| /external/llvm/test/CodeGen/AArch64/ |
| D | cxx-tlscc.ll | 1 ; RUN: llc < %s -mtriple=aarch64-apple-ios | FileCheck %s
2 ; RUN: llc < %s -mtriple=aarch64-apple-ios -enable-shrink-wrap=true | FileCheck %s
6 ; RUN: llc < %s -mtriple=aarch64-apple-ios -O0 | FileCheck --check-prefix=CHECK-O0 %s
33 ; CHECK-LABEL: _ZTW2sg
34 ; CHECK-NOT: stp d31, d30
35 ; CHECK-NOT: stp d29, d28
36 ; CHECK-NOT: stp d27, d26
37 ; CHECK-NOT: stp d25, d24
38 ; CHECK-NOT: stp d23, d22
39 ; CHECK-NOT: stp d21, d20
[all …]
|
| D | swifterror.ll | 1 ; RUN: llc -verify-machineinstrs -disable-fp-elim < %s -mtriple=aarch64-apple-ios -disable-post-ra …
2 ; RUN: llc -verify-machineinstrs -disable-fp-elim -O0 < %s -mtriple=aarch64-apple-ios -disable-post…
11 ; CHECK-APPLE-LABEL: foo:
12 ; CHECK-APPLE: orr w0, wzr, #0x10
13 ; CHECK-APPLE: malloc
14 ; CHECK-APPLE: orr [[ID:w[0-9]+]], wzr, #0x1
15 ; CHECK-APPLE: strb [[ID]], [x0, #8]
16 ; CHECK-APPLE: mov x19, x0
17 ; CHECK-APPLE-NOT: x19
19 ; CHECK-O0-LABEL: foo:
[all …]
|
| /external/llvm/test/CodeGen/X86/ |
| D | cxx_tlscc64.ll | 1 ; RUN: llc -verify-machineinstrs < %s -mtriple=x86_64-apple-darwin | FileCheck %s
2 ; TLS function were wrongly model and after fixing that, shrink-wrapping
5 ; Applying tricks on x86-64 similar to r255821.
6 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -enable-shrink-wrap=true | FileCheck %s
7 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -O0 | FileCheck %s --check-prefix=CHECK-O0
19 ; Every GPR should be saved - except rdi, rax, and rsp
20 ; CHECK-LABEL: _ZTW2sg
21 ; CHECK-NOT: pushq %r11
22 ; CHECK-NOT: pushq %r10
23 ; CHECK-NOT: pushq %r9
[all …]
|
| D | swifterror.ll | 1 ; RUN: llc -verify-machineinstrs < %s -mtriple=x86_64-apple-darwin | FileCheck --check-prefix=CHECK…
2 ; RUN: llc -verify-machineinstrs -O0 < %s -mtriple=x86_64-apple-darwin | FileCheck --check-prefix=C…
11 ; CHECK-APPLE-LABEL: foo:
12 ; CHECK-APPLE: movl $16, %edi
13 ; CHECK-APPLE: malloc
14 ; CHECK-APPLE: movb $1, 8(%rax)
15 ; CHECK-APPLE: movq %rax, %r12
17 ; CHECK-O0-LABEL: foo:
18 ; CHECK-O0: movl $16
19 ; CHECK-O0: malloc
[all …]
|
| D | swift-return.ll | 1 ; RUN: llc -verify-machineinstrs < %s -mtriple=x86_64-unknown-unknown | FileCheck %s
2 ; RUN: llc -verify-machineinstrs < %s -mtriple=x86_64-unknown-unknown -O0 | FileCheck --check-prefi…
8 ; CHECK-LABEL: test:
13 ; CHECK-O0-LABEL: test
14 ; CHECK-O0: movl %edi
15 ; CHECK-O0: callq gen
16 ; CHECK-O0: movswl %ax
17 ; CHECK-O0: movsbl %dl
18 ; CHECK-O0: addl
19 ; CHECK-O0: movw %{{.*}}, %ax
[all …]
|
| /external/llvm/test/CodeGen/SystemZ/ |
| D | swifterror.ll | 1 ; RUN: llc < %s -mtriple=s390x-linux-gnu| FileCheck %s
2 ; RUN: llc < %s -O0 -mtriple=s390x-linux-gnu | FileCheck --check-prefix=CHECK-O0 %s
11 ; CHECK-LABEL: foo:
16 ; CHECK-O0-LABEL: foo:
17 ; CHECK-O0: lghi %r2, 16
18 ; CHECK-O0: brasl %r14, malloc
19 ; CHECK-O0: lgr %r[[REG1:[0-9]+]], %r2
20 ; CHECK-O0: mvi 8(%r2), 1
21 ; CHECK-O0: lgr %r9, %r[[REG1]]
33 ; CHECK-LABEL: caller:
[all …]
|
| D | swift-return.ll | 1 ; RUN: llc < %s -mtriple=s390x-linux-gnu -verify-machineinstrs | FileCheck %s
2 ; RUN: llc < %s -mtriple=s390x-linux-gnu -O0 -verify-machineinstrs | FileCheck --check-prefix=CHECK…
8 ; CHECK-LABEL: test:
11 ; CHECK-DAG: lhr %r2, %r2
12 ; CHECK-DAG: lbr %[[REG1:r[0-9]+]], %r3
14 ; CHECK-O0-LABEL: test
15 ; CHECK-O0: st %r2
16 ; CHECK-O0: brasl %r14, gen
17 ; CHECK-O0-DAG: lhr %[[REG1:r[0-9]+]], %r2
18 ; CHECK-O0-DAG: lbr %[[REG2:r[0-9]+]], %r3
[all …]
|
| /external/llvm/test/CodeGen/ARM/ |
| D | swifterror.ll | 1 ; RUN: llc -verify-machineinstrs < %s -mtriple=armv7-apple-ios | FileCheck --check-prefix=CHECK-APP…
2 ; RUN: llc -verify-machineinstrs -O0 < %s -mtriple=armv7-apple-ios | FileCheck --check-prefix=CHECK…
12 ; CHECK-APPLE-LABEL: foo:
13 ; CHECK-APPLE: mov r0, #16
14 ; CHECK-APPLE: malloc
15 ; CHECK-APPLE-DAG: mov [[ID:r[0-9]+]], #1
16 ; CHECK-APPLE-DAG: mov r6, r{{.*}}
17 ; CHECK-APPLE-DAG: strb [[ID]], [r{{.*}}, #8]
19 ; CHECK-O0-LABEL: foo:
20 ; CHECK-O0: mov r{{.*}}, #16
[all …]
|
| D | cxx-tlscc.ll | 1 ; RUN: llc < %s -mtriple=armv7k-apple-watchos2.0 | FileCheck %s
2 ; RUN: llc < %s -mtriple=armv7k-apple-watchos2.0 -enable-shrink-wrap=true | FileCheck %s
3 ; RUN: llc < %s -mtriple=armv7-apple-ios8.0 | FileCheck %s
4 ; RUN: llc < %s -mtriple=armv7-apple-ios8.0 -enable-shrink-wrap=true | FileCheck %s
6 ; RUN: llc < %s -mtriple=armv7k-apple-watchos2.0 -O0 | FileCheck --check-prefix=CHECK-O0 --check-pr…
7 ; RUN: llc < %s -mtriple=armv7-apple-ios8.0 -O0 | FileCheck --check-prefix=CHECK-O0 --check-prefix=…
9 ; RUN: llc < %s -mtriple=thumbv7-apple-ios8.0 | FileCheck --check-prefix=THUMB %s
25 ; Make sure Epilog does not overwrite an explicitly-handled CSR in CXX_FAST_TLS.
26 ; THUMB-LABEL: _ZTW2sg
29 ; THUMB: bne [[TH_end:.?LBB0_[0-9]+]]
[all …]
|
| D | swift-return.ll | 1 ; RUN: llc -mtriple=armv7k-apple-ios8.0 -mcpu=cortex-a7 -verify-machineinstrs < %s | FileCheck %s
2 ; RUN: llc -mtriple=armv7k-apple-ios8.0 -mcpu=cortex-a7 -verify-machineinstrs < %s -O0 | FileCheck …
4 ; RUN: llc -mtriple=armv7-apple-ios -verify-machineinstrs < %s | FileCheck %s
5 ; RUN: llc -mtriple=armv7-apple-ios -verify-machineinstrs < %s -O0 | FileCheck --check-prefix=CHECK…
9 ; CHECK-LABEL: test:
13 ; CHECK-O0-LABEL: test:
14 ; CHECK-O0: bl {{.*}}gen
15 ; CHECK-O0: sxth r0, r0
16 ; CHECK-O0: sxtb r1, r1
17 ; CHECK-O0: add r0, r0, r1
[all …]
|
| /external/XNNPACK/src/x24-transposec/gen/ |
| D | 4x1-scalar.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/x24-transposec/scalar.c.in
7 // This source code is licensed under the BSD-style license found in the
28 const size_t input_reset = 3 - round_down_po2(block_height, 4) * input_stride; in xnn_x24_transposec_ukernel__4x1_scalar()
29 const size_t output_reset = 1 * output_stride - block_height * 3; in xnn_x24_transposec_ukernel__4x1_scalar()
37 uint8_t* o0 = (uint8_t*) output; in xnn_x24_transposec_ukernel__4x1_scalar() local
41 for (; bh >= 4; bh -= 4) { in xnn_x24_transposec_ukernel__4x1_scalar()
42 o0[0] = i0[0]; in xnn_x24_transposec_ukernel__4x1_scalar()
43 o0[1] = i0[1]; in xnn_x24_transposec_ukernel__4x1_scalar()
44 o0[2] = i0[2]; in xnn_x24_transposec_ukernel__4x1_scalar()
[all …]
|
| D | 4x2-scalar.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/x24-transposec/scalar.c.in
7 // This source code is licensed under the BSD-style license found in the
28 const size_t input_reset = 6 - round_down_po2(block_height, 4) * input_stride; in xnn_x24_transposec_ukernel__4x2_scalar()
29 const size_t output_reset = 2 * output_stride - block_height * 3; in xnn_x24_transposec_ukernel__4x2_scalar()
37 uint8_t* o0 = (uint8_t*) output; in xnn_x24_transposec_ukernel__4x2_scalar() local
38 uint8_t* o1 = (uint8_t*) ((uintptr_t) o0 + output_stride); in xnn_x24_transposec_ukernel__4x2_scalar()
42 o1 = o0; in xnn_x24_transposec_ukernel__4x2_scalar()
45 for (; bh >= 4; bh -= 4) { in xnn_x24_transposec_ukernel__4x2_scalar()
59 o0[0] = i0[0]; in xnn_x24_transposec_ukernel__4x2_scalar()
[all …]
|
| D | 4x4-scalar.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/x24-transposec/scalar.c.in
7 // This source code is licensed under the BSD-style license found in the
28 const size_t input_reset = 12 - round_down_po2(block_height, 4) * input_stride; in xnn_x24_transposec_ukernel__4x4_scalar()
29 const size_t output_reset = 4 * output_stride - block_height * 3; in xnn_x24_transposec_ukernel__4x4_scalar()
37 uint8_t* o0 = (uint8_t*) output; in xnn_x24_transposec_ukernel__4x4_scalar() local
38 uint8_t* o1 = (uint8_t*) ((uintptr_t) o0 + output_stride); in xnn_x24_transposec_ukernel__4x4_scalar()
44 o1 = o0; in xnn_x24_transposec_ukernel__4x4_scalar()
47 o2 = o0; in xnn_x24_transposec_ukernel__4x4_scalar()
50 o3 = o0; in xnn_x24_transposec_ukernel__4x4_scalar()
[all …]
|
| D | 2x2-scalar.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/x24-transposec/scalar.c.in
7 // This source code is licensed under the BSD-style license found in the
28 const size_t input_reset = 6 - round_down_po2(block_height, 2) * input_stride; in xnn_x24_transposec_ukernel__2x2_scalar()
29 const size_t output_reset = 2 * output_stride - block_height * 3; in xnn_x24_transposec_ukernel__2x2_scalar()
35 uint8_t* o0 = (uint8_t*) output; in xnn_x24_transposec_ukernel__2x2_scalar() local
36 uint8_t* o1 = (uint8_t*) ((uintptr_t) o0 + output_stride); in xnn_x24_transposec_ukernel__2x2_scalar()
40 o1 = o0; in xnn_x24_transposec_ukernel__2x2_scalar()
43 for (; bh >= 2; bh -= 2) { in xnn_x24_transposec_ukernel__2x2_scalar()
51 o0[0] = i0[0]; in xnn_x24_transposec_ukernel__2x2_scalar()
[all …]
|
| D | 2x1-scalar.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/x24-transposec/scalar.c.in
7 // This source code is licensed under the BSD-style license found in the
28 const size_t input_reset = 3 - round_down_po2(block_height, 2) * input_stride; in xnn_x24_transposec_ukernel__2x1_scalar()
29 const size_t output_reset = 1 * output_stride - block_height * 3; in xnn_x24_transposec_ukernel__2x1_scalar()
35 uint8_t* o0 = (uint8_t*) output; in xnn_x24_transposec_ukernel__2x1_scalar() local
39 for (; bh >= 2; bh -= 2) { in xnn_x24_transposec_ukernel__2x1_scalar()
40 o0[0] = i0[0]; in xnn_x24_transposec_ukernel__2x1_scalar()
41 o0[1] = i0[1]; in xnn_x24_transposec_ukernel__2x1_scalar()
42 o0[2] = i0[2]; in xnn_x24_transposec_ukernel__2x1_scalar()
[all …]
|
| D | 2x4-scalar.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/x24-transposec/scalar.c.in
7 // This source code is licensed under the BSD-style license found in the
28 const size_t input_reset = 12 - round_down_po2(block_height, 2) * input_stride; in xnn_x24_transposec_ukernel__2x4_scalar()
29 const size_t output_reset = 4 * output_stride - block_height * 3; in xnn_x24_transposec_ukernel__2x4_scalar()
35 uint8_t* o0 = (uint8_t*) output; in xnn_x24_transposec_ukernel__2x4_scalar() local
36 uint8_t* o1 = (uint8_t*) ((uintptr_t) o0 + output_stride); in xnn_x24_transposec_ukernel__2x4_scalar()
42 o1 = o0; in xnn_x24_transposec_ukernel__2x4_scalar()
45 o2 = o0; in xnn_x24_transposec_ukernel__2x4_scalar()
48 o3 = o0; in xnn_x24_transposec_ukernel__2x4_scalar()
[all …]
|
| /external/llvm/test/CodeGen/SPARC/ |
| D | basictest.ll | 1 ; RUN: llc < %s -march=sparc | FileCheck %s
6 ; CHECK-LABEL: test0:
7 ; CHECK: add %o0, 1, %o0
14 %B = xor i32 %A, -1
16 ; CHECK-LABEL: test1:
17 ; CHECK: xnor %o0, %o1, %o0
21 %A = xor i32 %X, -1
24 ; CHECK-LABEL: test2:
25 ; CHECK: xnor %o0, %o1, %o0
28 ; CHECK-LABEL: store_zero:
[all …]
|
| /external/libffi/src/sparc/ |
| D | v8.S | 1 /* -----------------------------------------------------------------------
2 v8.S - Copyright (c) 2013 The Written Word, Inc.
26 ----------------------------------------------------------------------- */
54 1: iflush %o0
63 .size C(ffi_flush_icache), . - C(ffi_flush_icache)
82 mov %i0, %o0 ! copy cif
89 and %o0, SPARC_FLAG_RET_MASK, %l0 ! save return type
90 srl %o0, SPARC_SIZEMASK_SHIFT, %l1 ! save return size
91 ld [%sp+64+4], %o0 ! load all argument registers
107 jmp %o7+(2f-0b)
[all …]
|
| D | v9.S | 1 /* -----------------------------------------------------------------------
2 v9.S - Copyright (c) 2000, 2003, 2004, 2008 Red Hat, Inc.
4 SPARC 64-bit Foreign Function Interface
25 ----------------------------------------------------------------------- */
62 mov %i0, %o0 ! copy cif
68 andcc %o0, SPARC_FLAG_FP_ARGS, %g0 ! need fp regs?
71 mov %o0, %l0 ! save flags
90 1: ldx [%sp+STACK_BIAS+128], %o0 ! load all int arg regs
103 jmp %o7+(2f-0b)
112 add %sp, STACK_BIAS-64+128+48, %l2
[all …]
|
| /external/XNNPACK/src/f32-prelu/gen/ |
| D | neon-1x16.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/f32-prelu/neon.c.in
7 // This source code is licensed under the BSD-style license found in the
32 float* o0 = output; in xnn_f32_prelu_ukernel__neon_1x16() local
34 const size_t input_increment = input_stride * 1 - channels; in xnn_f32_prelu_ukernel__neon_1x16()
35 const size_t output_increment = output_stride * 1 - channels; in xnn_f32_prelu_ukernel__neon_1x16()
41 for (; c >= 16 * sizeof(float); c -= 16 * sizeof(float)) { in xnn_f32_prelu_ukernel__neon_1x16()
66 vst1q_f32(o0, vacc0x0123); o0 += 4; in xnn_f32_prelu_ukernel__neon_1x16()
67 vst1q_f32(o0, vacc0x4567); o0 += 4; in xnn_f32_prelu_ukernel__neon_1x16()
68 vst1q_f32(o0, vacc0x89AB); o0 += 4; in xnn_f32_prelu_ukernel__neon_1x16()
[all …]
|
| D | neon-2x16.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/f32-prelu/neon.c.in
7 // This source code is licensed under the BSD-style license found in the
32 float* o0 = output; in xnn_f32_prelu_ukernel__neon_2x16() local
34 float* o1 = (float*) ((uintptr_t) o0 + output_stride); in xnn_f32_prelu_ukernel__neon_2x16()
36 const size_t input_increment = input_stride * 2 - channels; in xnn_f32_prelu_ukernel__neon_2x16()
37 const size_t output_increment = output_stride * 2 - channels; in xnn_f32_prelu_ukernel__neon_2x16()
42 o1 = o0; in xnn_f32_prelu_ukernel__neon_2x16()
47 for (; c >= 16 * sizeof(float); c -= 16 * sizeof(float)) { in xnn_f32_prelu_ukernel__neon_2x16()
88 vst1q_f32(o0, vacc0x0123); o0 += 4; in xnn_f32_prelu_ukernel__neon_2x16()
[all …]
|
| D | avx-2x16.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/f32-prelu/avx.c.in
7 // This source code is licensed under the BSD-style license found in the
18 static const int32_t mask_table[14] = {-1, -1, -1, -1, -1, -1, -1, 0, 0, 0, 0, 0, 0, 0};
34 float* o0 = output; in xnn_f32_prelu_ukernel__avx_2x16() local
36 float* o1 = (float*) ((uintptr_t) o0 + output_stride); in xnn_f32_prelu_ukernel__avx_2x16()
38 const size_t input_increment = input_stride * 2 - channels; in xnn_f32_prelu_ukernel__avx_2x16()
39 const size_t output_increment = output_stride * 2 - channels; in xnn_f32_prelu_ukernel__avx_2x16()
44 o1 = o0; in xnn_f32_prelu_ukernel__avx_2x16()
49 for (; c >= 16 * sizeof(float); c -= 16 * sizeof(float)) { in xnn_f32_prelu_ukernel__avx_2x16()
[all …]
|
| D | wasmsimd-bitselect-1x16.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/f32-prelu/wasmsimd-bitselect.c.in
7 // This source code is licensed under the BSD-style license found in the
32 float* o0 = output; in xnn_f32_prelu_ukernel__wasmsimd_bitselect_1x16() local
34 const size_t input_increment = input_stride * 1 - channels; in xnn_f32_prelu_ukernel__wasmsimd_bitselect_1x16()
35 const size_t output_increment = output_stride * 1 - channels; in xnn_f32_prelu_ukernel__wasmsimd_bitselect_1x16()
41 for (; c >= 16 * sizeof(float); c -= 16 * sizeof(float)) { in xnn_f32_prelu_ukernel__wasmsimd_bitselect_1x16()
68 wasm_v128_store(o0, vacc0x0123); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_1x16()
69 wasm_v128_store(o0 + 4, vacc0x4567); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_1x16()
70 wasm_v128_store(o0 + 8, vacc0x89AB); in xnn_f32_prelu_ukernel__wasmsimd_bitselect_1x16()
[all …]
|
| /external/XNNPACK/src/x24-transposec/ |
| D | 2x2-neon-tbl.c | 3 // This source code is licensed under the BSD-style license found in the
31 const size_t tile_wbytes_minus_4 = tile_wbytes - 4; in xnn_x24_transposec_ukernel__2x2_neon_tbl()
32 const size_t input_reset = tile_wbytes - round_down_po2(block_height, tile_height) * input_stride; in xnn_x24_transposec_ukernel__2x2_neon_tbl()
33 const size_t output_reset = tile_height * output_stride - block_height * 3; in xnn_x24_transposec_ukernel__2x2_neon_tbl()
40 uint8_t* o0 = (uint8_t*) output; in xnn_x24_transposec_ukernel__2x2_neon_tbl() local
41 uint8_t* o1 = (uint8_t*) ((uintptr_t) o0 + output_stride); in xnn_x24_transposec_ukernel__2x2_neon_tbl()
47 o1 = o0; in xnn_x24_transposec_ukernel__2x2_neon_tbl()
50 for (; bh >= 2; bh -= 2) { in xnn_x24_transposec_ukernel__2x2_neon_tbl()
59 … vst1_lane_u32((void*) o0, vreinterpret_u32_u8(vres0), 0); o0 = (uint8_t*) ((uintptr_t) o0 + 4); in xnn_x24_transposec_ukernel__2x2_neon_tbl()
61 …vst1_lane_u16((void*) o0, vreinterpret_u16_u8(vres0), 2); o0 = (uint8_t*) ((uintptr_t) o0 + tile_w… in xnn_x24_transposec_ukernel__2x2_neon_tbl()
[all …]
|
| /external/XNNPACK/src/f16-prelu/gen/ |
| D | neonfp16arith-2x16.c | 1 // Auto-generated file. Do not edit!
2 // Template: src/f16-prelu/neonfp16arith.c.in
7 // This source code is licensed under the BSD-style license found in the
32 __fp16* o0 = (__fp16*) output; in xnn_f16_prelu_ukernel__neonfp16arith_2x16() local
34 __fp16* o1 = (__fp16*) ((uintptr_t) o0 + output_stride); in xnn_f16_prelu_ukernel__neonfp16arith_2x16()
36 const size_t input_increment = input_stride * 2 - channels; in xnn_f16_prelu_ukernel__neonfp16arith_2x16()
37 const size_t output_increment = output_stride * 2 - channels; in xnn_f16_prelu_ukernel__neonfp16arith_2x16()
42 o1 = o0; in xnn_f16_prelu_ukernel__neonfp16arith_2x16()
47 for (; c >= 16 * sizeof(__fp16); c -= 16 * sizeof(__fp16)) { in xnn_f16_prelu_ukernel__neonfp16arith_2x16()
70 vst1q_f16(o0, vacc0x001234567); o0 += 8; in xnn_f16_prelu_ukernel__neonfp16arith_2x16()
[all …]
|