a0, 2047
|
|
; RV32I-NEXT: slli a1, a1, 2
|
|
; RV32I-NEXT: slli a2, a2, 2
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a2, a0
|
|
; RV32I-NEXT: lw a1, 753(a1)
|
|
; RV32I-NEXT: lw a0, 793(a0)
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_medium_offset:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: addi a0, a0, 2047
|
|
; RV64I-NEXT: slli a1, a1, 2
|
|
; RV64I-NEXT: slli a2, a2, 2
|
|
; RV64I-NEXT: add a1, a0, a1
|
|
; RV64I-NEXT: add a0, a2, a0
|
|
; RV64I-NEXT: lw a1, 753(a1)
|
|
; RV64I-NEXT: lw a0, 793(a0)
|
|
; RV64I-NEXT: addw a0, a0, a1
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_medium_offset:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: addi a0, a0, 2047
|
|
; RV32ZBA-NEXT: sh2add a1, a1, a0
|
|
; RV32ZBA-NEXT: sh2add a0, a2, a0
|
|
; RV32ZBA-NEXT: lw a1, 753(a1)
|
|
; RV32ZBA-NEXT: lw a0, 793(a0)
|
|
; RV32ZBA-NEXT: add a0, a0, a1
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_medium_offset:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: addi a0, a0, 2047
|
|
; RV64ZBA-NEXT: sh2add a1, a1, a0
|
|
; RV64ZBA-NEXT: sh2add a0, a2, a0
|
|
; RV64ZBA-NEXT: lw a1, 753(a1)
|
|
; RV64ZBA-NEXT: lw a0, 793(a0)
|
|
; RV64ZBA-NEXT: addw a0, a0, a1
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%f = getelementptr inbounds nuw i8, ptr %p, i64 2800
|
|
%arrayidx = getelementptr inbounds nuw [1000 x i32], ptr %f, i64 0, iXLen %x
|
|
%0 = load i32, ptr %arrayidx, align 4
|
|
%add = add iXLen %y, 10
|
|
%arrayidx2 = getelementptr inbounds nuw [1000 x i32], ptr %f, i64 0, iXLen %add
|
|
%1 = load i32, ptr %arrayidx2, align 4
|
|
%add3 = add nsw i32 %1, %0
|
|
ret i32 %add3
|
|
}
|
|
|
|
; Offset is a lui+addiw. We can't fold this on RV64.
|
|
define signext i32 @test_large_offset(ptr %p, iXLen %x, iXLen %y) {
|
|
; RV32I-LABEL: test_large_offset:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: lui a3, 2
|
|
; RV32I-NEXT: slli a1, a1, 2
|
|
; RV32I-NEXT: slli a2, a2, 2
|
|
; RV32I-NEXT: add a0, a0, a3
|
|
; RV32I-NEXT: add a1, a0, a1
|
|
; RV32I-NEXT: add a0, a2, a0
|
|
; RV32I-NEXT: lw a1, -1392(a1)
|
|
; RV32I-NEXT: lw a0, -1352(a0)
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_large_offset:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: lui a3, 2
|
|
; RV64I-NEXT: slli a1, a1, 2
|
|
; RV64I-NEXT: slli a2, a2, 2
|
|
; RV64I-NEXT: addiw a3, a3, -1392
|
|
; RV64I-NEXT: add a0, a0, a3
|
|
; RV64I-NEXT: add a1, a0, a1
|
|
; RV64I-NEXT: add a0, a2, a0
|
|
; RV64I-NEXT: lw a1, 0(a1)
|
|
; RV64I-NEXT: lw a0, 40(a0)
|
|
; RV64I-NEXT: addw a0, a0, a1
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_large_offset:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: li a3, 1700
|
|
; RV32ZBA-NEXT: sh2add a0, a3, a0
|
|
; RV32ZBA-NEXT: sh2add a1, a1, a0
|
|
; RV32ZBA-NEXT: sh2add a0, a2, a0
|
|
; RV32ZBA-NEXT: lw a1, 0(a1)
|
|
; RV32ZBA-NEXT: lw a0, 40(a0)
|
|
; RV32ZBA-NEXT: add a0, a0, a1
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_large_offset:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: li a3, 1700
|
|
; RV64ZBA-NEXT: sh2add a0, a3, a0
|
|
; RV64ZBA-NEXT: sh2add a1, a1, a0
|
|
; RV64ZBA-NEXT: sh2add a0, a2, a0
|
|
; RV64ZBA-NEXT: lw a1, 0(a1)
|
|
; RV64ZBA-NEXT: lw a0, 40(a0)
|
|
; RV64ZBA-NEXT: addw a0, a0, a1
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%g = getelementptr inbounds nuw i8, ptr %p, i64 6800
|
|
%arrayidx = getelementptr inbounds nuw [200 x i32], ptr %g, i64 0, iXLen %x
|
|
%0 = load i32, ptr %arrayidx, align 4
|
|
%add = add iXLen %y, 10
|
|
%arrayidx2 = getelementptr inbounds nuw [200 x i32], ptr %g, i64 0, iXLen %add
|
|
%1 = load i32, ptr %arrayidx2, align 4
|
|
%add3 = add nsw i32 %1, %0
|
|
ret i32 %add3
|
|
}
|
|
|
|
; After folding we can CSE the sh2add
|
|
define signext i32 @test_cse(ptr %p, iXLen %x) {
|
|
; RV32I-LABEL: test_cse:
|
|
; RV32I: # %bb.0: # %entry
|
|
; RV32I-NEXT: slli a1, a1, 2
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: lw a1, 1200(a0)
|
|
; RV32I-NEXT: addi a0, a0, 2047
|
|
; RV32I-NEXT: lw a0, 753(a0)
|
|
; RV32I-NEXT: add a0, a0, a1
|
|
; RV32I-NEXT: ret
|
|
;
|
|
; RV64I-LABEL: test_cse:
|
|
; RV64I: # %bb.0: # %entry
|
|
; RV64I-NEXT: slli a1, a1, 2
|
|
; RV64I-NEXT: add a0, a0, a1
|
|
; RV64I-NEXT: lw a1, 1200(a0)
|
|
; RV64I-NEXT: addi a0, a0, 2047
|
|
; RV64I-NEXT: lw a0, 753(a0)
|
|
; RV64I-NEXT: addw a0, a0, a1
|
|
; RV64I-NEXT: ret
|
|
;
|
|
; RV32ZBA-LABEL: test_cse:
|
|
; RV32ZBA: # %bb.0: # %entry
|
|
; RV32ZBA-NEXT: sh2add a0, a1, a0
|
|
; RV32ZBA-NEXT: lw a1, 1200(a0)
|
|
; RV32ZBA-NEXT: addi a0, a0, 2047
|
|
; RV32ZBA-NEXT: lw a0, 753(a0)
|
|
; RV32ZBA-NEXT: add a0, a0, a1
|
|
; RV32ZBA-NEXT: ret
|
|
;
|
|
; RV64ZBA-LABEL: test_cse:
|
|
; RV64ZBA: # %bb.0: # %entry
|
|
; RV64ZBA-NEXT: sh2add a0, a1, a0
|
|
; RV64ZBA-NEXT: lw a1, 1200(a0)
|
|
; RV64ZBA-NEXT: addi a0, a0, 2047
|
|
; RV64ZBA-NEXT: lw a0, 753(a0)
|
|
; RV64ZBA-NEXT: addw a0, a0, a1
|
|
; RV64ZBA-NEXT: ret
|
|
entry:
|
|
%c = getelementptr inbounds nuw i8, ptr %p, i64 1200
|
|
%arrayidx = getelementptr inbounds nuw [100 x i32], ptr %c, i64 0, iXLen %x
|
|
%0 = load i32, ptr %arrayidx, align 4
|
|
%f = getelementptr inbounds nuw i8, ptr %p, i64 2800
|
|
%arrayidx1 = getelementptr inbounds nuw [1000 x i32], ptr %f, i64 0, iXLen %x
|
|
%1 = load i32, ptr %arrayidx1, align 4
|
|
%add = add nsw i32 %1, %0
|
|
ret i32 %add
|
|
}
|
|
|
|
define zeroext i8 @test_optsize(ptr %p, iXLen %x, iXLen %y) optsize {
|
|
; CHECK-LABEL: test_optsize:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: addi a0, a0, 1800
|
|
; CHECK-NEXT: add a1, a0, a1
|
|
; CHECK-NEXT: add a0, a2, a0
|
|
; CHECK-NEXT: lbu a1, 0(a1)
|
|
; CHECK-NEXT: lbu a0, 10(a0)
|
|
; CHECK-NEXT: add a0, a0, a1
|
|
; CHECK-NEXT: andi a0, a0, 255
|
|
; CHECK-NEXT: ret
|
|
;
|
|
; ZBA-LABEL: test_optsize:
|
|
; ZBA: # %bb.0: # %entry
|
|
; ZBA-NEXT: addi a0, a0, 1800
|
|
; ZBA-NEXT: add a1, a0, a1
|
|
; ZBA-NEXT: add a0, a2, a0
|
|
; ZBA-NEXT: lbu a1, 0(a1)
|
|
; ZBA-NEXT: lbu a0, 10(a0)
|
|
; ZBA-NEXT: add a0, a0, a1
|
|
; ZBA-NEXT: andi a0, a0, 255
|
|
; ZBA-NEXT: ret
|
|
entry:
|
|
%e = getelementptr inbounds nuw i8, ptr %p, i64 1800
|
|
%arrayidx = getelementptr inbounds nuw [1000 x i8], ptr %e, i64 0, iXLen %x
|
|
%0 = load i8, ptr %arrayidx, align 1
|
|
%add = add iXLen %y, 10
|
|
%arrayidx2 = getelementptr inbounds nuw [1000 x i8], ptr %e, i64 0, iXLen %add
|
|
%1 = load i8, ptr %arrayidx2, align 1
|
|
%add4 = add i8 %1, %0
|
|
ret i8 %add4
|
|
}
|
|
|
|
define zeroext i8 @test_minsize(ptr %p, iXLen %x, iXLen %y) minsize {
|
|
; CHECK-LABEL: test_minsize:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: addi a0, a0, 1800
|
|
; CHECK-NEXT: add a1, a0, a1
|
|
; CHECK-NEXT: add a0, a2, a0
|
|
; CHECK-NEXT: lbu a1, 0(a1)
|
|
; CHECK-NEXT: lbu a0, 10(a0)
|
|
; CHECK-NEXT: add a0, a0, a1
|
|
; CHECK-NEXT: andi a0, a0, 255
|
|
; CHECK-NEXT: ret
|
|
;
|
|
; ZBA-LABEL: test_minsize:
|
|
; ZBA: # %bb.0: # %entry
|
|
; ZBA-NEXT: addi a0, a0, 1800
|
|
; ZBA-NEXT: add a1, a0, a1
|
|
; ZBA-NEXT: add a0, a2, a0
|
|
; ZBA-NEXT: lbu a1, 0(a1)
|
|
; ZBA-NEXT: lbu a0, 10(a0)
|
|
; ZBA-NEXT: add a0, a0, a1
|
|
; ZBA-NEXT: andi a0, a0, 255
|
|
; ZBA-NEXT: ret
|
|
entry:
|
|
%e = getelementptr inbounds nuw i8, ptr %p, i64 1800
|
|
%arrayidx = getelementptr inbounds nuw [1000 x i8], ptr %e, i64 0, iXLen %x
|
|
%0 = load i8, ptr %arrayidx, align 1
|
|
%add = add iXLen %y, 10
|
|
%arrayidx2 = getelementptr inbounds nuw [1000 x i8], ptr %e, i64 0, iXLen %add
|
|
%1 = load i8, ptr %arrayidx2, align 1
|
|
%add4 = add i8 %1, %0
|
|
ret i8 %add4
|
|
}
|