| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8q_mopa_2VLx2VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL() function
73 const int8_t *const A; in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL()
74 const int8_t *const B; in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL()
75 const long kstride_bytes; in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL()
76 int8_t *const C; in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL()
77 const long ldcb; in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL()
78 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL()
79 int32_t min = std::numeric_limits<int8_t>::min(); in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL()
80 int32_t max = std::numeric_limits<int8_t>::max(); in sme2_interleaved_nomerge_s8q_mopa_2VLx2VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8q_mopa_1VLx4VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL() function
73 const int8_t *const A; in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL()
74 const int8_t *const B; in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL()
75 const long kstride_bytes; in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL()
76 int8_t *const C; in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL()
77 const long ldcb; in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL()
78 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL()
79 int32_t min = std::numeric_limits<int8_t>::min(); in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL()
80 int32_t max = std::numeric_limits<int8_t>::max(); in sme2_interleaved_nomerge_s8q_mopa_1VLx4VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_u8q_mopa_1VLx4VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL() argument
73 const uint8_t *const A; in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL()
74 const uint8_t *const B; in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL()
75 const long kstride_bytes; in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL()
76 uint8_t *const C; in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL()
77 const long ldcb; in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL()
78 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL()
79 int32_t min = std::numeric_limits<uint8_t>::min(); in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL()
80 int32_t max = std::numeric_limits<uint8_t>::max(); in sme2_interleaved_nomerge_u8q_mopa_1VLx4VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_u8q_mopa_4VLx1VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL() argument
73 const uint8_t *const A; in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL()
74 const uint8_t *const B; in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL()
75 const long kstride_bytes; in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL()
76 uint8_t *const C; in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL()
77 const long ldcb; in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL()
78 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL()
79 int32_t min = std::numeric_limits<uint8_t>::min(); in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL()
80 int32_t max = std::numeric_limits<uint8_t>::max(); in sme2_interleaved_nomerge_u8q_mopa_4VLx1VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_u8q_mopa_2VLx2VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL() function
73 const uint8_t *const A; in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL()
74 const uint8_t *const B; in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL()
75 const long kstride_bytes; in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL()
76 uint8_t *const C; in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL()
77 const long ldcb; in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL()
78 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL()
79 int32_t min = std::numeric_limits<uint8_t>::min(); in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL()
80 int32_t max = std::numeric_limits<uint8_t>::max(); in sme2_interleaved_nomerge_u8q_mopa_2VLx2VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8q_mopa_4VLx1VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL() argument
73 const int8_t *const A; in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL()
74 const int8_t *const B; in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL()
75 const long kstride_bytes; in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL()
76 int8_t *const C; in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL()
77 const long ldcb; in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL()
78 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL()
79 int32_t min = std::numeric_limits<int8_t>::min(); in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL()
80 int32_t max = std::numeric_limits<int8_t>::max(); in sme2_interleaved_nomerge_s8q_mopa_4VLx1VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL() function
87 const bfloat16 *const A; in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL()
88 const bfloat16 *const B; in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL()
89 const long kstride_bytes; in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL()
90 float *const C; in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL()
91 const long ldcb; in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL()
92 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL()
93 float min = -std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL()
94 float max = std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_bf16fp32_mopa_1VLx4VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_fp32_mopa_2VLx2VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL() function
87 const float *const A; in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL()
88 const float *const B; in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL()
89 const long kstride_bytes; in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL()
90 float *const C; in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL()
91 const long ldcb; in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL()
92 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL()
93 float min = -std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL()
94 float max = std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_fp32_mopa_2VLx2VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL() argument
87 const bfloat16 *const A; in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL()
88 const bfloat16 *const B; in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL()
89 const long kstride_bytes; in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL()
90 float *const C; in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL()
91 const long ldcb; in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL()
92 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL()
93 float min = -std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL()
94 float max = std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_bf16fp32_mopa_2VLx2VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_fp32_mopa_1VLx4VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL() argument
87 const float *const A; in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL()
88 const float *const B; in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL()
89 const long kstride_bytes; in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL()
90 float *const C; in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL()
91 const long ldcb; in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL()
92 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL()
93 float min = -std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL()
94 float max = std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_fp32_mopa_1VLx4VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL() argument
87 const bfloat16 *const A; in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL()
88 const bfloat16 *const B; in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL()
89 const long kstride_bytes; in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL()
90 float *const C; in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL()
91 const long ldcb; in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL()
92 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL()
93 float min = -std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL()
94 float max = std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_bf16fp32_mopa_4VLx1VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_fp32_mopa_4VLx1VL/ |
| H A D | generic.cpp | 37 struct KernelArgs in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL() struct
39 KernelArgs( in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL() function
87 const float *const A; in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL()
88 const float *const B; in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL()
89 const long kstride_bytes; in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL()
90 float *const C; in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL()
91 const long ldcb; in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL()
92 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL()
93 float min = -std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL()
94 float max = std::numeric_limits<float>::infinity(); in sme2_interleaved_nomerge_fp32_mopa_4VLx1VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL/ |
| H A D | generic.cpp | 39 struct KernelArgs in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL() struct
41 KernelArgs( in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL() function
70 const int8_t *const A; in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL()
71 const int8_t *const B; in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL()
72 const long kstride_bytes; in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL()
73 int32_t *const C; in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL()
74 const long ldcb; in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL()
75 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL()
77 const int32_t *const bias; in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL()
79 int32_t *const accumulator_buffer; in sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL/ |
| H A D | generic.cpp | 39 struct KernelArgs in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL() struct
41 KernelArgs( in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL() argument
70 const int8_t *const A; in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL()
71 const int8_t *const B; in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL()
72 const long kstride_bytes; in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL()
73 int32_t *const C; in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL()
74 const long ldcb; in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL()
75 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL()
77 const int32_t *const bias; in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL()
79 int32_t *const accumulator_buffer; in sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL/ |
| H A D | generic.cpp | 39 struct KernelArgs in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL() struct
41 KernelArgs( in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL() function
70 const int8_t *const A; in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL()
71 const int8_t *const B; in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL()
72 const long kstride_bytes; in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL()
73 int32_t *const C; in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL()
74 const long ldcb; in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL()
75 const long M, N, K, n_loops, n_tail_iters; in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL()
77 const int32_t *const bias; in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL()
79 int32_t *const accumulator_buffer; in sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sve_ffhybrid_fp32_mla_6x4VL/ |
| H A D | a64fx.cpp | 40 struct KernelArgs { in sve_ffhybrid_fp32_mla_6x4VL_a64fx() struct
41 float maxval = static_cast<float>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp32_mla_6x4VL_a64fx()
42 float minval = - static_cast<float>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp32_mla_6x4VL_a64fx()
43 unsigned int num_strings = {}; in sve_ffhybrid_fp32_mla_6x4VL_a64fx()
44 const unsigned int *string_lengths = {}; in sve_ffhybrid_fp32_mla_6x4VL_a64fx()
45 size_t N = {}; in sve_ffhybrid_fp32_mla_6x4VL_a64fx()
46 const float *B_ptr = {}; in sve_ffhybrid_fp32_mla_6x4VL_a64fx()
47 const float *cur_B_ptr = {}; in sve_ffhybrid_fp32_mla_6x4VL_a64fx()
48 size_t B_stride = {}; in sve_ffhybrid_fp32_mla_6x4VL_a64fx()
49 size_t output_offset = {}; in sve_ffhybrid_fp32_mla_6x4VL_a64fx()
[all …]
|
| H A D | generic.cpp | 40 struct KernelArgs { in sve_ffhybrid_fp32_mla_6x4VL() struct
41 float maxval = static_cast<float>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp32_mla_6x4VL()
42 float minval = - static_cast<float>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp32_mla_6x4VL()
43 unsigned int num_strings = {}; in sve_ffhybrid_fp32_mla_6x4VL()
44 const unsigned int *string_lengths = {}; in sve_ffhybrid_fp32_mla_6x4VL()
45 size_t N = {}; in sve_ffhybrid_fp32_mla_6x4VL()
46 const float *B_ptr = {}; in sve_ffhybrid_fp32_mla_6x4VL()
47 const float *cur_B_ptr = {}; in sve_ffhybrid_fp32_mla_6x4VL()
48 size_t B_stride = {}; in sve_ffhybrid_fp32_mla_6x4VL()
49 size_t output_offset = {}; in sve_ffhybrid_fp32_mla_6x4VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sve_ffhybrid_fp16_mla_6x4VL/ |
| H A D | a64fx.cpp | 40 struct KernelArgs { in sve_ffhybrid_fp16_mla_6x4VL_a64fx() struct
41 __fp16 maxval = static_cast<__fp16>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp16_mla_6x4VL_a64fx()
42 __fp16 minval = - static_cast<__fp16>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp16_mla_6x4VL_a64fx()
43 unsigned int num_strings = {}; in sve_ffhybrid_fp16_mla_6x4VL_a64fx()
44 const unsigned int *string_lengths = {}; in sve_ffhybrid_fp16_mla_6x4VL_a64fx()
45 size_t N = {}; in sve_ffhybrid_fp16_mla_6x4VL_a64fx()
46 const __fp16 *B_ptr = {}; in sve_ffhybrid_fp16_mla_6x4VL_a64fx()
47 const __fp16 *cur_B_ptr = {}; in sve_ffhybrid_fp16_mla_6x4VL_a64fx()
48 size_t B_stride = {}; in sve_ffhybrid_fp16_mla_6x4VL_a64fx()
49 size_t output_offset = {}; in sve_ffhybrid_fp16_mla_6x4VL_a64fx()
[all …]
|
| H A D | generic.cpp | 40 struct KernelArgs { in sve_ffhybrid_fp16_mla_6x4VL() struct
41 __fp16 maxval = static_cast<__fp16>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp16_mla_6x4VL()
42 __fp16 minval = - static_cast<__fp16>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp16_mla_6x4VL()
43 unsigned int num_strings = {}; in sve_ffhybrid_fp16_mla_6x4VL()
44 const unsigned int *string_lengths = {}; in sve_ffhybrid_fp16_mla_6x4VL()
45 size_t N = {}; in sve_ffhybrid_fp16_mla_6x4VL()
46 const __fp16 *B_ptr = {}; in sve_ffhybrid_fp16_mla_6x4VL()
47 const __fp16 *cur_B_ptr = {}; in sve_ffhybrid_fp16_mla_6x4VL()
48 size_t B_stride = {}; in sve_ffhybrid_fp16_mla_6x4VL()
49 size_t output_offset = {}; in sve_ffhybrid_fp16_mla_6x4VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/a64_ffhybrid_fp32bf16fp32_mmla_4x24/ |
| H A D | generic.cpp | 41 struct KernelArgs { in a64_ffhybrid_fp32bf16fp32_mmla_4x24() struct
42 float maxval = static_cast<float>(std::numeric_limits<float>::infinity()); in a64_ffhybrid_fp32bf16fp32_mmla_4x24()
43 float minval = - static_cast<float>(std::numeric_limits<float>::infinity()); in a64_ffhybrid_fp32bf16fp32_mmla_4x24()
44 unsigned int num_strings = {}; in a64_ffhybrid_fp32bf16fp32_mmla_4x24()
45 const unsigned int *string_lengths = {}; in a64_ffhybrid_fp32bf16fp32_mmla_4x24()
46 size_t N = {}; in a64_ffhybrid_fp32bf16fp32_mmla_4x24()
47 const bfloat16 *B_ptr = {}; in a64_ffhybrid_fp32bf16fp32_mmla_4x24()
48 const bfloat16 *cur_B_ptr = {}; in a64_ffhybrid_fp32bf16fp32_mmla_4x24()
49 size_t B_stride = {}; in a64_ffhybrid_fp32bf16fp32_mmla_4x24()
50 size_t output_offset = {}; in a64_ffhybrid_fp32bf16fp32_mmla_4x24()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/a64_ffhybrid_fp32_mla_6x16/ |
| H A D | generic.cpp | 40 struct KernelArgs { in a64_ffhybrid_fp32_mla_6x16() struct
41 float maxval = static_cast<float>(std::numeric_limits<float>::infinity()); in a64_ffhybrid_fp32_mla_6x16()
42 float minval = - static_cast<float>(std::numeric_limits<float>::infinity()); in a64_ffhybrid_fp32_mla_6x16()
43 unsigned int num_strings = {}; in a64_ffhybrid_fp32_mla_6x16()
44 const unsigned int *string_lengths = {}; in a64_ffhybrid_fp32_mla_6x16()
45 size_t N = {}; in a64_ffhybrid_fp32_mla_6x16()
46 const float *B_ptr = {}; in a64_ffhybrid_fp32_mla_6x16()
47 const float *cur_B_ptr = {}; in a64_ffhybrid_fp32_mla_6x16()
48 size_t B_stride = {}; in a64_ffhybrid_fp32_mla_6x16()
49 size_t output_offset = {}; in a64_ffhybrid_fp32_mla_6x16()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sve_ffhybrid_bf16fp32_mmla_6x4VL/ |
| H A D | generic.cpp | 41 struct KernelArgs { in sve_ffhybrid_bf16fp32_mmla_6x4VL() struct
42 float maxval = static_cast<float>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_bf16fp32_mmla_6x4VL()
43 float minval = - static_cast<float>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_bf16fp32_mmla_6x4VL()
44 unsigned int num_strings = {}; in sve_ffhybrid_bf16fp32_mmla_6x4VL()
45 const unsigned int *string_lengths = {}; in sve_ffhybrid_bf16fp32_mmla_6x4VL()
46 size_t N = {}; in sve_ffhybrid_bf16fp32_mmla_6x4VL()
47 const bfloat16 *B_ptr = {}; in sve_ffhybrid_bf16fp32_mmla_6x4VL()
48 const bfloat16 *cur_B_ptr = {}; in sve_ffhybrid_bf16fp32_mmla_6x4VL()
49 size_t B_stride = {}; in sve_ffhybrid_bf16fp32_mmla_6x4VL()
50 size_t output_offset = {}; in sve_ffhybrid_bf16fp32_mmla_6x4VL()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/a64_ffhybrid_bf16fp32_mmla_6x16/ |
| H A D | generic.cpp | 41 struct KernelArgs { in a64_ffhybrid_bf16fp32_mmla_6x16() struct
42 float maxval = static_cast<float>(std::numeric_limits<float>::infinity()); in a64_ffhybrid_bf16fp32_mmla_6x16()
43 float minval = - static_cast<float>(std::numeric_limits<float>::infinity()); in a64_ffhybrid_bf16fp32_mmla_6x16()
44 unsigned int num_strings = {}; in a64_ffhybrid_bf16fp32_mmla_6x16()
45 const unsigned int *string_lengths = {}; in a64_ffhybrid_bf16fp32_mmla_6x16()
46 size_t N = {}; in a64_ffhybrid_bf16fp32_mmla_6x16()
47 const bfloat16 *B_ptr = {}; in a64_ffhybrid_bf16fp32_mmla_6x16()
48 const bfloat16 *cur_B_ptr = {}; in a64_ffhybrid_bf16fp32_mmla_6x16()
49 size_t B_stride = {}; in a64_ffhybrid_bf16fp32_mmla_6x16()
50 size_t output_offset = {}; in a64_ffhybrid_bf16fp32_mmla_6x16()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/a64_ffhybrid_fp16_mla_6x32/ |
| H A D | generic.cpp | 40 struct KernelArgs { in a64_ffhybrid_fp16_mla_6x32() struct
41 __fp16 maxval = static_cast<__fp16>(std::numeric_limits<float>::infinity()); in a64_ffhybrid_fp16_mla_6x32()
42 __fp16 minval = - static_cast<__fp16>(std::numeric_limits<float>::infinity()); in a64_ffhybrid_fp16_mla_6x32()
43 unsigned int num_strings = {}; in a64_ffhybrid_fp16_mla_6x32()
44 const unsigned int *string_lengths = {}; in a64_ffhybrid_fp16_mla_6x32()
45 size_t N = {}; in a64_ffhybrid_fp16_mla_6x32()
46 const __fp16 *B_ptr = {}; in a64_ffhybrid_fp16_mla_6x32()
47 const __fp16 *cur_B_ptr = {}; in a64_ffhybrid_fp16_mla_6x32()
48 size_t B_stride = {}; in a64_ffhybrid_fp16_mla_6x32()
49 size_t output_offset = {}; in a64_ffhybrid_fp16_mla_6x32()
[all …]
|
| /aosp_15_r20/external/ComputeLibrary/src/core/NEON/kernels/arm_gemm/kernels/sve_ffhybrid_fp32bf16fp32_mmla_4x6VL/ |
| H A D | generic.cpp | 41 struct KernelArgs { in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL() struct
42 float maxval = static_cast<float>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL()
43 float minval = - static_cast<float>(std::numeric_limits<float>::infinity()); in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL()
44 unsigned int num_strings = {}; in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL()
45 const unsigned int *string_lengths = {}; in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL()
46 size_t N = {}; in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL()
47 const bfloat16 *B_ptr = {}; in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL()
48 const bfloat16 *cur_B_ptr = {}; in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL()
49 size_t B_stride = {}; in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL()
50 size_t output_offset = {}; in sve_ffhybrid_fp32bf16fp32_mmla_4x6VL()
[all …]
|