| /aosp_15_r20/external/javassist/src/test/test3/ |
| H A D | Switch.java | 13 int j = 4; in foo()
14 j = 4; in foo()
15 j = 4; in foo()
16 j = 4; in foo()
17 j = 4; in foo()
18 j = 4; in foo()
19 j = 4; in foo()
20 j = 4; in foo()
21 j = 4; in foo()
22 j = 4; in foo()
[all …]
|
| /aosp_15_r20/external/rust/android-crates-io/crates/unicode-ident/tests/trie/ |
| D | trie.rs | 28 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
29 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 4, 32, 33, 34, 4, 4, 4, 4, 4,
30 35, 36, 37, 38, 39, 40, 41, 42, 4, 4, 4, 4, 4, 4, 4, 4, 43, 44, 45, 46,
31 47, 4, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 4, 61, 4, 62,
32 63, 64, 65, 66, 4, 4, 4, 4, 4, 4, 4, 4, 67, 68, 69, 70, 71, 72, 73, 74,
35 78, 78, 78, 78, 78, 78, 78, 78, 4, 4, 4, 79, 80, 81, 82, 83, 78, 78, 78,
36 78, 78, 78, 78, 78, 84, 42, 85, 86, 87, 4, 88, 89, 78, 78, 78, 78, 78, 78,
37 78, 78, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
38 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
39 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
[all …]
|
| /aosp_15_r20/external/cronet/third_party/rust/chromium_crates_io/vendor/unicode-ident-1.0.12/tests/trie/ |
| H A D | trie.rs | 28 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
29 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 4, 32, 33, 34, 4, 4, 4, 4, 4,
30 35, 36, 37, 38, 39, 40, 41, 42, 4, 4, 4, 4, 4, 4, 4, 4, 43, 44, 45, 46,
31 47, 4, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 4, 61, 4, 62,
32 63, 64, 65, 66, 4, 4, 4, 4, 4, 4, 4, 4, 67, 68, 69, 70, 71, 72, 73, 74,
35 78, 78, 78, 78, 78, 78, 78, 78, 4, 4, 4, 79, 80, 81, 82, 83, 78, 78, 78,
36 78, 78, 78, 78, 78, 84, 42, 85, 4, 86, 4, 87, 88, 78, 78, 78, 78, 78, 78,
37 78, 78, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
38 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
39 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4,
[all …]
|
| /aosp_15_r20/packages/modules/NeuralNetworks/runtime/test/generated/spec_V1_0/ |
| D | avg_pool_quant8_3.example.cpp | 21 …4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4,… in get_test_model()
25 .numberOfConsumers = 4, in get_test_model()
41 .data = TestBuffer::createFromVector<int32_t>({4}) in get_test_model()
75 .inputs = {0, 1, 1, 1, 1, 2, 2, 3, 3, 4}, in get_test_model()
112 .numberOfConsumers = 4, in get_test_model_all_inputs_as_internal()
128 .data = TestBuffer::createFromVector<int32_t>({4}) in get_test_model_all_inputs_as_internal()
168 …4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4, 6, 0, 2, 4,… in get_test_model_all_inputs_as_internal()
196 .inputs = {0, 1, 1, 1, 1, 2, 2, 3, 3, 4}, in get_test_model_all_inputs_as_internal()
|
| /aosp_15_r20/external/gemmlowp/meta/ |
| H A D | transform_kernels_arm_64.h | 41 "dup v4.4s, %w[input_range_min]\n" in Transform()
42 "dup v5.4s, %w[output_range_min]\n" in Transform()
43 "dup v6.4s, %w[input_range_offset]\n" in Transform()
44 "dup v7.4s, %w[input_range_scale]\n" in Transform()
45 "dup v8.4s, %w[one_over_output_range_scale]\n" in Transform()
46 "fsub v4.4s, v4.4s, v5.4s\n" in Transform()
52 "ld1 {v0.4s, v1.4s, v2.4s, v3.4s}, [%x[input]], #64\n" in Transform()
54 "scvtf v0.4s, v0.4s\n" in Transform()
55 "scvtf v1.4s, v1.4s\n" in Transform()
56 "scvtf v2.4s, v2.4s\n" in Transform()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/a64_smallK_hybrid_fp32_mla_6x4/ |
| H A D | generic.cpp | 37 const long loops_count = iceildiv(N, (int)4) - 1; in a64_smallK_hybrid_fp32_mla_6x4()
40 float nullbias[4]; in a64_smallK_hybrid_fp32_mla_6x4()
42 memset(nullbias, 0, (4 * sizeof(float))); in a64_smallK_hybrid_fp32_mla_6x4()
67 const uint64_t biasinc = bias ? 4*sizeof(float) : 0; in a64_smallK_hybrid_fp32_mla_6x4()
158 "fmla v26.4s, v18.4s, v0.s[0]\n" in a64_smallK_hybrid_fp32_mla_6x4()
159 "fmla v27.4s, v18.4s, v3.s[0]\n" in a64_smallK_hybrid_fp32_mla_6x4()
160 "fmla v28.4s, v18.4s, v6.s[0]\n" in a64_smallK_hybrid_fp32_mla_6x4()
161 "fmla v29.4s, v18.4s, v9.s[0]\n" in a64_smallK_hybrid_fp32_mla_6x4()
162 "fmla v30.4s, v18.4s, v12.s[0]\n" in a64_smallK_hybrid_fp32_mla_6x4()
163 "fmla v31.4s, v18.4s, v15.s[0]\n" in a64_smallK_hybrid_fp32_mla_6x4()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_conv/depthwise/kernels/a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst/ |
| H A D | generic.cpp | 50 "ld1r { v14.4s }, [x19]\n" in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
52 "ld1r { v13.4s }, [x19]\n" in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
58 "ld1r { v10.4s }, [x19]\n" in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
60 "ld1r { v9.4s }, [x19]\n" in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
62 "ld1r { v8.4s }, [x19]\n" in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
64 "ld1r { v7.4s }, [x19]\n" in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
68 "movi v16.4s, #0x0\n" in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
115 "4:" // Output channel loop: Kernel loop in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
116 "smlal v6.4s, v17.4h, v3.h[0]\n" in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
119 "smlal v5.4s, v17.4h, v3.h[1]\n" in a64_u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_conv/depthwise/kernels/a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst/ |
| H A D | generic.cpp | 50 "ld1r { v14.4s }, [x19]\n" in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
52 "ld1r { v13.4s }, [x19]\n" in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
58 "ld1r { v10.4s }, [x19]\n" in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
60 "ld1r { v9.4s }, [x19]\n" in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
62 "ld1r { v8.4s }, [x19]\n" in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
64 "ld1r { v7.4s }, [x19]\n" in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
68 "movi v16.4s, #0x0\n" in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
115 "4:" // Output channel loop: Kernel loop in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
116 "smlal v6.4s, v17.4h, v3.h[0]\n" in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
119 "smlal v5.4s, v17.4h, v3.h[1]\n" in a64_s8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_conv/depthwise/kernels/a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst/ |
| H A D | generic.cpp | 50 "ld1r { v14.4s }, [x19]\n" in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
52 "ld1r { v13.4s }, [x19]\n" in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
58 "ld1r { v10.4s }, [x19]\n" in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
60 "ld1r { v9.4s }, [x19]\n" in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
62 "ld1r { v8.4s }, [x19]\n" in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
64 "ld1r { v7.4s }, [x19]\n" in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
68 "movi v16.4s, #0x0\n" in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
115 "4:" // Output channel loop: Kernel loop in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
116 "smlal v6.4s, v17.4h, v3.h[0]\n" in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
119 "smlal v5.4s, v17.4h, v3.h[1]\n" in a64_u8s8u8q_packed_to_nhwc_generic_with_multiplier_output2x8_mla_depthfirst_impl()
[all …]
|
| /aosp_15_r20/external/executorch/extension/llm/custom_ops/spinquant/third-party/FFHT/ |
| H A D | fht_neon.c | 16 for (int j = 0; j < 4; j += 4) { in helper_float_2()
18 "LD1 {v0.4S}, [%0]\n" in helper_float_2()
19 "TRN1 v16.4S, v0.4S, v0.4S\n" in helper_float_2()
20 "FNEG v17.4S, v0.4S\n" in helper_float_2()
21 "TRN2 v17.4S, v0.4S, v17.4S\n" in helper_float_2()
22 "FADD v0.4S, v16.4S, v17.4S\n" in helper_float_2()
24 "FNEG v17.4S, v0.4S\n" in helper_float_2()
26 "FADD v0.4S, v16.4S, v17.4S\n" in helper_float_2()
27 "ST1 {v0.4S}, [%0]\n" ::"r"(buf + j) in helper_float_2()
71 helper_float_3_recursive(buf + 4, 2); in helper_float_3_recursive()
[all …]
|
| /aosp_15_r20/external/hyphenation-patterns/TeX/el/ |
| H A D | hyph-el.pat.txt | 125 4β.
126 4γ.
127 4γκ.
128 4δ.
129 4ζ.
130 4θ.
131 4κ.
132 4λ.
133 4μ.
134 4μπ.
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_conv/depthwise/kernels/a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst/ |
| H A D | generic_direct.cpp | 111 "ld1r { v15.4s }, [x23]\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
113 "ld1r { v14.4s }, [x21]\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
136 "cbz x19, 4f\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
139 "cmp x21, x19, LSL #4\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
150 "ld1 { v10.4s }, [x8]\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
155 "mov v31.16b, v13.16b\n fmla v31.4s, v8.4s, v9.4s\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
157 "mov v30.16b, v13.16b\n fmla v30.4s, v7.4s, v9.4s\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
159 "mov v29.16b, v13.16b\n fmla v29.4s, v6.4s, v9.4s\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
161 "mov v27.16b, v13.16b\n fmla v27.4s, v5.4s, v9.4s\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
162 "cmp x21, x19, LSL #4\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_direct_impl()
[all …]
|
| H A D | generic_indirect.cpp | 61 inptrs[4] = input_ptrs[30]; in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
65 inptrs[8] = input_ptrs[4]; in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
106 "ld1r { v15.4s }, [x20]\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
107 "ld1r { v14.4s }, [x19]\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
115 "cmp x13, x11, LSL #4\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
133 "mov v31.16b, v13.16b\n fmla v31.4s, v8.4s, v9.4s\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
136 "mov v30.16b, v13.16b\n fmla v30.4s, v7.4s, v9.4s\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
138 "mov v29.16b, v13.16b\n fmla v29.4s, v6.4s, v9.4s\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
140 "mov v27.16b, v13.16b\n fmla v27.4s, v5.4s, v9.4s\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
142 "mov v26.16b, v13.16b\n fmla v26.4s, v4.4s, v9.4s\n" in a64_fp32_nhwc_3x3_s1_output4x4_mla_depthfirst_indirect_impl()
[all …]
|
| /aosp_15_r20/external/libhevc/common/arm64/ |
| H A D | ihevc_itrans_recon_32x32.s | 125 //d5[3]= 38 d7[3]=4
128 .align 4
169 ld1 {v0.4h, v1.4h, v2.4h, v3.4h},[x14],#32
170 ld1 {v4.4h, v5.4h, v6.4h, v7.4h},[x14],#32
195 mov x20,#4
208 ld1 {v10.4h},[x0],x6
209 ld1 {v8.4h},[x0],x6
210 ld1 {v11.4h},[x0],x6
211 ld1 {v9.4h},[x0],x6
213 smull v24.4s, v8.4h, v0.h[1] //// y1 * cos1(part of b0)
[all …]
|
| /aosp_15_r20/external/llvm/test/CodeGen/PowerPC/ |
| H A D | 2007-03-30-SpillerCrash.ll | 3 define void @test(<4 x float>*, { { i16, i16, i32 } }*) {
5 …%.sub7896 = getelementptr [4 x <4 x i32>], [4 x <4 x i32>]* null, i32 0, i32 0 ; <<4 x i32>*> [#u…
6 …getelementptr [193 x [4 x <4 x float>]], [193 x [4 x <4 x float>]]* null, i32 0, i32 175, i32 3 ;…
7 …getelementptr [193 x [4 x <4 x float>]], [193 x [4 x <4 x float>]]* null, i32 0, i32 174, i32 2 ;…
8 …getelementptr [193 x [4 x <4 x float>]], [193 x [4 x <4 x float>]]* null, i32 0, i32 174, i32 3 ;…
9 …getelementptr [193 x [4 x <4 x float>]], [193 x [4 x <4 x float>]]* null, i32 0, i32 173, i32 1 ;…
10 …getelementptr [193 x [4 x <4 x float>]], [193 x [4 x <4 x float>]]* null, i32 0, i32 173, i32 2 ;…
11 …getelementptr [193 x [4 x <4 x float>]], [193 x [4 x <4 x float>]]* null, i32 0, i32 173, i32 3 ;…
12 …getelementptr [193 x [4 x <4 x float>]], [193 x [4 x <4 x float>]]* null, i32 0, i32 172, i32 1 ;…
13 …getelementptr [193 x [4 x <4 x float>]], [193 x [4 x <4 x float>]]* null, i32 0, i32 172, i32 2 ;…
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_conv/depthwise/kernels/a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst/ |
| H A D | generic.cpp | 71 "ld1r { v21.4s }, [%x[clamps]]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
72 "ld1r { v20.4s }, [x11]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
91 "fmla v12.4s, v31.4s, v0.s[0]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
93 "fmla v13.4s, v31.4s, v0.s[1]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
95 "fmla v14.4s, v31.4s, v0.s[2]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
96 "fmla v15.4s, v31.4s, v0.s[3]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
97 "fmla v16.4s, v31.4s, v2.s[0]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
98 "fmla v17.4s, v31.4s, v2.s[1]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
99 "fmla v18.4s, v31.4s, v2.s[2]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
100 "fmla v19.4s, v31.4s, v2.s[3]\n" in a64_fp32_packed_to_nhwc_5x5_s1_with_multiplier_output2x4_mla_depthfirst_impl()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/a64_smallK_hybrid_fp32_mla_8x4/ |
| H A D | generic.cpp | 37 const long loops_count = iceildiv(N, (int)4) - 1; in a64_smallK_hybrid_fp32_mla_8x4()
40 float nullbias[4]; in a64_smallK_hybrid_fp32_mla_8x4()
42 memset(nullbias, 0, (4 * sizeof(float))); in a64_smallK_hybrid_fp32_mla_8x4()
67 const uint64_t biasinc = bias ? 4*sizeof(float) : 0; in a64_smallK_hybrid_fp32_mla_8x4()
159 "fmla v24.4s, v16.4s, v0.s[0]\n" in a64_smallK_hybrid_fp32_mla_8x4()
160 "fmla v25.4s, v16.4s, v1.s[0]\n" in a64_smallK_hybrid_fp32_mla_8x4()
161 "fmla v26.4s, v16.4s, v2.s[0]\n" in a64_smallK_hybrid_fp32_mla_8x4()
162 "fmla v27.4s, v16.4s, v3.s[0]\n" in a64_smallK_hybrid_fp32_mla_8x4()
163 "fmla v28.4s, v16.4s, v4.s[0]\n" in a64_smallK_hybrid_fp32_mla_8x4()
164 "fmla v29.4s, v16.4s, v5.s[0]\n" in a64_smallK_hybrid_fp32_mla_8x4()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/transforms/ |
| H A D | a64_transpose_interleave_24_2x4_fp32bf16.hpp | 35 if (height % 4) { in a64_transpose_interleave_24_2x4_fp32bf16()
39 size_t out_stride = 24 * roundup<size_t>(height, 4) * sizeof(bfloat16); in a64_transpose_interleave_24_2x4_fp32bf16()
65 "zip1 v28.4s, v3.4s, v26.4s\n" in a64_transpose_interleave_24_2x4_fp32bf16()
67 "zip2 v12.4s, v3.4s, v26.4s\n" in a64_transpose_interleave_24_2x4_fp32bf16()
70 "zip1 v11.4s, v1.4s, v17.4s\n" in a64_transpose_interleave_24_2x4_fp32bf16()
72 "zip2 v23.4s, v1.4s, v17.4s\n" in a64_transpose_interleave_24_2x4_fp32bf16()
75 "zip1 v10.4s, v31.4s, v1.4s\n" in a64_transpose_interleave_24_2x4_fp32bf16()
77 "zip2 v17.4s, v31.4s, v1.4s\n" in a64_transpose_interleave_24_2x4_fp32bf16()
80 "zip1 v5.4s, v14.4s, v1.4s\n" in a64_transpose_interleave_24_2x4_fp32bf16()
82 "zip2 v8.4s, v14.4s, v1.4s\n" in a64_transpose_interleave_24_2x4_fp32bf16()
[all …]
|
| H A D | a64_transpose_interleave_12_2x4_fp32bf16.hpp | 35 if (height % 4) { in a64_transpose_interleave_12_2x4_fp32bf16()
39 size_t out_stride = 12 * roundup<size_t>(height, 4) * sizeof(bfloat16); in a64_transpose_interleave_12_2x4_fp32bf16()
65 "zip1 v29.4s, v12.4s, v11.4s\n" in a64_transpose_interleave_12_2x4_fp32bf16()
67 "zip2 v0.4s, v12.4s, v11.4s\n" in a64_transpose_interleave_12_2x4_fp32bf16()
70 "zip1 v23.4s, v5.4s, v17.4s\n" in a64_transpose_interleave_12_2x4_fp32bf16()
72 "zip2 v18.4s, v5.4s, v17.4s\n" in a64_transpose_interleave_12_2x4_fp32bf16()
75 "zip1 v21.4s, v25.4s, v31.4s\n" in a64_transpose_interleave_12_2x4_fp32bf16()
77 "zip2 v10.4s, v25.4s, v31.4s\n" in a64_transpose_interleave_12_2x4_fp32bf16()
80 "zip1 v13.4s, v16.4s, v1.4s\n" in a64_transpose_interleave_12_2x4_fp32bf16()
82 "zip2 v24.4s, v16.4s, v1.4s\n" in a64_transpose_interleave_12_2x4_fp32bf16()
[all …]
|
| /aosp_15_r20/external/libxaac/decoder/armv8/ |
| H A D | ixheaacd_post_twiddle.s | 57 MOV x6, #4
58 dup v10.4h, w4
65 dup v10.4h, w4
72 LDR w7, [x1], #4
73 LDR w8, [x1], #4
107 SUB x7, x7, #4
109 STR w11, [x7], #-4
111 STR w9, [x0], #4
119 ASR w3, w3, #4
139 LD4 {v0.4h, v1.4h, v2.4h, v3.4h}, [x5], x8
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/convolution/winograd/input_transforms/ |
| H A D | a64_fp32_6x6.cpp | 42 const float pcoeffs[4] = {1.0f, 2.0f, 4.0f, 5.0f}; in a64_fp32_6x6()
63 "cmp %w[n_channels], #4\n" in a64_fp32_6x6()
72 "fmla v14.4s, v9.4s, v0.s[2]\n" in a64_fp32_6x6()
76 "fmls v10.4s, v12.4s, v0.s[2]\n" in a64_fp32_6x6()
78 "fmls v14.4s, v2.4s, v0.s[3]\n" in a64_fp32_6x6()
80 "fmla v9.4s, v12.4s, v0.s[2]\n" in a64_fp32_6x6()
82 "fmls v10.4s, v2.4s, v0.s[2]\n" in a64_fp32_6x6()
86 "fmls v9.4s, v2.4s, v0.s[2]\n" in a64_fp32_6x6()
88 "fadd v10.4s, v10.4s, v4.4s\n" in a64_fp32_6x6()
90 "fmls v7.4s, v12.4s, v0.s[1]\n" in a64_fp32_6x6()
[all …]
|
| /aosp_15_r20/external/llvm/test/CodeGen/X86/ |
| H A D | 2008-07-19-movups-spills.ll | 7 @0 = external global <4 x float>, align 1 ; <<4 x float>*>:0 [#uses=2]
8 @1 = external global <4 x float>, align 1 ; <<4 x float>*>:1 [#uses=1]
9 @2 = external global <4 x float>, align 1 ; <<4 x float>*>:2 [#uses=1]
10 @3 = external global <4 x float>, align 1 ; <<4 x float>*>:3 [#uses=1]
11 @4 = external global <4 x float>, align 1 ; <<4 x float>*>:4 [#uses=1]
12 @5 = external global <4 x float>, align 1 ; <<4 x float>*>:5 [#uses=1]
13 @6 = external global <4 x float>, align 1 ; <<4 x float>*>:6 [#uses=1]
14 @7 = external global <4 x float>, align 1 ; <<4 x float>*>:7 [#uses=1]
15 @8 = external global <4 x float>, align 1 ; <<4 x float>*>:8 [#uses=1]
16 @9 = external global <4 x float>, align 1 ; <<4 x float>*>:9 [#uses=1]
[all …]
|
| /aosp_15_r20/external/tensorflow/tensorflow/lite/kernels/internal/optimized/integer_ops/ |
| H A D | depthwise_conv_hybrid_3x3_filter.h | 32 // Enable for arm64 except for the Nvidia Linux 4 Tegra (L4T) running on
139 const int64_t output_height_increment = 2 * 4 * params_ptr->output_row_size;
145 #define DEPTHWISECONV_LABEL_HEIGHT_2_WIDTH_2_LEFTOVER "4"
165 // height, 1 width) output window (4x3 input window).
192 // But we need 4 more registers to hold the output multiplier & output
197 // the first 4 values of the output_multiplier_ptr (we have 8 in total);
199 // the first 4 values of the output_shift_ptr (we have 8 in total);
200 // lastly, v28 will hold the last 4 values of output_mulitplier and v31
201 // (previously occupied by activations) will hold the last 4 values of
215 "dup v25.4s, w4\n"
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/a64_hybrid_s8qs_dot_6x16/ |
| H A D | generic.cpp | 105 "movi v8.4s, #0x0\n" in a64_hybrid_s8qs_dot_6x16()
106 "movi v9.4s, #0x0\n" in a64_hybrid_s8qs_dot_6x16()
107 "movi v10.4s, #0x0\n" in a64_hybrid_s8qs_dot_6x16()
108 "movi v11.4s, #0x0\n" in a64_hybrid_s8qs_dot_6x16()
111 "4:" // Height 1: String loop in a64_hybrid_s8qs_dot_6x16()
133 ".inst 0x4f80e0c8 // sdot v8.4s, v6.16b, v0.4b[0]\n" in a64_hybrid_s8qs_dot_6x16()
136 ".inst 0x4f80e0e9 // sdot v9.4s, v7.16b, v0.4b[0]\n" in a64_hybrid_s8qs_dot_6x16()
139 ".inst 0x4f80e0ca // sdot v10.4s, v6.16b, v0.4b[0]\n" in a64_hybrid_s8qs_dot_6x16()
142 ".inst 0x4f80e0eb // sdot v11.4s, v7.16b, v0.4b[0]\n" in a64_hybrid_s8qs_dot_6x16()
145 ".inst 0x4fa0e0c8 // sdot v8.4s, v6.16b, v0.4b[1]\n" in a64_hybrid_s8qs_dot_6x16()
[all …]
|
| H A D | a55.cpp | 105 "movi v8.4s, #0x0\n" in a64_hybrid_s8qs_dot_6x16_a55()
106 "movi v9.4s, #0x0\n" in a64_hybrid_s8qs_dot_6x16_a55()
107 "movi v10.4s, #0x0\n" in a64_hybrid_s8qs_dot_6x16_a55()
108 "movi v11.4s, #0x0\n" in a64_hybrid_s8qs_dot_6x16_a55()
111 "4:" // Height 1: String loop in a64_hybrid_s8qs_dot_6x16_a55()
133 ".inst 0x4f80e0c8 // sdot v8.4s, v6.16b, v0.4b[0]\n" in a64_hybrid_s8qs_dot_6x16_a55()
144 ".inst 0x4f80e0e9 // sdot v9.4s, v7.16b, v0.4b[0]\n" in a64_hybrid_s8qs_dot_6x16_a55()
147 ".inst 0x4f80e0ca // sdot v10.4s, v6.16b, v0.4b[0]\n" in a64_hybrid_s8qs_dot_6x16_a55()
153 ".inst 0x4f80e0eb // sdot v11.4s, v7.16b, v0.4b[0]\n" in a64_hybrid_s8qs_dot_6x16_a55()
156 ".inst 0x4fa0e0c8 // sdot v8.4s, v6.16b, v0.4b[1]\n" in a64_hybrid_s8qs_dot_6x16_a55()
[all …]
|