| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5 |
| ; RUN: llc < %s -mtriple=aarch64-unknown-unknown -global-isel=0 | FileCheck %s --check-prefixes=CHECK,CHECK-SD |
| ; RUN: llc < %s -mtriple=aarch64-unknown-unknown -global-isel=1 | FileCheck %s --check-prefixes=CHECK,CHECK-GI |
| |
| ; These tests are taken from the combine-sdiv.ll in X86. |
| define i32 @combine_sdiv_by_one(i32 %x) { |
| ; CHECK-LABEL: combine_sdiv_by_one: |
| ; CHECK: // %bb.0: |
| ; CHECK-NEXT: ret |
| %1 = sdiv i32 %x, 1 |
| ret i32 %1 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_by_one(<4 x i32> %x) { |
| ; CHECK-LABEL: combine_vec_sdiv_by_one: |
| ; CHECK: // %bb.0: |
| ; CHECK-NEXT: ret |
| %1 = sdiv <4 x i32> %x, <i32 1, i32 1, i32 1, i32 1> |
| ret <4 x i32> %1 |
| } |
| |
| define i32 @combine_sdiv_by_negone(i32 %x) { |
| ; CHECK-LABEL: combine_sdiv_by_negone: |
| ; CHECK: // %bb.0: |
| ; CHECK-NEXT: neg w0, w0 |
| ; CHECK-NEXT: ret |
| %1 = sdiv i32 %x, -1 |
| ret i32 %1 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_by_negone(<4 x i32> %x) { |
| ; CHECK-LABEL: combine_vec_sdiv_by_negone: |
| ; CHECK: // %bb.0: |
| ; CHECK-NEXT: neg v0.4s, v0.4s |
| ; CHECK-NEXT: ret |
| %1 = sdiv <4 x i32> %x, <i32 -1, i32 -1, i32 -1, i32 -1> |
| ret <4 x i32> %1 |
| } |
| |
| define i32 @combine_sdiv_by_minsigned(i32 %x) { |
| ; CHECK-SD-LABEL: combine_sdiv_by_minsigned: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: mov w8, #-2147483648 // =0x80000000 |
| ; CHECK-SD-NEXT: cmp w0, w8 |
| ; CHECK-SD-NEXT: cset w0, eq |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_sdiv_by_minsigned: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: asr w8, w0, #31 |
| ; CHECK-GI-NEXT: add w8, w0, w8, lsr #1 |
| ; CHECK-GI-NEXT: neg w0, w8, asr #31 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i32 %x, -2147483648 |
| ret i32 %1 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_by_minsigned(<4 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_minsigned: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: movi v1.4s, #128, lsl #24 |
| ; CHECK-SD-NEXT: movi v2.4s, #1 |
| ; CHECK-SD-NEXT: cmeq v0.4s, v0.4s, v1.4s |
| ; CHECK-SD-NEXT: and v0.16b, v0.16b, v2.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_minsigned: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sshr v1.4s, v0.4s, #31 |
| ; CHECK-GI-NEXT: usra v0.4s, v1.4s, #1 |
| ; CHECK-GI-NEXT: sshr v0.4s, v0.4s, #31 |
| ; CHECK-GI-NEXT: neg v0.4s, v0.4s |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <4 x i32> %x, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648> |
| ret <4 x i32> %1 |
| } |
| |
| define i32 @combine_sdiv_zero(i32 %x) { |
| ; CHECK-LABEL: combine_sdiv_zero: |
| ; CHECK: // %bb.0: |
| ; CHECK-NEXT: mov w0, wzr |
| ; CHECK-NEXT: ret |
| %1 = sdiv i32 0, %x |
| ret i32 %1 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_zero(<4 x i32> %x) { |
| ; CHECK-LABEL: combine_vec_sdiv_zero: |
| ; CHECK: // %bb.0: |
| ; CHECK-NEXT: movi v0.2d, #0000000000000000 |
| ; CHECK-NEXT: ret |
| %1 = sdiv <4 x i32> zeroinitializer, %x |
| ret <4 x i32> %1 |
| } |
| |
| define i32 @combine_sdiv_dupe(i32 %x) { |
| ; CHECK-SD-LABEL: combine_sdiv_dupe: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: mov w0, #1 // =0x1 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_sdiv_dupe: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sdiv w0, w0, w0 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i32 %x, %x |
| ret i32 %1 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_dupe(<4 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_dupe: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: movi v0.4s, #1 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_dupe: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: fmov w8, s0 |
| ; CHECK-GI-NEXT: mov w9, v0.s[1] |
| ; CHECK-GI-NEXT: mov w10, v0.s[2] |
| ; CHECK-GI-NEXT: mov w11, v0.s[3] |
| ; CHECK-GI-NEXT: sdiv w8, w8, w8 |
| ; CHECK-GI-NEXT: sdiv w9, w9, w9 |
| ; CHECK-GI-NEXT: fmov s0, w8 |
| ; CHECK-GI-NEXT: sdiv w10, w10, w10 |
| ; CHECK-GI-NEXT: mov v0.s[1], w9 |
| ; CHECK-GI-NEXT: sdiv w8, w11, w11 |
| ; CHECK-GI-NEXT: mov v0.s[2], w10 |
| ; CHECK-GI-NEXT: mov v0.s[3], w8 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <4 x i32> %x, %x |
| ret <4 x i32> %1 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_by_pos0(<4 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pos0: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: movi v1.2d, #0x0000ff000000ff |
| ; CHECK-SD-NEXT: and v0.16b, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: ushr v0.4s, v0.4s, #2 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pos0: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: movi v1.2d, #0x0000ff000000ff |
| ; CHECK-GI-NEXT: and v0.16b, v0.16b, v1.16b |
| ; CHECK-GI-NEXT: sshr v0.4s, v0.4s, #2 |
| ; CHECK-GI-NEXT: ret |
| %1 = and <4 x i32> %x, <i32 255, i32 255, i32 255, i32 255> |
| %2 = sdiv <4 x i32> %1, <i32 4, i32 4, i32 4, i32 4> |
| ret <4 x i32> %2 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_by_pos1(<4 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pos1: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: movi v1.2d, #0x0000ff000000ff |
| ; CHECK-SD-NEXT: adrp x8, .LCPI11_0 |
| ; CHECK-SD-NEXT: and v0.16b, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI11_0] |
| ; CHECK-SD-NEXT: ushl v0.4s, v0.4s, v1.4s |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pos1: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #1 // =0x1 |
| ; CHECK-GI-NEXT: mov w9, #0 // =0x0 |
| ; CHECK-GI-NEXT: movi v2.2d, #0x0000ff000000ff |
| ; CHECK-GI-NEXT: fmov s1, w8 |
| ; CHECK-GI-NEXT: adrp x8, .LCPI11_0 |
| ; CHECK-GI-NEXT: ldr q3, [x8, :lo12:.LCPI11_0] |
| ; CHECK-GI-NEXT: mov v1.s[1], w9 |
| ; CHECK-GI-NEXT: and v0.16b, v0.16b, v2.16b |
| ; CHECK-GI-NEXT: neg v2.4s, v3.4s |
| ; CHECK-GI-NEXT: sshl v2.4s, v0.4s, v2.4s |
| ; CHECK-GI-NEXT: mov v1.s[2], w9 |
| ; CHECK-GI-NEXT: mov v1.s[3], w9 |
| ; CHECK-GI-NEXT: shl v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: sshr v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: bif v0.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = and <4 x i32> %x, <i32 255, i32 255, i32 255, i32 255> |
| %2 = sdiv <4 x i32> %1, <i32 1, i32 4, i32 8, i32 16> |
| ret <4 x i32> %2 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_by_pow2a(<4 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2a: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: cmlt v1.4s, v0.4s, #0 |
| ; CHECK-SD-NEXT: usra v0.4s, v1.4s, #30 |
| ; CHECK-SD-NEXT: sshr v0.4s, v0.4s, #2 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2a: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sshr v1.4s, v0.4s, #31 |
| ; CHECK-GI-NEXT: usra v0.4s, v1.4s, #30 |
| ; CHECK-GI-NEXT: sshr v0.4s, v0.4s, #2 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <4 x i32> %x, <i32 4, i32 4, i32 4, i32 4> |
| ret <4 x i32> %1 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_by_pow2a_neg(<4 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2a_neg: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: cmlt v1.4s, v0.4s, #0 |
| ; CHECK-SD-NEXT: usra v0.4s, v1.4s, #30 |
| ; CHECK-SD-NEXT: sshr v0.4s, v0.4s, #2 |
| ; CHECK-SD-NEXT: neg v0.4s, v0.4s |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2a_neg: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sshr v1.4s, v0.4s, #31 |
| ; CHECK-GI-NEXT: usra v0.4s, v1.4s, #30 |
| ; CHECK-GI-NEXT: sshr v0.4s, v0.4s, #2 |
| ; CHECK-GI-NEXT: neg v0.4s, v0.4s |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <4 x i32> %x, <i32 -4, i32 -4, i32 -4, i32 -4> |
| ret <4 x i32> %1 |
| } |
| |
| define <16 x i8> @combine_vec_sdiv_by_pow2b_v16i8(<16 x i8> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v16i8: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI14_0 |
| ; CHECK-SD-NEXT: cmlt v1.16b, v0.16b, #0 |
| ; CHECK-SD-NEXT: movi v3.2d, #0x000000000000ff |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI14_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI14_1 |
| ; CHECK-SD-NEXT: movi v4.2d, #0xffffffffffffff00 |
| ; CHECK-SD-NEXT: ushl v1.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI14_1] |
| ; CHECK-SD-NEXT: add v1.16b, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: and v0.16b, v0.16b, v3.16b |
| ; CHECK-SD-NEXT: sshl v1.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: and v1.16b, v1.16b, v4.16b |
| ; CHECK-SD-NEXT: orr v0.16b, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v16i8: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI14_1 |
| ; CHECK-GI-NEXT: sshr v2.16b, v0.16b, #7 |
| ; CHECK-GI-NEXT: adrp x9, .LCPI14_0 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI14_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI14_2 |
| ; CHECK-GI-NEXT: ldr q3, [x9, :lo12:.LCPI14_0] |
| ; CHECK-GI-NEXT: neg v1.16b, v1.16b |
| ; CHECK-GI-NEXT: shl v3.16b, v3.16b, #7 |
| ; CHECK-GI-NEXT: ushl v1.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI14_2] |
| ; CHECK-GI-NEXT: neg v2.16b, v2.16b |
| ; CHECK-GI-NEXT: add v1.16b, v0.16b, v1.16b |
| ; CHECK-GI-NEXT: sshl v1.16b, v1.16b, v2.16b |
| ; CHECK-GI-NEXT: sshr v2.16b, v3.16b, #7 |
| ; CHECK-GI-NEXT: bif v0.16b, v1.16b, v2.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <16 x i8> %x, <i8 1, i8 4, i8 2, i8 16, i8 8, i8 32, i8 64, i8 2, i8 1, i8 4, i8 2, i8 16, i8 8, i8 32, i8 64, i8 2> |
| ret <16 x i8> %1 |
| } |
| |
| define <8 x i16> @combine_vec_sdiv_by_pow2b_v8i16(<8 x i16> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v8i16: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI15_1 |
| ; CHECK-SD-NEXT: cmlt v1.8h, v0.8h, #0 |
| ; CHECK-SD-NEXT: adrp x9, .LCPI15_3 |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI15_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI15_2 |
| ; CHECK-SD-NEXT: ldr q3, [x9, :lo12:.LCPI15_3] |
| ; CHECK-SD-NEXT: ushl v1.8h, v1.8h, v2.8h |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI15_2] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI15_0 |
| ; CHECK-SD-NEXT: add v1.8h, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: sshl v1.8h, v1.8h, v2.8h |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI15_0] |
| ; CHECK-SD-NEXT: and v0.16b, v0.16b, v2.16b |
| ; CHECK-SD-NEXT: and v1.16b, v1.16b, v3.16b |
| ; CHECK-SD-NEXT: orr v0.16b, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v8i16: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI15_1 |
| ; CHECK-GI-NEXT: sshr v2.8h, v0.8h, #15 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI15_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI15_0 |
| ; CHECK-GI-NEXT: ldr d3, [x8, :lo12:.LCPI15_0] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI15_2 |
| ; CHECK-GI-NEXT: neg v1.8h, v1.8h |
| ; CHECK-GI-NEXT: ushl v1.8h, v2.8h, v1.8h |
| ; CHECK-GI-NEXT: ushll v2.8h, v3.8b, #0 |
| ; CHECK-GI-NEXT: ldr q3, [x8, :lo12:.LCPI15_2] |
| ; CHECK-GI-NEXT: neg v3.8h, v3.8h |
| ; CHECK-GI-NEXT: add v1.8h, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: shl v2.8h, v2.8h, #15 |
| ; CHECK-GI-NEXT: sshl v1.8h, v1.8h, v3.8h |
| ; CHECK-GI-NEXT: sshr v2.8h, v2.8h, #15 |
| ; CHECK-GI-NEXT: bif v0.16b, v1.16b, v2.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <8 x i16> %x, <i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2> |
| ret <8 x i16> %1 |
| } |
| |
| define <16 x i16> @combine_vec_sdiv_by_pow2b_v16i16(<16 x i16> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v16i16: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI16_1 |
| ; CHECK-SD-NEXT: cmlt v2.8h, v0.8h, #0 |
| ; CHECK-SD-NEXT: cmlt v3.8h, v1.8h, #0 |
| ; CHECK-SD-NEXT: ldr q4, [x8, :lo12:.LCPI16_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI16_2 |
| ; CHECK-SD-NEXT: ushl v2.8h, v2.8h, v4.8h |
| ; CHECK-SD-NEXT: ushl v3.8h, v3.8h, v4.8h |
| ; CHECK-SD-NEXT: ldr q4, [x8, :lo12:.LCPI16_2] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI16_0 |
| ; CHECK-SD-NEXT: ldr q5, [x8, :lo12:.LCPI16_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI16_3 |
| ; CHECK-SD-NEXT: add v2.8h, v0.8h, v2.8h |
| ; CHECK-SD-NEXT: add v3.8h, v1.8h, v3.8h |
| ; CHECK-SD-NEXT: and v0.16b, v0.16b, v5.16b |
| ; CHECK-SD-NEXT: and v1.16b, v1.16b, v5.16b |
| ; CHECK-SD-NEXT: sshl v2.8h, v2.8h, v4.8h |
| ; CHECK-SD-NEXT: sshl v3.8h, v3.8h, v4.8h |
| ; CHECK-SD-NEXT: ldr q4, [x8, :lo12:.LCPI16_3] |
| ; CHECK-SD-NEXT: and v2.16b, v2.16b, v4.16b |
| ; CHECK-SD-NEXT: and v3.16b, v3.16b, v4.16b |
| ; CHECK-SD-NEXT: orr v0.16b, v0.16b, v2.16b |
| ; CHECK-SD-NEXT: orr v1.16b, v1.16b, v3.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v16i16: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI16_1 |
| ; CHECK-GI-NEXT: sshr v3.8h, v0.8h, #15 |
| ; CHECK-GI-NEXT: sshr v4.8h, v1.8h, #15 |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI16_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI16_0 |
| ; CHECK-GI-NEXT: ldr d5, [x8, :lo12:.LCPI16_0] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI16_2 |
| ; CHECK-GI-NEXT: neg v2.8h, v2.8h |
| ; CHECK-GI-NEXT: ushll v5.8h, v5.8b, #0 |
| ; CHECK-GI-NEXT: ushl v3.8h, v3.8h, v2.8h |
| ; CHECK-GI-NEXT: ushl v2.8h, v4.8h, v2.8h |
| ; CHECK-GI-NEXT: ldr q4, [x8, :lo12:.LCPI16_2] |
| ; CHECK-GI-NEXT: shl v5.8h, v5.8h, #15 |
| ; CHECK-GI-NEXT: neg v4.8h, v4.8h |
| ; CHECK-GI-NEXT: add v3.8h, v0.8h, v3.8h |
| ; CHECK-GI-NEXT: add v2.8h, v1.8h, v2.8h |
| ; CHECK-GI-NEXT: sshl v3.8h, v3.8h, v4.8h |
| ; CHECK-GI-NEXT: sshl v2.8h, v2.8h, v4.8h |
| ; CHECK-GI-NEXT: sshr v4.8h, v5.8h, #15 |
| ; CHECK-GI-NEXT: bif v0.16b, v3.16b, v4.16b |
| ; CHECK-GI-NEXT: bif v1.16b, v2.16b, v4.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <16 x i16> %x, <i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2, i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2> |
| ret <16 x i16> %1 |
| } |
| |
| define <32 x i16> @combine_vec_sdiv_by_pow2b_v32i16(<32 x i16> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v32i16: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI17_1 |
| ; CHECK-SD-NEXT: cmlt v4.8h, v0.8h, #0 |
| ; CHECK-SD-NEXT: cmlt v5.8h, v1.8h, #0 |
| ; CHECK-SD-NEXT: cmlt v7.8h, v2.8h, #0 |
| ; CHECK-SD-NEXT: cmlt v16.8h, v3.8h, #0 |
| ; CHECK-SD-NEXT: ldr q6, [x8, :lo12:.LCPI17_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI17_2 |
| ; CHECK-SD-NEXT: ushl v4.8h, v4.8h, v6.8h |
| ; CHECK-SD-NEXT: ushl v5.8h, v5.8h, v6.8h |
| ; CHECK-SD-NEXT: ushl v7.8h, v7.8h, v6.8h |
| ; CHECK-SD-NEXT: ushl v6.8h, v16.8h, v6.8h |
| ; CHECK-SD-NEXT: ldr q16, [x8, :lo12:.LCPI17_2] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI17_0 |
| ; CHECK-SD-NEXT: add v4.8h, v0.8h, v4.8h |
| ; CHECK-SD-NEXT: add v5.8h, v1.8h, v5.8h |
| ; CHECK-SD-NEXT: ldr q17, [x8, :lo12:.LCPI17_0] |
| ; CHECK-SD-NEXT: add v7.8h, v2.8h, v7.8h |
| ; CHECK-SD-NEXT: add v6.8h, v3.8h, v6.8h |
| ; CHECK-SD-NEXT: adrp x8, .LCPI17_3 |
| ; CHECK-SD-NEXT: and v0.16b, v0.16b, v17.16b |
| ; CHECK-SD-NEXT: and v1.16b, v1.16b, v17.16b |
| ; CHECK-SD-NEXT: and v2.16b, v2.16b, v17.16b |
| ; CHECK-SD-NEXT: sshl v4.8h, v4.8h, v16.8h |
| ; CHECK-SD-NEXT: sshl v5.8h, v5.8h, v16.8h |
| ; CHECK-SD-NEXT: and v3.16b, v3.16b, v17.16b |
| ; CHECK-SD-NEXT: sshl v7.8h, v7.8h, v16.8h |
| ; CHECK-SD-NEXT: sshl v6.8h, v6.8h, v16.8h |
| ; CHECK-SD-NEXT: ldr q16, [x8, :lo12:.LCPI17_3] |
| ; CHECK-SD-NEXT: and v4.16b, v4.16b, v16.16b |
| ; CHECK-SD-NEXT: and v5.16b, v5.16b, v16.16b |
| ; CHECK-SD-NEXT: and v7.16b, v7.16b, v16.16b |
| ; CHECK-SD-NEXT: and v6.16b, v6.16b, v16.16b |
| ; CHECK-SD-NEXT: orr v0.16b, v0.16b, v4.16b |
| ; CHECK-SD-NEXT: orr v1.16b, v1.16b, v5.16b |
| ; CHECK-SD-NEXT: orr v2.16b, v2.16b, v7.16b |
| ; CHECK-SD-NEXT: orr v3.16b, v3.16b, v6.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v32i16: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI17_1 |
| ; CHECK-GI-NEXT: sshr v5.8h, v0.8h, #15 |
| ; CHECK-GI-NEXT: sshr v6.8h, v1.8h, #15 |
| ; CHECK-GI-NEXT: ldr q4, [x8, :lo12:.LCPI17_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI17_0 |
| ; CHECK-GI-NEXT: sshr v7.8h, v2.8h, #15 |
| ; CHECK-GI-NEXT: sshr v16.8h, v3.8h, #15 |
| ; CHECK-GI-NEXT: ldr d17, [x8, :lo12:.LCPI17_0] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI17_2 |
| ; CHECK-GI-NEXT: neg v4.8h, v4.8h |
| ; CHECK-GI-NEXT: ushll v17.8h, v17.8b, #0 |
| ; CHECK-GI-NEXT: ushl v5.8h, v5.8h, v4.8h |
| ; CHECK-GI-NEXT: ushl v6.8h, v6.8h, v4.8h |
| ; CHECK-GI-NEXT: ushl v7.8h, v7.8h, v4.8h |
| ; CHECK-GI-NEXT: ushl v4.8h, v16.8h, v4.8h |
| ; CHECK-GI-NEXT: ldr q16, [x8, :lo12:.LCPI17_2] |
| ; CHECK-GI-NEXT: shl v17.8h, v17.8h, #15 |
| ; CHECK-GI-NEXT: neg v16.8h, v16.8h |
| ; CHECK-GI-NEXT: add v5.8h, v0.8h, v5.8h |
| ; CHECK-GI-NEXT: add v6.8h, v1.8h, v6.8h |
| ; CHECK-GI-NEXT: add v7.8h, v2.8h, v7.8h |
| ; CHECK-GI-NEXT: add v4.8h, v3.8h, v4.8h |
| ; CHECK-GI-NEXT: sshr v17.8h, v17.8h, #15 |
| ; CHECK-GI-NEXT: sshl v5.8h, v5.8h, v16.8h |
| ; CHECK-GI-NEXT: sshl v6.8h, v6.8h, v16.8h |
| ; CHECK-GI-NEXT: sshl v7.8h, v7.8h, v16.8h |
| ; CHECK-GI-NEXT: sshl v4.8h, v4.8h, v16.8h |
| ; CHECK-GI-NEXT: bif v0.16b, v5.16b, v17.16b |
| ; CHECK-GI-NEXT: bif v1.16b, v6.16b, v17.16b |
| ; CHECK-GI-NEXT: bif v2.16b, v7.16b, v17.16b |
| ; CHECK-GI-NEXT: bif v3.16b, v4.16b, v17.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <32 x i16> %x, <i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2, i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2, i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2, i16 1, i16 4, i16 2, i16 16, i16 8, i16 32, i16 64, i16 2> |
| ret <32 x i16> %1 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_by_pow2b_v4i32(<4 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v4i32: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI18_0 |
| ; CHECK-SD-NEXT: cmlt v1.4s, v0.4s, #0 |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI18_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI18_1 |
| ; CHECK-SD-NEXT: ushl v1.4s, v1.4s, v2.4s |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI18_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI18_2 |
| ; CHECK-SD-NEXT: add v1.4s, v0.4s, v1.4s |
| ; CHECK-SD-NEXT: sshl v1.4s, v1.4s, v2.4s |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI18_2] |
| ; CHECK-SD-NEXT: bif v0.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v4i32: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #1 // =0x1 |
| ; CHECK-GI-NEXT: mov w9, #0 // =0x0 |
| ; CHECK-GI-NEXT: sshr v3.4s, v0.4s, #31 |
| ; CHECK-GI-NEXT: fmov s1, w8 |
| ; CHECK-GI-NEXT: adrp x8, .LCPI18_0 |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI18_0] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI18_1 |
| ; CHECK-GI-NEXT: mov v1.s[1], w9 |
| ; CHECK-GI-NEXT: neg v2.4s, v2.4s |
| ; CHECK-GI-NEXT: ushl v2.4s, v3.4s, v2.4s |
| ; CHECK-GI-NEXT: ldr q3, [x8, :lo12:.LCPI18_1] |
| ; CHECK-GI-NEXT: mov v1.s[2], w9 |
| ; CHECK-GI-NEXT: neg v3.4s, v3.4s |
| ; CHECK-GI-NEXT: add v2.4s, v0.4s, v2.4s |
| ; CHECK-GI-NEXT: mov v1.s[3], w9 |
| ; CHECK-GI-NEXT: sshl v2.4s, v2.4s, v3.4s |
| ; CHECK-GI-NEXT: shl v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: sshr v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: bif v0.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <4 x i32> %x, <i32 1, i32 4, i32 8, i32 16> |
| ret <4 x i32> %1 |
| } |
| |
| define <8 x i32> @combine_vec_sdiv_by_pow2b_v8i32(<8 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v8i32: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI19_0 |
| ; CHECK-SD-NEXT: cmlt v2.4s, v0.4s, #0 |
| ; CHECK-SD-NEXT: cmlt v3.4s, v1.4s, #0 |
| ; CHECK-SD-NEXT: ldr q4, [x8, :lo12:.LCPI19_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI19_1 |
| ; CHECK-SD-NEXT: ushl v2.4s, v2.4s, v4.4s |
| ; CHECK-SD-NEXT: ushl v3.4s, v3.4s, v4.4s |
| ; CHECK-SD-NEXT: ldr q4, [x8, :lo12:.LCPI19_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI19_2 |
| ; CHECK-SD-NEXT: add v2.4s, v0.4s, v2.4s |
| ; CHECK-SD-NEXT: add v3.4s, v1.4s, v3.4s |
| ; CHECK-SD-NEXT: sshl v2.4s, v2.4s, v4.4s |
| ; CHECK-SD-NEXT: sshl v3.4s, v3.4s, v4.4s |
| ; CHECK-SD-NEXT: ldr q4, [x8, :lo12:.LCPI19_2] |
| ; CHECK-SD-NEXT: bif v0.16b, v2.16b, v4.16b |
| ; CHECK-SD-NEXT: bif v1.16b, v3.16b, v4.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v8i32: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #1 // =0x1 |
| ; CHECK-GI-NEXT: mov w9, #0 // =0x0 |
| ; CHECK-GI-NEXT: sshr v4.4s, v0.4s, #31 |
| ; CHECK-GI-NEXT: fmov s2, w8 |
| ; CHECK-GI-NEXT: adrp x8, .LCPI19_0 |
| ; CHECK-GI-NEXT: sshr v5.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: ldr q3, [x8, :lo12:.LCPI19_0] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI19_1 |
| ; CHECK-GI-NEXT: mov v2.h[1], w9 |
| ; CHECK-GI-NEXT: neg v3.4s, v3.4s |
| ; CHECK-GI-NEXT: ushl v4.4s, v4.4s, v3.4s |
| ; CHECK-GI-NEXT: ushl v3.4s, v5.4s, v3.4s |
| ; CHECK-GI-NEXT: ldr q5, [x8, :lo12:.LCPI19_1] |
| ; CHECK-GI-NEXT: mov v2.h[2], w9 |
| ; CHECK-GI-NEXT: neg v5.4s, v5.4s |
| ; CHECK-GI-NEXT: add v4.4s, v0.4s, v4.4s |
| ; CHECK-GI-NEXT: add v3.4s, v1.4s, v3.4s |
| ; CHECK-GI-NEXT: mov v2.h[3], w9 |
| ; CHECK-GI-NEXT: sshl v4.4s, v4.4s, v5.4s |
| ; CHECK-GI-NEXT: sshl v3.4s, v3.4s, v5.4s |
| ; CHECK-GI-NEXT: ushll v2.4s, v2.4h, #0 |
| ; CHECK-GI-NEXT: shl v2.4s, v2.4s, #31 |
| ; CHECK-GI-NEXT: sshr v2.4s, v2.4s, #31 |
| ; CHECK-GI-NEXT: bif v0.16b, v4.16b, v2.16b |
| ; CHECK-GI-NEXT: bif v1.16b, v3.16b, v2.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <8 x i32> %x, <i32 1, i32 4, i32 8, i32 16, i32 1, i32 4, i32 8, i32 16> |
| ret <8 x i32> %1 |
| } |
| |
| define <16 x i32> @combine_vec_sdiv_by_pow2b_v16i32(<16 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v16i32: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI20_0 |
| ; CHECK-SD-NEXT: cmlt v4.4s, v0.4s, #0 |
| ; CHECK-SD-NEXT: cmlt v5.4s, v1.4s, #0 |
| ; CHECK-SD-NEXT: cmlt v7.4s, v2.4s, #0 |
| ; CHECK-SD-NEXT: cmlt v16.4s, v3.4s, #0 |
| ; CHECK-SD-NEXT: ldr q6, [x8, :lo12:.LCPI20_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI20_1 |
| ; CHECK-SD-NEXT: ushl v4.4s, v4.4s, v6.4s |
| ; CHECK-SD-NEXT: ushl v5.4s, v5.4s, v6.4s |
| ; CHECK-SD-NEXT: ushl v7.4s, v7.4s, v6.4s |
| ; CHECK-SD-NEXT: ushl v6.4s, v16.4s, v6.4s |
| ; CHECK-SD-NEXT: ldr q16, [x8, :lo12:.LCPI20_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI20_2 |
| ; CHECK-SD-NEXT: add v4.4s, v0.4s, v4.4s |
| ; CHECK-SD-NEXT: add v5.4s, v1.4s, v5.4s |
| ; CHECK-SD-NEXT: add v7.4s, v2.4s, v7.4s |
| ; CHECK-SD-NEXT: add v6.4s, v3.4s, v6.4s |
| ; CHECK-SD-NEXT: sshl v4.4s, v4.4s, v16.4s |
| ; CHECK-SD-NEXT: sshl v5.4s, v5.4s, v16.4s |
| ; CHECK-SD-NEXT: sshl v7.4s, v7.4s, v16.4s |
| ; CHECK-SD-NEXT: sshl v6.4s, v6.4s, v16.4s |
| ; CHECK-SD-NEXT: ldr q16, [x8, :lo12:.LCPI20_2] |
| ; CHECK-SD-NEXT: bif v0.16b, v4.16b, v16.16b |
| ; CHECK-SD-NEXT: bif v1.16b, v5.16b, v16.16b |
| ; CHECK-SD-NEXT: bif v2.16b, v7.16b, v16.16b |
| ; CHECK-SD-NEXT: bif v3.16b, v6.16b, v16.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v16i32: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #1 // =0x1 |
| ; CHECK-GI-NEXT: mov w9, #0 // =0x0 |
| ; CHECK-GI-NEXT: sshr v6.4s, v0.4s, #31 |
| ; CHECK-GI-NEXT: fmov s4, w8 |
| ; CHECK-GI-NEXT: adrp x8, .LCPI20_0 |
| ; CHECK-GI-NEXT: sshr v7.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: ldr q5, [x8, :lo12:.LCPI20_0] |
| ; CHECK-GI-NEXT: sshr v16.4s, v2.4s, #31 |
| ; CHECK-GI-NEXT: sshr v17.4s, v3.4s, #31 |
| ; CHECK-GI-NEXT: adrp x8, .LCPI20_1 |
| ; CHECK-GI-NEXT: mov v4.h[1], w9 |
| ; CHECK-GI-NEXT: neg v5.4s, v5.4s |
| ; CHECK-GI-NEXT: ushl v6.4s, v6.4s, v5.4s |
| ; CHECK-GI-NEXT: ushl v7.4s, v7.4s, v5.4s |
| ; CHECK-GI-NEXT: ushl v16.4s, v16.4s, v5.4s |
| ; CHECK-GI-NEXT: mov v4.h[2], w9 |
| ; CHECK-GI-NEXT: ushl v5.4s, v17.4s, v5.4s |
| ; CHECK-GI-NEXT: ldr q17, [x8, :lo12:.LCPI20_1] |
| ; CHECK-GI-NEXT: neg v17.4s, v17.4s |
| ; CHECK-GI-NEXT: add v6.4s, v0.4s, v6.4s |
| ; CHECK-GI-NEXT: add v7.4s, v1.4s, v7.4s |
| ; CHECK-GI-NEXT: add v16.4s, v2.4s, v16.4s |
| ; CHECK-GI-NEXT: add v5.4s, v3.4s, v5.4s |
| ; CHECK-GI-NEXT: mov v4.h[3], w9 |
| ; CHECK-GI-NEXT: sshl v6.4s, v6.4s, v17.4s |
| ; CHECK-GI-NEXT: sshl v7.4s, v7.4s, v17.4s |
| ; CHECK-GI-NEXT: sshl v16.4s, v16.4s, v17.4s |
| ; CHECK-GI-NEXT: sshl v5.4s, v5.4s, v17.4s |
| ; CHECK-GI-NEXT: ushll v4.4s, v4.4h, #0 |
| ; CHECK-GI-NEXT: shl v4.4s, v4.4s, #31 |
| ; CHECK-GI-NEXT: sshr v4.4s, v4.4s, #31 |
| ; CHECK-GI-NEXT: bif v0.16b, v6.16b, v4.16b |
| ; CHECK-GI-NEXT: bif v1.16b, v7.16b, v4.16b |
| ; CHECK-GI-NEXT: bif v2.16b, v16.16b, v4.16b |
| ; CHECK-GI-NEXT: bif v3.16b, v5.16b, v4.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <16 x i32> %x, <i32 1, i32 4, i32 8, i32 16, i32 1, i32 4, i32 8, i32 16, i32 1, i32 4, i32 8, i32 16, i32 1, i32 4, i32 8, i32 16> |
| ret <16 x i32> %1 |
| } |
| |
| define <2 x i64> @combine_vec_sdiv_by_pow2b_v2i64(<2 x i64> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v2i64: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI21_0 |
| ; CHECK-SD-NEXT: cmlt v1.2d, v0.2d, #0 |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI21_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI21_1 |
| ; CHECK-SD-NEXT: ushl v1.2d, v1.2d, v2.2d |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI21_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI21_2 |
| ; CHECK-SD-NEXT: add v1.2d, v0.2d, v1.2d |
| ; CHECK-SD-NEXT: sshl v1.2d, v1.2d, v2.2d |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI21_2] |
| ; CHECK-SD-NEXT: bif v0.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v2i64: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI21_1 |
| ; CHECK-GI-NEXT: sshr v2.2d, v0.2d, #63 |
| ; CHECK-GI-NEXT: adrp x9, .LCPI21_0 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI21_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI21_2 |
| ; CHECK-GI-NEXT: ldr q3, [x9, :lo12:.LCPI21_0] |
| ; CHECK-GI-NEXT: neg v1.2d, v1.2d |
| ; CHECK-GI-NEXT: shl v3.2d, v3.2d, #63 |
| ; CHECK-GI-NEXT: ushl v1.2d, v2.2d, v1.2d |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI21_2] |
| ; CHECK-GI-NEXT: neg v2.2d, v2.2d |
| ; CHECK-GI-NEXT: add v1.2d, v0.2d, v1.2d |
| ; CHECK-GI-NEXT: sshl v1.2d, v1.2d, v2.2d |
| ; CHECK-GI-NEXT: sshr v2.2d, v3.2d, #63 |
| ; CHECK-GI-NEXT: bif v0.16b, v1.16b, v2.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <2 x i64> %x, <i64 1, i64 4> |
| ret <2 x i64> %1 |
| } |
| |
| define <4 x i64> @combine_vec_sdiv_by_pow2b_v4i64(<4 x i64> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v4i64: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI22_0 |
| ; CHECK-SD-NEXT: cmlt v2.2d, v0.2d, #0 |
| ; CHECK-SD-NEXT: ldr q3, [x8, :lo12:.LCPI22_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI22_3 |
| ; CHECK-SD-NEXT: ldr q4, [x8, :lo12:.LCPI22_3] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI22_1 |
| ; CHECK-SD-NEXT: ushl v2.2d, v2.2d, v3.2d |
| ; CHECK-SD-NEXT: cmlt v3.2d, v1.2d, #0 |
| ; CHECK-SD-NEXT: add v2.2d, v0.2d, v2.2d |
| ; CHECK-SD-NEXT: ushl v3.2d, v3.2d, v4.2d |
| ; CHECK-SD-NEXT: ldr q4, [x8, :lo12:.LCPI22_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI22_2 |
| ; CHECK-SD-NEXT: sshl v2.2d, v2.2d, v4.2d |
| ; CHECK-SD-NEXT: ldr q4, [x8, :lo12:.LCPI22_2] |
| ; CHECK-SD-NEXT: add v1.2d, v1.2d, v3.2d |
| ; CHECK-SD-NEXT: adrp x8, .LCPI22_4 |
| ; CHECK-SD-NEXT: ldr q3, [x8, :lo12:.LCPI22_4] |
| ; CHECK-SD-NEXT: bif v0.16b, v2.16b, v4.16b |
| ; CHECK-SD-NEXT: sshl v1.2d, v1.2d, v3.2d |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v4i64: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI22_2 |
| ; CHECK-GI-NEXT: sshr v3.2d, v0.2d, #63 |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI22_2] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI22_1 |
| ; CHECK-GI-NEXT: ldr q4, [x8, :lo12:.LCPI22_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI22_4 |
| ; CHECK-GI-NEXT: neg v2.2d, v2.2d |
| ; CHECK-GI-NEXT: ldr q5, [x8, :lo12:.LCPI22_4] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI22_0 |
| ; CHECK-GI-NEXT: neg v4.2d, v4.2d |
| ; CHECK-GI-NEXT: ldr q6, [x8, :lo12:.LCPI22_0] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI22_3 |
| ; CHECK-GI-NEXT: neg v5.2d, v5.2d |
| ; CHECK-GI-NEXT: ushl v2.2d, v3.2d, v2.2d |
| ; CHECK-GI-NEXT: sshr v3.2d, v1.2d, #63 |
| ; CHECK-GI-NEXT: shl v6.2d, v6.2d, #63 |
| ; CHECK-GI-NEXT: add v2.2d, v0.2d, v2.2d |
| ; CHECK-GI-NEXT: ushl v3.2d, v3.2d, v4.2d |
| ; CHECK-GI-NEXT: ldr q4, [x8, :lo12:.LCPI22_3] |
| ; CHECK-GI-NEXT: sshl v2.2d, v2.2d, v5.2d |
| ; CHECK-GI-NEXT: sshr v5.2d, v6.2d, #63 |
| ; CHECK-GI-NEXT: add v1.2d, v1.2d, v3.2d |
| ; CHECK-GI-NEXT: neg v3.2d, v4.2d |
| ; CHECK-GI-NEXT: bif v0.16b, v2.16b, v5.16b |
| ; CHECK-GI-NEXT: sshl v1.2d, v1.2d, v3.2d |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <4 x i64> %x, <i64 1, i64 4, i64 8, i64 16> |
| ret <4 x i64> %1 |
| } |
| |
| define <8 x i64> @combine_vec_sdiv_by_pow2b_v8i64(<8 x i64> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_v8i64: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI23_0 |
| ; CHECK-SD-NEXT: cmlt v4.2d, v0.2d, #0 |
| ; CHECK-SD-NEXT: cmlt v6.2d, v2.2d, #0 |
| ; CHECK-SD-NEXT: ldr q5, [x8, :lo12:.LCPI23_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI23_3 |
| ; CHECK-SD-NEXT: cmlt v7.2d, v3.2d, #0 |
| ; CHECK-SD-NEXT: ldr q16, [x8, :lo12:.LCPI23_3] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI23_1 |
| ; CHECK-SD-NEXT: ushl v4.2d, v4.2d, v5.2d |
| ; CHECK-SD-NEXT: ushl v5.2d, v6.2d, v5.2d |
| ; CHECK-SD-NEXT: cmlt v6.2d, v1.2d, #0 |
| ; CHECK-SD-NEXT: ldr q17, [x8, :lo12:.LCPI23_1] |
| ; CHECK-SD-NEXT: ushl v7.2d, v7.2d, v16.2d |
| ; CHECK-SD-NEXT: adrp x8, .LCPI23_2 |
| ; CHECK-SD-NEXT: add v4.2d, v0.2d, v4.2d |
| ; CHECK-SD-NEXT: add v5.2d, v2.2d, v5.2d |
| ; CHECK-SD-NEXT: ushl v6.2d, v6.2d, v16.2d |
| ; CHECK-SD-NEXT: ldr q16, [x8, :lo12:.LCPI23_2] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI23_4 |
| ; CHECK-SD-NEXT: add v3.2d, v3.2d, v7.2d |
| ; CHECK-SD-NEXT: sshl v4.2d, v4.2d, v17.2d |
| ; CHECK-SD-NEXT: sshl v5.2d, v5.2d, v17.2d |
| ; CHECK-SD-NEXT: add v1.2d, v1.2d, v6.2d |
| ; CHECK-SD-NEXT: ldr q6, [x8, :lo12:.LCPI23_4] |
| ; CHECK-SD-NEXT: bif v0.16b, v4.16b, v16.16b |
| ; CHECK-SD-NEXT: bif v2.16b, v5.16b, v16.16b |
| ; CHECK-SD-NEXT: sshl v1.2d, v1.2d, v6.2d |
| ; CHECK-SD-NEXT: sshl v3.2d, v3.2d, v6.2d |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_v8i64: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #1 // =0x1 |
| ; CHECK-GI-NEXT: mov w9, #0 // =0x0 |
| ; CHECK-GI-NEXT: sshr v7.2d, v0.2d, #63 |
| ; CHECK-GI-NEXT: fmov s4, w8 |
| ; CHECK-GI-NEXT: adrp x8, .LCPI23_1 |
| ; CHECK-GI-NEXT: sshr v16.2d, v1.2d, #63 |
| ; CHECK-GI-NEXT: ldr q5, [x8, :lo12:.LCPI23_1] |
| ; CHECK-GI-NEXT: sshr v17.2d, v2.2d, #63 |
| ; CHECK-GI-NEXT: sshr v18.2d, v3.2d, #63 |
| ; CHECK-GI-NEXT: adrp x8, .LCPI23_3 |
| ; CHECK-GI-NEXT: mov v4.h[1], w9 |
| ; CHECK-GI-NEXT: neg v5.2d, v5.2d |
| ; CHECK-GI-NEXT: ldr q19, [x8, :lo12:.LCPI23_3] |
| ; CHECK-GI-NEXT: neg v19.2d, v19.2d |
| ; CHECK-GI-NEXT: ushl v7.2d, v7.2d, v5.2d |
| ; CHECK-GI-NEXT: ushl v5.2d, v17.2d, v5.2d |
| ; CHECK-GI-NEXT: mov v4.h[2], w9 |
| ; CHECK-GI-NEXT: add v7.2d, v0.2d, v7.2d |
| ; CHECK-GI-NEXT: add v5.2d, v2.2d, v5.2d |
| ; CHECK-GI-NEXT: mov v4.h[3], w9 |
| ; CHECK-GI-NEXT: adrp x9, .LCPI23_0 |
| ; CHECK-GI-NEXT: ldr q6, [x9, :lo12:.LCPI23_0] |
| ; CHECK-GI-NEXT: adrp x9, .LCPI23_2 |
| ; CHECK-GI-NEXT: sshl v7.2d, v7.2d, v19.2d |
| ; CHECK-GI-NEXT: ldr q20, [x9, :lo12:.LCPI23_2] |
| ; CHECK-GI-NEXT: sshl v5.2d, v5.2d, v19.2d |
| ; CHECK-GI-NEXT: neg v6.2d, v6.2d |
| ; CHECK-GI-NEXT: ushll v4.4s, v4.4h, #0 |
| ; CHECK-GI-NEXT: neg v20.2d, v20.2d |
| ; CHECK-GI-NEXT: ushl v16.2d, v16.2d, v6.2d |
| ; CHECK-GI-NEXT: ushl v6.2d, v18.2d, v6.2d |
| ; CHECK-GI-NEXT: ushll v17.2d, v4.2s, #0 |
| ; CHECK-GI-NEXT: ushll2 v18.2d, v4.4s, #0 |
| ; CHECK-GI-NEXT: ushll v4.2d, v4.2s, #0 |
| ; CHECK-GI-NEXT: add v16.2d, v1.2d, v16.2d |
| ; CHECK-GI-NEXT: add v6.2d, v3.2d, v6.2d |
| ; CHECK-GI-NEXT: shl v17.2d, v17.2d, #63 |
| ; CHECK-GI-NEXT: shl v18.2d, v18.2d, #63 |
| ; CHECK-GI-NEXT: shl v4.2d, v4.2d, #63 |
| ; CHECK-GI-NEXT: sshl v16.2d, v16.2d, v20.2d |
| ; CHECK-GI-NEXT: sshl v6.2d, v6.2d, v20.2d |
| ; CHECK-GI-NEXT: sshr v17.2d, v17.2d, #63 |
| ; CHECK-GI-NEXT: sshr v18.2d, v18.2d, #63 |
| ; CHECK-GI-NEXT: sshr v4.2d, v4.2d, #63 |
| ; CHECK-GI-NEXT: bif v0.16b, v7.16b, v17.16b |
| ; CHECK-GI-NEXT: bif v1.16b, v16.16b, v18.16b |
| ; CHECK-GI-NEXT: bif v2.16b, v5.16b, v4.16b |
| ; CHECK-GI-NEXT: bif v3.16b, v6.16b, v18.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <8 x i64> %x, <i64 1, i64 4, i64 8, i64 16, i64 1, i64 4, i64 8, i64 16> |
| ret <8 x i64> %1 |
| } |
| |
| define <4 x i32> @combine_vec_sdiv_by_pow2b_PosAndNeg(<4 x i32> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_by_pow2b_PosAndNeg: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI24_0 |
| ; CHECK-SD-NEXT: cmlt v1.4s, v0.4s, #0 |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI24_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI24_1 |
| ; CHECK-SD-NEXT: ushl v1.4s, v1.4s, v2.4s |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI24_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI24_2 |
| ; CHECK-SD-NEXT: add v1.4s, v0.4s, v1.4s |
| ; CHECK-SD-NEXT: sshl v1.4s, v1.4s, v2.4s |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI24_2] |
| ; CHECK-SD-NEXT: bif v0.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: movi v1.2d, #0xffffffff00000000 |
| ; CHECK-SD-NEXT: neg v2.4s, v0.4s |
| ; CHECK-SD-NEXT: bit v0.16b, v2.16b, v1.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_by_pow2b_PosAndNeg: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #1 // =0x1 |
| ; CHECK-GI-NEXT: mov w9, #0 // =0x0 |
| ; CHECK-GI-NEXT: adrp x10, .LCPI24_0 |
| ; CHECK-GI-NEXT: fmov s1, w8 |
| ; CHECK-GI-NEXT: ldr q2, [x10, :lo12:.LCPI24_0] |
| ; CHECK-GI-NEXT: sshr v3.4s, v0.4s, #31 |
| ; CHECK-GI-NEXT: fmov s4, w9 |
| ; CHECK-GI-NEXT: adrp x10, .LCPI24_1 |
| ; CHECK-GI-NEXT: neg v2.4s, v2.4s |
| ; CHECK-GI-NEXT: mov v1.s[1], w9 |
| ; CHECK-GI-NEXT: mov v4.s[1], w8 |
| ; CHECK-GI-NEXT: ushl v2.4s, v3.4s, v2.4s |
| ; CHECK-GI-NEXT: ldr q3, [x10, :lo12:.LCPI24_1] |
| ; CHECK-GI-NEXT: mov v1.s[2], w9 |
| ; CHECK-GI-NEXT: neg v3.4s, v3.4s |
| ; CHECK-GI-NEXT: add v2.4s, v0.4s, v2.4s |
| ; CHECK-GI-NEXT: mov v4.d[1], v4.d[0] |
| ; CHECK-GI-NEXT: mov v1.s[3], w9 |
| ; CHECK-GI-NEXT: sshl v2.4s, v2.4s, v3.4s |
| ; CHECK-GI-NEXT: shl v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: sshr v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: bif v0.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: shl v1.4s, v4.4s, #31 |
| ; CHECK-GI-NEXT: sshr v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: neg v2.4s, v0.4s |
| ; CHECK-GI-NEXT: bit v0.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <4 x i32> %x, <i32 1, i32 -4, i32 8, i32 -16> |
| ret <4 x i32> %1 |
| } |
| |
| ; PR37119 |
| define <16 x i8> @non_splat_minus_one_divisor_0(<16 x i8> %A) { |
| ; CHECK-SD-LABEL: non_splat_minus_one_divisor_0: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: umov w9, v0.b[0] |
| ; CHECK-SD-NEXT: mov w8, wzr |
| ; CHECK-SD-NEXT: umov w10, v0.b[1] |
| ; CHECK-SD-NEXT: sub w9, w8, w9, sxtb |
| ; CHECK-SD-NEXT: sub w10, w8, w10, sxtb |
| ; CHECK-SD-NEXT: fmov s1, w9 |
| ; CHECK-SD-NEXT: smov w9, v0.b[2] |
| ; CHECK-SD-NEXT: mov v1.b[1], w10 |
| ; CHECK-SD-NEXT: umov w10, v0.b[3] |
| ; CHECK-SD-NEXT: mov v1.b[2], w9 |
| ; CHECK-SD-NEXT: sub w9, w8, w10, sxtb |
| ; CHECK-SD-NEXT: umov w10, v0.b[4] |
| ; CHECK-SD-NEXT: mov v1.b[3], w9 |
| ; CHECK-SD-NEXT: sub w9, w8, w10, sxtb |
| ; CHECK-SD-NEXT: umov w10, v0.b[5] |
| ; CHECK-SD-NEXT: mov v1.b[4], w9 |
| ; CHECK-SD-NEXT: sub w9, w8, w10, sxtb |
| ; CHECK-SD-NEXT: umov w10, v0.b[7] |
| ; CHECK-SD-NEXT: mov v1.b[5], w9 |
| ; CHECK-SD-NEXT: smov w9, v0.b[6] |
| ; CHECK-SD-NEXT: mov v1.b[6], w9 |
| ; CHECK-SD-NEXT: sub w9, w8, w10, sxtb |
| ; CHECK-SD-NEXT: umov w10, v0.b[8] |
| ; CHECK-SD-NEXT: mov v1.b[7], w9 |
| ; CHECK-SD-NEXT: sub w8, w8, w10, sxtb |
| ; CHECK-SD-NEXT: mov v1.b[8], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.b[9] |
| ; CHECK-SD-NEXT: mov v1.b[9], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.b[10] |
| ; CHECK-SD-NEXT: mov v1.b[10], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.b[11] |
| ; CHECK-SD-NEXT: mov v1.b[11], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.b[12] |
| ; CHECK-SD-NEXT: mov v1.b[12], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.b[13] |
| ; CHECK-SD-NEXT: mov v1.b[13], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.b[14] |
| ; CHECK-SD-NEXT: mov v1.b[14], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.b[15] |
| ; CHECK-SD-NEXT: mov v1.b[15], w8 |
| ; CHECK-SD-NEXT: mov v0.16b, v1.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: non_splat_minus_one_divisor_0: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI25_0 |
| ; CHECK-GI-NEXT: neg v2.16b, v0.16b |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI25_0] |
| ; CHECK-GI-NEXT: shl v1.16b, v1.16b, #7 |
| ; CHECK-GI-NEXT: sshr v1.16b, v1.16b, #7 |
| ; CHECK-GI-NEXT: bit v0.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: ret |
| %div = sdiv <16 x i8> %A, <i8 -1, i8 -1, i8 1, i8 -1, i8 -1, i8 -1, i8 1, i8 -1, i8 -1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> |
| ret <16 x i8> %div |
| } |
| |
| define <16 x i8> @non_splat_minus_one_divisor_1(<16 x i8> %A) { |
| ; CHECK-SD-LABEL: non_splat_minus_one_divisor_1: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI26_1 |
| ; CHECK-SD-NEXT: cmlt v1.16b, v0.16b, #0 |
| ; CHECK-SD-NEXT: adrp x9, .LCPI26_3 |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI26_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI26_2 |
| ; CHECK-SD-NEXT: ldr q3, [x9, :lo12:.LCPI26_3] |
| ; CHECK-SD-NEXT: adrp x9, .LCPI26_5 |
| ; CHECK-SD-NEXT: ushl v1.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI26_2] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI26_0 |
| ; CHECK-SD-NEXT: add v1.16b, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: sshl v1.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI26_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI26_4 |
| ; CHECK-SD-NEXT: and v0.16b, v0.16b, v2.16b |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI26_4] |
| ; CHECK-SD-NEXT: and v1.16b, v1.16b, v3.16b |
| ; CHECK-SD-NEXT: ldr q3, [x9, :lo12:.LCPI26_5] |
| ; CHECK-SD-NEXT: orr v0.16b, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: neg v1.16b, v0.16b |
| ; CHECK-SD-NEXT: and v0.16b, v0.16b, v2.16b |
| ; CHECK-SD-NEXT: and v1.16b, v1.16b, v3.16b |
| ; CHECK-SD-NEXT: orr v0.16b, v1.16b, v0.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: non_splat_minus_one_divisor_1: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI26_2 |
| ; CHECK-GI-NEXT: sshr v2.16b, v0.16b, #7 |
| ; CHECK-GI-NEXT: adrp x9, .LCPI26_1 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI26_2] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI26_3 |
| ; CHECK-GI-NEXT: ldr q3, [x9, :lo12:.LCPI26_1] |
| ; CHECK-GI-NEXT: neg v1.16b, v1.16b |
| ; CHECK-GI-NEXT: shl v3.16b, v3.16b, #7 |
| ; CHECK-GI-NEXT: ushl v1.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI26_3] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI26_0 |
| ; CHECK-GI-NEXT: neg v2.16b, v2.16b |
| ; CHECK-GI-NEXT: add v1.16b, v0.16b, v1.16b |
| ; CHECK-GI-NEXT: sshl v1.16b, v1.16b, v2.16b |
| ; CHECK-GI-NEXT: sshr v2.16b, v3.16b, #7 |
| ; CHECK-GI-NEXT: ldr q3, [x8, :lo12:.LCPI26_0] |
| ; CHECK-GI-NEXT: bif v0.16b, v1.16b, v2.16b |
| ; CHECK-GI-NEXT: shl v1.16b, v3.16b, #7 |
| ; CHECK-GI-NEXT: sshr v1.16b, v1.16b, #7 |
| ; CHECK-GI-NEXT: neg v2.16b, v0.16b |
| ; CHECK-GI-NEXT: bit v0.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: ret |
| %div = sdiv <16 x i8> %A, <i8 -1, i8 -1, i8 2, i8 -1, i8 -1, i8 -1, i8 2, i8 -1, i8 -1, i8 2, i8 2, i8 2, i8 2, i8 -128, i8 2, i8 -128> |
| ret <16 x i8> %div |
| } |
| |
| define <4 x i32> @non_splat_minus_one_divisor_2(<4 x i32> %A) { |
| ; CHECK-SD-LABEL: non_splat_minus_one_divisor_2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI27_0 |
| ; CHECK-SD-NEXT: cmlt v1.4s, v0.4s, #0 |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI27_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI27_1 |
| ; CHECK-SD-NEXT: ushl v1.4s, v1.4s, v2.4s |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI27_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI27_2 |
| ; CHECK-SD-NEXT: add v1.4s, v0.4s, v1.4s |
| ; CHECK-SD-NEXT: sshl v1.4s, v1.4s, v2.4s |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI27_2] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI27_3 |
| ; CHECK-SD-NEXT: bif v0.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI27_3] |
| ; CHECK-SD-NEXT: neg v1.4s, v0.4s |
| ; CHECK-SD-NEXT: bit v0.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: non_splat_minus_one_divisor_2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #1 // =0x1 |
| ; CHECK-GI-NEXT: adrp x9, .LCPI27_0 |
| ; CHECK-GI-NEXT: mov w10, #0 // =0x0 |
| ; CHECK-GI-NEXT: fmov s1, w8 |
| ; CHECK-GI-NEXT: ldr q2, [x9, :lo12:.LCPI27_0] |
| ; CHECK-GI-NEXT: fmov s4, w8 |
| ; CHECK-GI-NEXT: sshr v3.4s, v0.4s, #31 |
| ; CHECK-GI-NEXT: adrp x9, .LCPI27_1 |
| ; CHECK-GI-NEXT: neg v2.4s, v2.4s |
| ; CHECK-GI-NEXT: mov v1.s[1], w8 |
| ; CHECK-GI-NEXT: mov v4.s[1], w10 |
| ; CHECK-GI-NEXT: ushl v2.4s, v3.4s, v2.4s |
| ; CHECK-GI-NEXT: ldr q3, [x9, :lo12:.LCPI27_1] |
| ; CHECK-GI-NEXT: mov v1.s[2], w10 |
| ; CHECK-GI-NEXT: mov v4.s[2], w10 |
| ; CHECK-GI-NEXT: neg v3.4s, v3.4s |
| ; CHECK-GI-NEXT: add v2.4s, v0.4s, v2.4s |
| ; CHECK-GI-NEXT: mov v1.s[3], w10 |
| ; CHECK-GI-NEXT: sshl v2.4s, v2.4s, v3.4s |
| ; CHECK-GI-NEXT: mov v4.s[3], w8 |
| ; CHECK-GI-NEXT: shl v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: sshr v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: bif v0.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: shl v1.4s, v4.4s, #31 |
| ; CHECK-GI-NEXT: sshr v1.4s, v1.4s, #31 |
| ; CHECK-GI-NEXT: neg v2.4s, v0.4s |
| ; CHECK-GI-NEXT: bit v0.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: ret |
| %div = sdiv <4 x i32> %A, <i32 -1, i32 1, i32 2, i32 -2> |
| ret <4 x i32> %div |
| } |
| |
| define <8 x i16> @combine_vec_sdiv_nonuniform(<8 x i16> %x) { |
| ; CHECK-LABEL: combine_vec_sdiv_nonuniform: |
| ; CHECK: // %bb.0: |
| ; CHECK-NEXT: adrp x8, .LCPI28_0 |
| ; CHECK-NEXT: ldr q1, [x8, :lo12:.LCPI28_0] |
| ; CHECK-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-NEXT: smull v0.4s, v0.4h, v1.4h |
| ; CHECK-NEXT: uzp2 v0.8h, v0.8h, v2.8h |
| ; CHECK-NEXT: usra v0.8h, v0.8h, #15 |
| ; CHECK-NEXT: ret |
| %1 = sdiv <8 x i16> %x, <i16 3, i16 3, i16 3, i16 3, i16 22, i16 22, i16 22, i16 22> |
| ret <8 x i16> %1 |
| } |
| |
| define <8 x i16> @combine_vec_sdiv_nonuniform2(<8 x i16> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_nonuniform2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI29_0 |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI29_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI29_1 |
| ; CHECK-SD-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: smull v0.4s, v0.4h, v1.4h |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI29_1] |
| ; CHECK-SD-NEXT: uzp2 v0.8h, v0.8h, v2.8h |
| ; CHECK-SD-NEXT: sshl v0.8h, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: usra v0.8h, v0.8h, #15 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_nonuniform2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI29_1 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI29_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI29_0 |
| ; CHECK-GI-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: smull v0.4s, v0.4h, v1.4h |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI29_0] |
| ; CHECK-GI-NEXT: neg v1.8h, v1.8h |
| ; CHECK-GI-NEXT: uzp2 v0.8h, v0.8h, v2.8h |
| ; CHECK-GI-NEXT: sshl v0.8h, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: usra v0.8h, v0.8h, #15 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <8 x i16> %x, <i16 24, i16 24, i16 24, i16 24, i16 25, i16 25, i16 25, i16 25> |
| ret <8 x i16> %1 |
| } |
| |
| define <8 x i16> @combine_vec_sdiv_nonuniform3(<8 x i16> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_nonuniform3: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI30_0 |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI30_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI30_1 |
| ; CHECK-SD-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: smull v1.4s, v0.4h, v1.4h |
| ; CHECK-SD-NEXT: uzp2 v1.8h, v1.8h, v2.8h |
| ; CHECK-SD-NEXT: add v0.8h, v1.8h, v0.8h |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI30_1] |
| ; CHECK-SD-NEXT: sshl v0.8h, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: usra v0.8h, v0.8h, #15 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_nonuniform3: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI30_1 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI30_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI30_0 |
| ; CHECK-GI-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: smull v1.4s, v0.4h, v1.4h |
| ; CHECK-GI-NEXT: uzp2 v1.8h, v1.8h, v2.8h |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI30_0] |
| ; CHECK-GI-NEXT: add v0.8h, v1.8h, v0.8h |
| ; CHECK-GI-NEXT: neg v1.8h, v2.8h |
| ; CHECK-GI-NEXT: sshl v0.8h, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: usra v0.8h, v0.8h, #15 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <8 x i16> %x, <i16 23, i16 23, i16 23, i16 23, i16 511, i16 511, i16 511, i16 511> |
| ret <8 x i16> %1 |
| } |
| |
| define <8 x i16> @combine_vec_sdiv_nonuniform4(<8 x i16> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_nonuniform4: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI31_0 |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI31_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI31_1 |
| ; CHECK-SD-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: smull v1.4s, v0.4h, v1.4h |
| ; CHECK-SD-NEXT: uzp2 v1.8h, v1.8h, v2.8h |
| ; CHECK-SD-NEXT: sub v0.8h, v1.8h, v0.8h |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI31_1] |
| ; CHECK-SD-NEXT: sshl v0.8h, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: usra v0.8h, v0.8h, #15 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_nonuniform4: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI31_1 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI31_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI31_0 |
| ; CHECK-GI-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: smull v1.4s, v0.4h, v1.4h |
| ; CHECK-GI-NEXT: uzp2 v1.8h, v1.8h, v2.8h |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI31_0] |
| ; CHECK-GI-NEXT: sub v0.8h, v1.8h, v0.8h |
| ; CHECK-GI-NEXT: neg v1.8h, v2.8h |
| ; CHECK-GI-NEXT: sshl v0.8h, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: usra v0.8h, v0.8h, #15 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <8 x i16> %x, <i16 -23, i16 -23, i16 -23, i16 -23, i16 -510, i16 -510, i16 -510, i16 -510> |
| ret <8 x i16> %1 |
| } |
| |
| define <8 x i16> @combine_vec_sdiv_nonuniform5(<8 x i16> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_nonuniform5: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI32_0 |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI32_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI32_1 |
| ; CHECK-SD-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: smull v1.4s, v0.4h, v1.4h |
| ; CHECK-SD-NEXT: uzp2 v1.8h, v1.8h, v2.8h |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI32_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI32_2 |
| ; CHECK-SD-NEXT: mla v1.8h, v0.8h, v2.8h |
| ; CHECK-SD-NEXT: ldr q0, [x8, :lo12:.LCPI32_2] |
| ; CHECK-SD-NEXT: sshl v0.8h, v1.8h, v0.8h |
| ; CHECK-SD-NEXT: usra v0.8h, v0.8h, #15 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_nonuniform5: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI32_2 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI32_2] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI32_1 |
| ; CHECK-GI-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: smull v1.4s, v0.4h, v1.4h |
| ; CHECK-GI-NEXT: uzp2 v1.8h, v1.8h, v2.8h |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI32_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI32_0 |
| ; CHECK-GI-NEXT: ldr q3, [x8, :lo12:.LCPI32_0] |
| ; CHECK-GI-NEXT: mla v1.8h, v0.8h, v2.8h |
| ; CHECK-GI-NEXT: neg v0.8h, v3.8h |
| ; CHECK-GI-NEXT: sshl v0.8h, v1.8h, v0.8h |
| ; CHECK-GI-NEXT: usra v0.8h, v0.8h, #15 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <8 x i16> %x, <i16 -510, i16 -24, i16 -23, i16 3, i16 22, i16 25, i16 255, i16 511> |
| ret <8 x i16> %1 |
| } |
| |
| define <8 x i16> @combine_vec_sdiv_nonuniform6(<8 x i16> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_nonuniform6: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI33_0 |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI33_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI33_1 |
| ; CHECK-SD-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: smull v1.4s, v0.4h, v1.4h |
| ; CHECK-SD-NEXT: uzp2 v1.8h, v1.8h, v2.8h |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI33_1] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI33_2 |
| ; CHECK-SD-NEXT: mla v1.8h, v0.8h, v2.8h |
| ; CHECK-SD-NEXT: ldr q0, [x8, :lo12:.LCPI33_2] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI33_3 |
| ; CHECK-SD-NEXT: ldr q2, [x8, :lo12:.LCPI33_3] |
| ; CHECK-SD-NEXT: sshl v0.8h, v1.8h, v0.8h |
| ; CHECK-SD-NEXT: ushr v1.8h, v0.8h, #15 |
| ; CHECK-SD-NEXT: and v1.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: add v0.8h, v0.8h, v1.8h |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_nonuniform6: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI33_3 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI33_3] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI33_2 |
| ; CHECK-GI-NEXT: smull2 v2.4s, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: smull v1.4s, v0.4h, v1.4h |
| ; CHECK-GI-NEXT: uzp2 v1.8h, v1.8h, v2.8h |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI33_2] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI33_1 |
| ; CHECK-GI-NEXT: ldr q3, [x8, :lo12:.LCPI33_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI33_0 |
| ; CHECK-GI-NEXT: mla v1.8h, v0.8h, v2.8h |
| ; CHECK-GI-NEXT: neg v0.8h, v3.8h |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI33_0] |
| ; CHECK-GI-NEXT: sshl v0.8h, v1.8h, v0.8h |
| ; CHECK-GI-NEXT: ushr v1.8h, v0.8h, #15 |
| ; CHECK-GI-NEXT: and v1.16b, v1.16b, v2.16b |
| ; CHECK-GI-NEXT: add v0.8h, v0.8h, v1.8h |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <8 x i16> %x, <i16 -32768, i16 -512, i16 -511, i16 -1, i16 1, i16 255, i16 512, i16 32767> |
| ret <8 x i16> %1 |
| } |
| |
| define <8 x i16> @combine_vec_sdiv_nonuniform7(<8 x i16> %x) { |
| ; CHECK-SD-LABEL: combine_vec_sdiv_nonuniform7: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: umov w9, v0.h[0] |
| ; CHECK-SD-NEXT: mov w8, wzr |
| ; CHECK-SD-NEXT: umov w10, v0.h[1] |
| ; CHECK-SD-NEXT: umov w11, v0.h[2] |
| ; CHECK-SD-NEXT: sub w9, w8, w9, sxth |
| ; CHECK-SD-NEXT: sub w10, w8, w10, sxth |
| ; CHECK-SD-NEXT: fmov s1, w9 |
| ; CHECK-SD-NEXT: sub w9, w8, w11, sxth |
| ; CHECK-SD-NEXT: mov v1.h[1], w10 |
| ; CHECK-SD-NEXT: umov w10, v0.h[3] |
| ; CHECK-SD-NEXT: mov v1.h[2], w9 |
| ; CHECK-SD-NEXT: sub w8, w8, w10, sxth |
| ; CHECK-SD-NEXT: mov v1.h[3], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.h[4] |
| ; CHECK-SD-NEXT: mov v1.h[4], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.h[5] |
| ; CHECK-SD-NEXT: mov v1.h[5], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.h[6] |
| ; CHECK-SD-NEXT: mov v1.h[6], w8 |
| ; CHECK-SD-NEXT: smov w8, v0.h[7] |
| ; CHECK-SD-NEXT: mov v1.h[7], w8 |
| ; CHECK-SD-NEXT: mov v0.16b, v1.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_vec_sdiv_nonuniform7: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI34_0 |
| ; CHECK-GI-NEXT: neg v2.8h, v0.8h |
| ; CHECK-GI-NEXT: ldr d1, [x8, :lo12:.LCPI34_0] |
| ; CHECK-GI-NEXT: ushll v1.8h, v1.8b, #0 |
| ; CHECK-GI-NEXT: shl v1.8h, v1.8h, #15 |
| ; CHECK-GI-NEXT: sshr v1.8h, v1.8h, #15 |
| ; CHECK-GI-NEXT: bit v0.16b, v2.16b, v1.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <8 x i16> %x, <i16 -1, i16 -1, i16 -1, i16 -1, i16 1, i16 1, i16 1, i16 1> |
| ret <8 x i16> %1 |
| } |
| |
| define <16 x i8> @pr38658(<16 x i8> %x) { |
| ; CHECK-SD-LABEL: pr38658: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: adrp x8, .LCPI35_0 |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI35_0] |
| ; CHECK-SD-NEXT: adrp x8, .LCPI35_1 |
| ; CHECK-SD-NEXT: smull2 v2.8h, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: smull v1.8h, v0.8b, v1.8b |
| ; CHECK-SD-NEXT: uzp2 v1.16b, v1.16b, v2.16b |
| ; CHECK-SD-NEXT: add v0.16b, v1.16b, v0.16b |
| ; CHECK-SD-NEXT: ldr q1, [x8, :lo12:.LCPI35_1] |
| ; CHECK-SD-NEXT: sshl v0.16b, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: movi v1.2d, #0000000000000000 |
| ; CHECK-SD-NEXT: ushr v2.16b, v0.16b, #7 |
| ; CHECK-SD-NEXT: mov v1.b[15], v2.b[15] |
| ; CHECK-SD-NEXT: add v0.16b, v0.16b, v1.16b |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: pr38658: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: adrp x8, .LCPI35_2 |
| ; CHECK-GI-NEXT: ldr q1, [x8, :lo12:.LCPI35_2] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI35_1 |
| ; CHECK-GI-NEXT: smull2 v2.8h, v0.16b, v1.16b |
| ; CHECK-GI-NEXT: smull v1.8h, v0.8b, v1.8b |
| ; CHECK-GI-NEXT: uzp2 v1.16b, v1.16b, v2.16b |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI35_1] |
| ; CHECK-GI-NEXT: adrp x8, .LCPI35_0 |
| ; CHECK-GI-NEXT: add v0.16b, v1.16b, v0.16b |
| ; CHECK-GI-NEXT: neg v1.16b, v2.16b |
| ; CHECK-GI-NEXT: ldr q2, [x8, :lo12:.LCPI35_0] |
| ; CHECK-GI-NEXT: sshl v0.16b, v0.16b, v1.16b |
| ; CHECK-GI-NEXT: ushr v1.16b, v0.16b, #7 |
| ; CHECK-GI-NEXT: and v1.16b, v1.16b, v2.16b |
| ; CHECK-GI-NEXT: add v0.16b, v0.16b, v1.16b |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv <16 x i8> %x, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 7> |
| ret <16 x i8> %1 |
| } |
| |
| define i1 @bool_sdiv(i1 %x, i1 %y) { |
| ; CHECK-SD-LABEL: bool_sdiv: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: and w0, w0, #0x1 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: bool_sdiv: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sbfx w8, w0, #0, #1 |
| ; CHECK-GI-NEXT: sbfx w9, w1, #0, #1 |
| ; CHECK-GI-NEXT: sdiv w8, w8, w9 |
| ; CHECK-GI-NEXT: and w0, w8, #0x1 |
| ; CHECK-GI-NEXT: ret |
| %r = sdiv i1 %x, %y |
| ret i1 %r |
| } |
| |
| define <4 x i1> @boolvec_sdiv(<4 x i1> %x, <4 x i1> %y) { |
| ; CHECK-SD-LABEL: boolvec_sdiv: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: boolvec_sdiv: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: // kill: def $d0 killed $d0 def $q0 |
| ; CHECK-GI-NEXT: // kill: def $d1 killed $d1 def $q1 |
| ; CHECK-GI-NEXT: umov w8, v0.h[0] |
| ; CHECK-GI-NEXT: umov w9, v1.h[0] |
| ; CHECK-GI-NEXT: umov w10, v1.h[1] |
| ; CHECK-GI-NEXT: umov w11, v1.h[2] |
| ; CHECK-GI-NEXT: umov w12, v1.h[3] |
| ; CHECK-GI-NEXT: sbfx w8, w8, #0, #1 |
| ; CHECK-GI-NEXT: sbfx w9, w9, #0, #1 |
| ; CHECK-GI-NEXT: sbfx w10, w10, #0, #1 |
| ; CHECK-GI-NEXT: sbfx w11, w11, #0, #1 |
| ; CHECK-GI-NEXT: sbfx w12, w12, #0, #1 |
| ; CHECK-GI-NEXT: sdiv w8, w8, w9 |
| ; CHECK-GI-NEXT: umov w9, v0.h[1] |
| ; CHECK-GI-NEXT: sbfx w9, w9, #0, #1 |
| ; CHECK-GI-NEXT: sdiv w9, w9, w10 |
| ; CHECK-GI-NEXT: umov w10, v0.h[2] |
| ; CHECK-GI-NEXT: sbfx w10, w10, #0, #1 |
| ; CHECK-GI-NEXT: sdiv w10, w10, w11 |
| ; CHECK-GI-NEXT: umov w11, v0.h[3] |
| ; CHECK-GI-NEXT: fmov s0, w8 |
| ; CHECK-GI-NEXT: mov v0.h[1], w9 |
| ; CHECK-GI-NEXT: sbfx w11, w11, #0, #1 |
| ; CHECK-GI-NEXT: sdiv w8, w11, w12 |
| ; CHECK-GI-NEXT: mov v0.h[2], w10 |
| ; CHECK-GI-NEXT: mov v0.h[3], w8 |
| ; CHECK-GI-NEXT: // kill: def $d0 killed $d0 killed $q0 |
| ; CHECK-GI-NEXT: ret |
| %r = sdiv <4 x i1> %x, %y |
| ret <4 x i1> %r |
| } |
| |
| define i32 @combine_sdiv_two(i32 %x) { |
| ; CHECK-SD-LABEL: combine_sdiv_two: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: add w8, w0, w0, lsr #31 |
| ; CHECK-SD-NEXT: asr w0, w8, #1 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_sdiv_two: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: asr w8, w0, #31 |
| ; CHECK-GI-NEXT: add w8, w0, w8, lsr #31 |
| ; CHECK-GI-NEXT: asr w0, w8, #1 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i32 %x, 2 |
| ret i32 %1 |
| } |
| |
| define i32 @combine_sdiv_negtwo(i32 %x) { |
| ; CHECK-SD-LABEL: combine_sdiv_negtwo: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: add w8, w0, w0, lsr #31 |
| ; CHECK-SD-NEXT: neg w0, w8, asr #1 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_sdiv_negtwo: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: asr w8, w0, #31 |
| ; CHECK-GI-NEXT: add w8, w0, w8, lsr #31 |
| ; CHECK-GI-NEXT: neg w0, w8, asr #1 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i32 %x, -2 |
| ret i32 %1 |
| } |
| |
| define i8 @combine_i8_sdiv_pow2(i8 %x) { |
| ; CHECK-SD-LABEL: combine_i8_sdiv_pow2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: sxtb w8, w0 |
| ; CHECK-SD-NEXT: ubfx w8, w8, #11, #4 |
| ; CHECK-SD-NEXT: add w8, w0, w8 |
| ; CHECK-SD-NEXT: sbfx w0, w8, #4, #4 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i8_sdiv_pow2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sbfx w8, w0, #7, #1 |
| ; CHECK-GI-NEXT: ubfx w8, w8, #4, #4 |
| ; CHECK-GI-NEXT: add w8, w0, w8 |
| ; CHECK-GI-NEXT: sbfx w0, w8, #4, #4 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i8 %x, 16 |
| ret i8 %1 |
| } |
| |
| define i8 @combine_i8_sdiv_negpow2(i8 %x) { |
| ; CHECK-SD-LABEL: combine_i8_sdiv_negpow2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: sxtb w8, w0 |
| ; CHECK-SD-NEXT: ubfx w8, w8, #9, #6 |
| ; CHECK-SD-NEXT: add w8, w0, w8 |
| ; CHECK-SD-NEXT: sxtb w8, w8 |
| ; CHECK-SD-NEXT: neg w0, w8, asr #6 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i8_sdiv_negpow2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sbfx w8, w0, #7, #1 |
| ; CHECK-GI-NEXT: ubfx w8, w8, #2, #6 |
| ; CHECK-GI-NEXT: add w8, w0, w8 |
| ; CHECK-GI-NEXT: sxtb w8, w8 |
| ; CHECK-GI-NEXT: neg w0, w8, asr #6 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i8 %x, -64 |
| ret i8 %1 |
| } |
| |
| define i16 @combine_i16_sdiv_pow2(i16 %x) { |
| ; CHECK-SD-LABEL: combine_i16_sdiv_pow2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: sxth w8, w0 |
| ; CHECK-SD-NEXT: ubfx w8, w8, #27, #4 |
| ; CHECK-SD-NEXT: add w8, w0, w8 |
| ; CHECK-SD-NEXT: sbfx w0, w8, #4, #12 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i16_sdiv_pow2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sbfx w8, w0, #15, #1 |
| ; CHECK-GI-NEXT: ubfx w8, w8, #12, #4 |
| ; CHECK-GI-NEXT: add w8, w0, w8 |
| ; CHECK-GI-NEXT: sbfx w0, w8, #4, #12 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i16 %x, 16 |
| ret i16 %1 |
| } |
| |
| define i16 @combine_i16_sdiv_negpow2(i16 %x) { |
| ; CHECK-SD-LABEL: combine_i16_sdiv_negpow2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: sxth w8, w0 |
| ; CHECK-SD-NEXT: lsr w8, w8, #23 |
| ; CHECK-SD-NEXT: add w8, w0, w8, uxtb |
| ; CHECK-SD-NEXT: sxth w8, w8 |
| ; CHECK-SD-NEXT: neg w0, w8, asr #8 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i16_sdiv_negpow2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sbfx w8, w0, #15, #1 |
| ; CHECK-GI-NEXT: ubfx w8, w8, #8, #8 |
| ; CHECK-GI-NEXT: add w8, w0, w8 |
| ; CHECK-GI-NEXT: sxth w8, w8 |
| ; CHECK-GI-NEXT: neg w0, w8, asr #8 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i16 %x, -256 |
| ret i16 %1 |
| } |
| |
| define i32 @combine_i32_sdiv_pow2(i32 %x) { |
| ; CHECK-SD-LABEL: combine_i32_sdiv_pow2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: add w8, w0, #15 |
| ; CHECK-SD-NEXT: cmp w0, #0 |
| ; CHECK-SD-NEXT: csel w8, w8, w0, lt |
| ; CHECK-SD-NEXT: asr w0, w8, #4 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i32_sdiv_pow2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: asr w8, w0, #31 |
| ; CHECK-GI-NEXT: add w8, w0, w8, lsr #28 |
| ; CHECK-GI-NEXT: asr w0, w8, #4 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i32 %x, 16 |
| ret i32 %1 |
| } |
| |
| define i32 @combine_i32_sdiv_negpow2(i32 %x) { |
| ; CHECK-SD-LABEL: combine_i32_sdiv_negpow2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: add w8, w0, #255 |
| ; CHECK-SD-NEXT: cmp w0, #0 |
| ; CHECK-SD-NEXT: csel w8, w8, w0, lt |
| ; CHECK-SD-NEXT: neg w0, w8, asr #8 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i32_sdiv_negpow2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: asr w8, w0, #31 |
| ; CHECK-GI-NEXT: add w8, w0, w8, lsr #24 |
| ; CHECK-GI-NEXT: neg w0, w8, asr #8 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i32 %x, -256 |
| ret i32 %1 |
| } |
| |
| define i64 @combine_i64_sdiv_pow2(i64 %x) { |
| ; CHECK-SD-LABEL: combine_i64_sdiv_pow2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: add x8, x0, #15 |
| ; CHECK-SD-NEXT: cmp x0, #0 |
| ; CHECK-SD-NEXT: csel x8, x8, x0, lt |
| ; CHECK-SD-NEXT: asr x0, x8, #4 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i64_sdiv_pow2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: asr x8, x0, #63 |
| ; CHECK-GI-NEXT: add x8, x0, x8, lsr #60 |
| ; CHECK-GI-NEXT: asr x0, x8, #4 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i64 %x, 16 |
| ret i64 %1 |
| } |
| |
| define i64 @combine_i64_sdiv_negpow2(i64 %x) { |
| ; CHECK-SD-LABEL: combine_i64_sdiv_negpow2: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: add x8, x0, #255 |
| ; CHECK-SD-NEXT: cmp x0, #0 |
| ; CHECK-SD-NEXT: csel x8, x8, x0, lt |
| ; CHECK-SD-NEXT: neg x0, x8, asr #8 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i64_sdiv_negpow2: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: asr x8, x0, #63 |
| ; CHECK-GI-NEXT: add x8, x0, x8, lsr #56 |
| ; CHECK-GI-NEXT: neg x0, x8, asr #8 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i64 %x, -256 |
| ret i64 %1 |
| } |
| |
| define i5 @combine_i5_sdiv_const7(i5 %x) { |
| ; CHECK-SD-LABEL: combine_i5_sdiv_const7: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: // kill: def $w0 killed $w0 def $x0 |
| ; CHECK-SD-NEXT: mov x8, #-56173 // =0xffffffffffff2493 |
| ; CHECK-SD-NEXT: sbfx x9, x0, #0, #5 |
| ; CHECK-SD-NEXT: movk x8, #37449, lsl #16 |
| ; CHECK-SD-NEXT: smull x8, w9, w8 |
| ; CHECK-SD-NEXT: lsl w9, w0, #27 |
| ; CHECK-SD-NEXT: lsr x8, x8, #32 |
| ; CHECK-SD-NEXT: add w8, w8, w9, asr #27 |
| ; CHECK-SD-NEXT: asr w9, w8, #2 |
| ; CHECK-SD-NEXT: add w0, w9, w8, lsr #31 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i5_sdiv_const7: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #19 // =0x13 |
| ; CHECK-GI-NEXT: sbfx w9, w0, #0, #5 |
| ; CHECK-GI-NEXT: sbfx w8, w8, #0, #5 |
| ; CHECK-GI-NEXT: mul w8, w9, w8 |
| ; CHECK-GI-NEXT: sbfx w8, w8, #0, #10 |
| ; CHECK-GI-NEXT: add w8, w0, w8, asr #5 |
| ; CHECK-GI-NEXT: sbfx w8, w8, #0, #5 |
| ; CHECK-GI-NEXT: asr w8, w8, #2 |
| ; CHECK-GI-NEXT: ubfx w9, w8, #4, #1 |
| ; CHECK-GI-NEXT: add w0, w8, w9 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i5 %x, 7 |
| ret i5 %1 |
| } |
| |
| define i5 @combine_i5_sdiv_const100(i5 %x) { |
| ; CHECK-SD-LABEL: combine_i5_sdiv_const100: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: sbfx w8, w0, #4, #1 |
| ; CHECK-SD-NEXT: and w8, w8, #0x3 |
| ; CHECK-SD-NEXT: add w8, w0, w8 |
| ; CHECK-SD-NEXT: sbfx w0, w8, #2, #3 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i5_sdiv_const100: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sbfx w8, w0, #0, #5 |
| ; CHECK-GI-NEXT: asr w8, w8, #4 |
| ; CHECK-GI-NEXT: ubfx w8, w8, #3, #2 |
| ; CHECK-GI-NEXT: add w8, w0, w8 |
| ; CHECK-GI-NEXT: sbfx w8, w8, #0, #5 |
| ; CHECK-GI-NEXT: asr w0, w8, #2 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i5 %x, 100 |
| ret i5 %1 |
| } |
| |
| define i8 @combine_i8_sdiv_const7(i8 %x) { |
| ; CHECK-SD-LABEL: combine_i8_sdiv_const7: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: sxtb w8, w0 |
| ; CHECK-SD-NEXT: mov w9, #-109 // =0xffffff93 |
| ; CHECK-SD-NEXT: mul w8, w8, w9 |
| ; CHECK-SD-NEXT: add w8, w0, w8, lsr #8 |
| ; CHECK-SD-NEXT: sbfx w9, w8, #2, #6 |
| ; CHECK-SD-NEXT: and w8, w8, #0x80 |
| ; CHECK-SD-NEXT: add w0, w9, w8, lsr #7 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i8_sdiv_const7: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sxtb w8, w0 |
| ; CHECK-GI-NEXT: mov w9, #-109 // =0xffffff93 |
| ; CHECK-GI-NEXT: mul w8, w8, w9 |
| ; CHECK-GI-NEXT: sxth w8, w8 |
| ; CHECK-GI-NEXT: add w8, w0, w8, asr #8 |
| ; CHECK-GI-NEXT: sbfx w8, w8, #2, #6 |
| ; CHECK-GI-NEXT: ubfx w9, w8, #7, #1 |
| ; CHECK-GI-NEXT: add w0, w8, w9 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i8 %x, 7 |
| ret i8 %1 |
| } |
| |
| define i8 @combine_i8_sdiv_const100(i8 %x) { |
| ; CHECK-SD-LABEL: combine_i8_sdiv_const100: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: sxtb w8, w0 |
| ; CHECK-SD-NEXT: mov w9, #41 // =0x29 |
| ; CHECK-SD-NEXT: mul w8, w8, w9 |
| ; CHECK-SD-NEXT: asr w9, w8, #12 |
| ; CHECK-SD-NEXT: add w0, w9, w8, lsr #31 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i8_sdiv_const100: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sxtb w8, w0 |
| ; CHECK-GI-NEXT: mov w9, #41 // =0x29 |
| ; CHECK-GI-NEXT: mul w8, w8, w9 |
| ; CHECK-GI-NEXT: sbfx w8, w8, #8, #8 |
| ; CHECK-GI-NEXT: asr w8, w8, #4 |
| ; CHECK-GI-NEXT: ubfx w9, w8, #7, #1 |
| ; CHECK-GI-NEXT: add w0, w8, w9 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i8 %x, 100 |
| ret i8 %1 |
| } |
| |
| define i16 @combine_i16_sdiv_const7(i16 %x) { |
| ; CHECK-SD-LABEL: combine_i16_sdiv_const7: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: sxth w8, w0 |
| ; CHECK-SD-NEXT: mov w9, #18725 // =0x4925 |
| ; CHECK-SD-NEXT: mul w8, w8, w9 |
| ; CHECK-SD-NEXT: asr w9, w8, #17 |
| ; CHECK-SD-NEXT: add w0, w9, w8, lsr #31 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i16_sdiv_const7: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sxth w8, w0 |
| ; CHECK-GI-NEXT: mov w9, #18725 // =0x4925 |
| ; CHECK-GI-NEXT: mul w8, w8, w9 |
| ; CHECK-GI-NEXT: asr w8, w8, #16 |
| ; CHECK-GI-NEXT: asr w8, w8, #1 |
| ; CHECK-GI-NEXT: ubfx w9, w8, #15, #1 |
| ; CHECK-GI-NEXT: add w0, w8, w9 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i16 %x, 7 |
| ret i16 %1 |
| } |
| |
| define i16 @combine_i16_sdiv_const100(i16 %x) { |
| ; CHECK-SD-LABEL: combine_i16_sdiv_const100: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: sxth w8, w0 |
| ; CHECK-SD-NEXT: mov w9, #5243 // =0x147b |
| ; CHECK-SD-NEXT: mul w8, w8, w9 |
| ; CHECK-SD-NEXT: asr w9, w8, #19 |
| ; CHECK-SD-NEXT: add w0, w9, w8, lsr #31 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i16_sdiv_const100: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: sxth w8, w0 |
| ; CHECK-GI-NEXT: mov w9, #5243 // =0x147b |
| ; CHECK-GI-NEXT: mul w8, w8, w9 |
| ; CHECK-GI-NEXT: asr w8, w8, #16 |
| ; CHECK-GI-NEXT: asr w8, w8, #3 |
| ; CHECK-GI-NEXT: ubfx w9, w8, #15, #1 |
| ; CHECK-GI-NEXT: add w0, w8, w9 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i16 %x, 100 |
| ret i16 %1 |
| } |
| |
| define i32 @combine_i32_sdiv_const7(i32 %x) { |
| ; CHECK-SD-LABEL: combine_i32_sdiv_const7: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: mov w8, #9363 // =0x2493 |
| ; CHECK-SD-NEXT: movk w8, #37449, lsl #16 |
| ; CHECK-SD-NEXT: smull x8, w0, w8 |
| ; CHECK-SD-NEXT: lsr x8, x8, #32 |
| ; CHECK-SD-NEXT: add w8, w8, w0 |
| ; CHECK-SD-NEXT: asr w9, w8, #2 |
| ; CHECK-SD-NEXT: add w0, w9, w8, lsr #31 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i32_sdiv_const7: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #9363 // =0x2493 |
| ; CHECK-GI-NEXT: movk w8, #37449, lsl #16 |
| ; CHECK-GI-NEXT: smull x8, w0, w8 |
| ; CHECK-GI-NEXT: asr x8, x8, #32 |
| ; CHECK-GI-NEXT: add w8, w8, w0 |
| ; CHECK-GI-NEXT: asr w8, w8, #2 |
| ; CHECK-GI-NEXT: add w0, w8, w8, lsr #31 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i32 %x, 7 |
| ret i32 %1 |
| } |
| |
| define i32 @combine_i32_sdiv_const100(i32 %x) { |
| ; CHECK-SD-LABEL: combine_i32_sdiv_const100: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: mov w8, #34079 // =0x851f |
| ; CHECK-SD-NEXT: movk w8, #20971, lsl #16 |
| ; CHECK-SD-NEXT: smull x8, w0, w8 |
| ; CHECK-SD-NEXT: asr x8, x8, #37 |
| ; CHECK-SD-NEXT: add w0, w8, w8, lsr #31 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i32_sdiv_const100: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov w8, #34079 // =0x851f |
| ; CHECK-GI-NEXT: movk w8, #20971, lsl #16 |
| ; CHECK-GI-NEXT: smull x8, w0, w8 |
| ; CHECK-GI-NEXT: asr x8, x8, #32 |
| ; CHECK-GI-NEXT: asr w8, w8, #5 |
| ; CHECK-GI-NEXT: add w0, w8, w8, lsr #31 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i32 %x, 100 |
| ret i32 %1 |
| } |
| |
| define i64 @combine_i64_sdiv_const7(i64 %x) { |
| ; CHECK-SD-LABEL: combine_i64_sdiv_const7: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: mov x8, #18725 // =0x4925 |
| ; CHECK-SD-NEXT: movk x8, #9362, lsl #16 |
| ; CHECK-SD-NEXT: movk x8, #37449, lsl #32 |
| ; CHECK-SD-NEXT: movk x8, #18724, lsl #48 |
| ; CHECK-SD-NEXT: smulh x8, x0, x8 |
| ; CHECK-SD-NEXT: asr x9, x8, #1 |
| ; CHECK-SD-NEXT: add x0, x9, x8, lsr #63 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i64_sdiv_const7: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov x8, #18725 // =0x4925 |
| ; CHECK-GI-NEXT: movk x8, #9362, lsl #16 |
| ; CHECK-GI-NEXT: movk x8, #37449, lsl #32 |
| ; CHECK-GI-NEXT: movk x8, #18724, lsl #48 |
| ; CHECK-GI-NEXT: smulh x8, x0, x8 |
| ; CHECK-GI-NEXT: asr x8, x8, #1 |
| ; CHECK-GI-NEXT: add x0, x8, x8, lsr #63 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i64 %x, 7 |
| ret i64 %1 |
| } |
| |
| define i64 @combine_i64_sdiv_const100(i64 %x) { |
| ; CHECK-SD-LABEL: combine_i64_sdiv_const100: |
| ; CHECK-SD: // %bb.0: |
| ; CHECK-SD-NEXT: mov x8, #55051 // =0xd70b |
| ; CHECK-SD-NEXT: movk x8, #28835, lsl #16 |
| ; CHECK-SD-NEXT: movk x8, #2621, lsl #32 |
| ; CHECK-SD-NEXT: movk x8, #41943, lsl #48 |
| ; CHECK-SD-NEXT: smulh x8, x0, x8 |
| ; CHECK-SD-NEXT: add x8, x8, x0 |
| ; CHECK-SD-NEXT: asr x9, x8, #6 |
| ; CHECK-SD-NEXT: add x0, x9, x8, lsr #63 |
| ; CHECK-SD-NEXT: ret |
| ; |
| ; CHECK-GI-LABEL: combine_i64_sdiv_const100: |
| ; CHECK-GI: // %bb.0: |
| ; CHECK-GI-NEXT: mov x8, #55051 // =0xd70b |
| ; CHECK-GI-NEXT: movk x8, #28835, lsl #16 |
| ; CHECK-GI-NEXT: movk x8, #2621, lsl #32 |
| ; CHECK-GI-NEXT: movk x8, #41943, lsl #48 |
| ; CHECK-GI-NEXT: smulh x8, x0, x8 |
| ; CHECK-GI-NEXT: add x8, x8, x0 |
| ; CHECK-GI-NEXT: asr x8, x8, #6 |
| ; CHECK-GI-NEXT: add x0, x8, x8, lsr #63 |
| ; CHECK-GI-NEXT: ret |
| %1 = sdiv i64 %x, 100 |
| ret i64 %1 |
| } |
| |
| define i128 @combine_i128_sdiv_const7(i128 %x) { |
| ; CHECK-LABEL: combine_i128_sdiv_const7: |
| ; CHECK: // %bb.0: |
| ; CHECK-NEXT: str x30, [sp, #-16]! // 8-byte Folded Spill |
| ; CHECK-NEXT: .cfi_def_cfa_offset 16 |
| ; CHECK-NEXT: .cfi_offset w30, -16 |
| ; CHECK-NEXT: mov w2, #7 // =0x7 |
| ; CHECK-NEXT: mov x3, xzr |
| ; CHECK-NEXT: bl __divti3 |
| ; CHECK-NEXT: ldr x30, [sp], #16 // 8-byte Folded Reload |
| ; CHECK-NEXT: ret |
| %1 = sdiv i128 %x, 7 |
| ret i128 %1 |
| } |
| |
| define i128 @combine_i128_sdiv_const100(i128 %x) { |
| ; CHECK-LABEL: combine_i128_sdiv_const100: |
| ; CHECK: // %bb.0: |
| ; CHECK-NEXT: str x30, [sp, #-16]! // 8-byte Folded Spill |
| ; CHECK-NEXT: .cfi_def_cfa_offset 16 |
| ; CHECK-NEXT: .cfi_offset w30, -16 |
| ; CHECK-NEXT: mov w2, #100 // =0x64 |
| ; CHECK-NEXT: mov x3, xzr |
| ; CHECK-NEXT: bl __divti3 |
| ; CHECK-NEXT: ldr x30, [sp], #16 // 8-byte Folded Reload |
| ; CHECK-NEXT: ret |
| %1 = sdiv i128 %x, 100 |
| ret i128 %1 |
| } |
