| /external/gemmlowp/meta/ |
| D | single_thread_gemm.h | 2492 std::int32_t result_stride) { in mul_1x8_1x8_int32_rhsadd() argument
2528 : [count] "+r"(count), [result_stride] "+r"(result_stride), in mul_1x8_1x8_int32_rhsadd()
2537 std::int32_t result_stride) { in mul_1x8_2x8_int32_rhsadd() argument
2577 : [count] "+r"(count), [result_stride] "+r"(result_stride), in mul_1x8_2x8_int32_rhsadd()
2587 std::int32_t result_stride) { in mul_1x8_3x8_int32_rhsadd() argument
2637 : [count] "+r"(count), [result_stride] "+r"(result_stride), in mul_1x8_3x8_int32_rhsadd()
2647 std::int32_t result_stride) { in mul_2x8_1x8_int32_rhsadd() argument
2690 : [count] "+r"(count), [result_stride] "+r"(result_stride), in mul_2x8_1x8_int32_rhsadd()
2700 std::int32_t result_stride) { in mul_2x8_2x8_int32_rhsadd() argument
2751 : [count] "+r"(count), [result_stride] "+r"(result_stride), in mul_2x8_2x8_int32_rhsadd()
[all …]
|
| D | multi_thread_gemm.h | 36 std::int32_t result_stride, const F& operation) { in CacheFriendlyMatrixMatrix() argument
46 result + i * optimal_n, result_stride); in CacheFriendlyMatrixMatrix()
51 result + chunks_count_less_one * optimal_n, result_stride); in CacheFriendlyMatrixMatrix()
54 result, result_stride); in CacheFriendlyMatrixMatrix()
72 std::int32_t result_stride) const { in ExecuteMatrixMatrix() argument
73 CacheFriendlyMatrixMatrix(scratch, lhs, rhs, m, n, k, result, result_stride, in ExecuteMatrixMatrix()
82 std::int32_t result_stride) const { in ExecuteCacheFriendlyMatrixMatrix() argument
84 sum_offset, multiplier, shift, result, result_stride); in ExecuteCacheFriendlyMatrixMatrix()
111 std::int32_t result_stride) const { in ExecuteMatrixMatrix() argument
112 CacheFriendlyMatrixMatrix(scratch, lhs, rhs, m, n, k, result, result_stride, in ExecuteMatrixMatrix()
[all …]
|
| D | multi_thread_common.h | 49 std::int32_t result_stride; member
54 std::int32_t result_stride, const F& operation) in MetaTask()
61 result_stride(result_stride), in MetaTask()
68 result + task_rect.m_offset * result_stride + task_rect.n_offset; in Run()
70 task_rect.n, k, task_result, result_stride); in Run()
120 OUT_TYPE* result, std::int32_t result_stride, in MultiThreadedMatrixMatrix() argument
132 result_stride); in MultiThreadedMatrixMatrix()
143 task_scratch, lhs, rhs, task_rects[i], k, result, result_stride, in MultiThreadedMatrixMatrix()
151 task_scratch, lhs, rhs, task_rects.back(), k, result, result_stride, in MultiThreadedMatrixMatrix()
|
| /external/gemmlowp/eight_bit_int_gemm/ |
| D | eight_bit_int_gemm.cc | 243 std::int32_t result_stride, std::uint8_t* result) { in MetaGemmQuantized8Bit() argument
245 if (IsRowMajorOrVector(result_transpose, result_stride, m, n)) { in MetaGemmQuantized8Bit()
268 std::int32_t result_stride, float* result) { in MetaGemmFloat() argument
270 if (IsRowMajorOrVector(result_transpose, result_stride, m, n)) { in MetaGemmFloat()
|
| /external/gemmlowp/meta/generators/ |
| D | mul_Nx8_Mx8_neon.py | 242 result_stride): argument
246 result_stride)
250 result_stride)
256 0), emitter.Dereference(result_address, None), result_stride)
|