caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
082251bba4effea7f60191c6cbddacb3705c07db
clang-p2996/llvm/test/CodeGen/AArch64/split-vector-insert.ll
Ricardo Jesus 15fbdc2b96 [AArch64][SVE] Lower unpredicated loads/stores as LDR/STR. (#127837) 
Currently, given:
```cpp
svuint8_t foo(uint8_t *x) {
  return svld1(svptrue_b8(), x);
}
```
We generate:
```gas
foo:
  ptrue   p0.b
  ld1b    { z0.b }, p0/z, [x0]
  ret
```
However, on little-endian and with unaligned memory accesses allowed, we
could instead be using LDR as follows:
```gas
foo:
  ldr     z0, [x0]
  ret
```

The second form avoids the predicate dependency.
Likewise for other types and stores.
2025-02-26 13:56:35 +00:00

201 lines
8.3 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 4
; RUN: llc < %s -debug-only=legalize-types 2>&1 | FileCheck %s --check-prefix=CHECK-LEGALIZATION
; RUN: llc < %s | FileCheck %s
; REQUIRES: asserts
target triple = "aarch64-unknown-linux-gnu"
attributes #0 = {"target-features"="+sve" uwtable}
declare <vscale x 2 x i64> @llvm.vector.insert.nxv2i64.v8i64(<vscale x 2 x i64>, <8 x i64>, i64)
declare <vscale x 2 x double> @llvm.vector.insert.nxv2f64.v8f64(<vscale x 2 x double>, <8 x double>, i64)
define <vscale x 2 x i64> @test_nxv2i64_v8i64(<vscale x 2 x i64> %a, <8 x i64> %b) #0 {
; CHECK-LEGALIZATION-LABEL: test_nxv2i64_v8i64:
; CHECK-LEGALIZATION: // %bb.0:
; CHECK-LEGALIZATION-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
; CHECK-LEGALIZATION-NEXT: .cfi_def_cfa_offset 16
; CHECK-LEGALIZATION-NEXT: .cfi_offset w29, -16
; CHECK-LEGALIZATION-NEXT: addvl sp, sp, #-3
; CHECK-LEGALIZATION-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 24 * VG
; CHECK-LEGALIZATION-NEXT: cntd x8
; CHECK-LEGALIZATION-NEXT: ptrue p0.d, vl2
; CHECK-LEGALIZATION-NEXT: mov w9, #2 // =0x2
; CHECK-LEGALIZATION-NEXT: sub x8, x8, #2
; CHECK-LEGALIZATION-NEXT: // kill: def $q1 killed $q1 def $z1
; CHECK-LEGALIZATION-NEXT: mov x10, sp
; CHECK-LEGALIZATION-NEXT: cmp x8, #2
; CHECK-LEGALIZATION-NEXT: mov z0.d, p0/m, z1.d
; CHECK-LEGALIZATION-NEXT: csel x9, x8, x9, lo
; CHECK-LEGALIZATION-NEXT: cmp x8, #4
; CHECK-LEGALIZATION-NEXT: lsl x9, x9, #3
; CHECK-LEGALIZATION-NEXT: str z0, [sp]
; CHECK-LEGALIZATION-NEXT: str q2, [x10, x9]
; CHECK-LEGALIZATION-NEXT: mov w9, #4 // =0x4
; CHECK-LEGALIZATION-NEXT: addvl x10, sp, #1
; CHECK-LEGALIZATION-NEXT: ldr z0, [sp]
; CHECK-LEGALIZATION-NEXT: csel x9, x8, x9, lo
; CHECK-LEGALIZATION-NEXT: cmp x8, #6
; CHECK-LEGALIZATION-NEXT: lsl x9, x9, #3
; CHECK-LEGALIZATION-NEXT: str z0, [sp, #1, mul vl]
; CHECK-LEGALIZATION-NEXT: str q3, [x10, x9]
; CHECK-LEGALIZATION-NEXT: mov w9, #6 // =0x6
; CHECK-LEGALIZATION-NEXT: ldr z0, [sp, #1, mul vl]
; CHECK-LEGALIZATION-NEXT: csel x8, x8, x9, lo
; CHECK-LEGALIZATION-NEXT: addvl x9, sp, #2
; CHECK-LEGALIZATION-NEXT: lsl x8, x8, #3
; CHECK-LEGALIZATION-NEXT: str z0, [sp, #2, mul vl]
; CHECK-LEGALIZATION-NEXT: str q4, [x9, x8]
; CHECK-LEGALIZATION-NEXT: ldr z0, [sp, #2, mul vl]
; CHECK-LEGALIZATION-NEXT: addvl sp, sp, #3
; CHECK-LEGALIZATION-NEXT: .cfi_def_cfa wsp, 16
; CHECK-LEGALIZATION-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
; CHECK-LEGALIZATION-NEXT: .cfi_def_cfa_offset 0
; CHECK-LEGALIZATION-NEXT: .cfi_restore w29
; CHECK-LEGALIZATION-NEXT: ret
;
; CHECK-LABEL: test_nxv2i64_v8i64:
; CHECK: // %bb.0:
; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 16
; CHECK-NEXT: .cfi_offset w29, -16
; CHECK-NEXT: addvl sp, sp, #-3
; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 24 * VG
; CHECK-NEXT: cntd x8
; CHECK-NEXT: ptrue p0.d, vl2
; CHECK-NEXT: mov w9, #2 // =0x2
; CHECK-NEXT: sub x8, x8, #2
; CHECK-NEXT: // kill: def $q1 killed $q1 def $z1
; CHECK-NEXT: mov x10, sp
; CHECK-NEXT: cmp x8, #2
; CHECK-NEXT: mov z0.d, p0/m, z1.d
; CHECK-NEXT: csel x9, x8, x9, lo
; CHECK-NEXT: cmp x8, #4
; CHECK-NEXT: lsl x9, x9, #3
; CHECK-NEXT: str z0, [sp]
; CHECK-NEXT: str q2, [x10, x9]
; CHECK-NEXT: mov w9, #4 // =0x4
; CHECK-NEXT: addvl x10, sp, #1
; CHECK-NEXT: ldr z0, [sp]
; CHECK-NEXT: csel x9, x8, x9, lo
; CHECK-NEXT: cmp x8, #6
; CHECK-NEXT: lsl x9, x9, #3
; CHECK-NEXT: str z0, [sp, #1, mul vl]
; CHECK-NEXT: str q3, [x10, x9]
; CHECK-NEXT: mov w9, #6 // =0x6
; CHECK-NEXT: ldr z0, [sp, #1, mul vl]
; CHECK-NEXT: csel x8, x8, x9, lo
; CHECK-NEXT: addvl x9, sp, #2
; CHECK-NEXT: lsl x8, x8, #3
; CHECK-NEXT: str z0, [sp, #2, mul vl]
; CHECK-NEXT: str q4, [x9, x8]
; CHECK-NEXT: ldr z0, [sp, #2, mul vl]
; CHECK-NEXT: addvl sp, sp, #3
; CHECK-NEXT: .cfi_def_cfa wsp, 16
; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
; CHECK-NEXT: .cfi_def_cfa_offset 0
; CHECK-NEXT: .cfi_restore w29
; CHECK-NEXT: ret
%r = call <vscale x 2 x i64> @llvm.vector.insert.nxv2i64.v8i64(<vscale x 2 x i64> %a, <8 x i64> %b, i64 0)
ret <vscale x 2 x i64> %r
}
define <vscale x 2 x double> @test_nxv2f64_v8f64(<vscale x 2 x double> %a, <8 x double> %b) #0 {
; CHECK-LEGALIZATION-LABEL: test_nxv2f64_v8f64:
; CHECK-LEGALIZATION: // %bb.0:
; CHECK-LEGALIZATION-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
; CHECK-LEGALIZATION-NEXT: .cfi_def_cfa_offset 16
; CHECK-LEGALIZATION-NEXT: .cfi_offset w29, -16
; CHECK-LEGALIZATION-NEXT: addvl sp, sp, #-3
; CHECK-LEGALIZATION-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 24 * VG
; CHECK-LEGALIZATION-NEXT: cntd x8
; CHECK-LEGALIZATION-NEXT: ptrue p0.d, vl2
; CHECK-LEGALIZATION-NEXT: mov w9, #2 // =0x2
; CHECK-LEGALIZATION-NEXT: sub x8, x8, #2
; CHECK-LEGALIZATION-NEXT: // kill: def $q1 killed $q1 def $z1
; CHECK-LEGALIZATION-NEXT: mov x10, sp
; CHECK-LEGALIZATION-NEXT: cmp x8, #2
; CHECK-LEGALIZATION-NEXT: mov z0.d, p0/m, z1.d
; CHECK-LEGALIZATION-NEXT: csel x9, x8, x9, lo
; CHECK-LEGALIZATION-NEXT: cmp x8, #4
; CHECK-LEGALIZATION-NEXT: lsl x9, x9, #3
; CHECK-LEGALIZATION-NEXT: str z0, [sp]
; CHECK-LEGALIZATION-NEXT: str q2, [x10, x9]
; CHECK-LEGALIZATION-NEXT: mov w9, #4 // =0x4
; CHECK-LEGALIZATION-NEXT: addvl x10, sp, #1
; CHECK-LEGALIZATION-NEXT: ldr z0, [sp]
; CHECK-LEGALIZATION-NEXT: csel x9, x8, x9, lo
; CHECK-LEGALIZATION-NEXT: cmp x8, #6
; CHECK-LEGALIZATION-NEXT: lsl x9, x9, #3
; CHECK-LEGALIZATION-NEXT: str z0, [sp, #1, mul vl]
; CHECK-LEGALIZATION-NEXT: str q3, [x10, x9]
; CHECK-LEGALIZATION-NEXT: mov w9, #6 // =0x6
; CHECK-LEGALIZATION-NEXT: ldr z0, [sp, #1, mul vl]
; CHECK-LEGALIZATION-NEXT: csel x8, x8, x9, lo
; CHECK-LEGALIZATION-NEXT: addvl x9, sp, #2
; CHECK-LEGALIZATION-NEXT: lsl x8, x8, #3
; CHECK-LEGALIZATION-NEXT: str z0, [sp, #2, mul vl]
; CHECK-LEGALIZATION-NEXT: str q4, [x9, x8]
; CHECK-LEGALIZATION-NEXT: ldr z0, [sp, #2, mul vl]
; CHECK-LEGALIZATION-NEXT: addvl sp, sp, #3
; CHECK-LEGALIZATION-NEXT: .cfi_def_cfa wsp, 16
; CHECK-LEGALIZATION-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
; CHECK-LEGALIZATION-NEXT: .cfi_def_cfa_offset 0
; CHECK-LEGALIZATION-NEXT: .cfi_restore w29
; CHECK-LEGALIZATION-NEXT: ret
;
; CHECK-LABEL: test_nxv2f64_v8f64:
; CHECK: // %bb.0:
; CHECK-NEXT: str x29, [sp, #-16]! // 8-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 16
; CHECK-NEXT: .cfi_offset w29, -16
; CHECK-NEXT: addvl sp, sp, #-3
; CHECK-NEXT: .cfi_escape 0x0f, 0x0c, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0x18, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 24 * VG
; CHECK-NEXT: cntd x8
; CHECK-NEXT: ptrue p0.d, vl2
; CHECK-NEXT: mov w9, #2 // =0x2
; CHECK-NEXT: sub x8, x8, #2
; CHECK-NEXT: // kill: def $q1 killed $q1 def $z1
; CHECK-NEXT: mov x10, sp
; CHECK-NEXT: cmp x8, #2
; CHECK-NEXT: mov z0.d, p0/m, z1.d
; CHECK-NEXT: csel x9, x8, x9, lo
; CHECK-NEXT: cmp x8, #4
; CHECK-NEXT: lsl x9, x9, #3
; CHECK-NEXT: str z0, [sp]
; CHECK-NEXT: str q2, [x10, x9]
; CHECK-NEXT: mov w9, #4 // =0x4
; CHECK-NEXT: addvl x10, sp, #1
; CHECK-NEXT: ldr z0, [sp]
; CHECK-NEXT: csel x9, x8, x9, lo
; CHECK-NEXT: cmp x8, #6
; CHECK-NEXT: lsl x9, x9, #3
; CHECK-NEXT: str z0, [sp, #1, mul vl]
; CHECK-NEXT: str q3, [x10, x9]
; CHECK-NEXT: mov w9, #6 // =0x6
; CHECK-NEXT: ldr z0, [sp, #1, mul vl]
; CHECK-NEXT: csel x8, x8, x9, lo
; CHECK-NEXT: addvl x9, sp, #2
; CHECK-NEXT: lsl x8, x8, #3
; CHECK-NEXT: str z0, [sp, #2, mul vl]
; CHECK-NEXT: str q4, [x9, x8]
; CHECK-NEXT: ldr z0, [sp, #2, mul vl]
; CHECK-NEXT: addvl sp, sp, #3
; CHECK-NEXT: .cfi_def_cfa wsp, 16
; CHECK-NEXT: ldr x29, [sp], #16 // 8-byte Folded Reload
; CHECK-NEXT: .cfi_def_cfa_offset 0
; CHECK-NEXT: .cfi_restore w29
; CHECK-NEXT: ret
%r = call <vscale x 2 x double> @llvm.vector.insert.nxv2f64.v8f64(<vscale x 2 x double> %a, <8 x double> %b, i64 0)
ret <vscale x 2 x double> %r
}
Reference in New Issue View Git Blame Copy Permalink