/external/ruy/ruy/ |
D | pack_avx512.cc | 63 std::int8_t* packed_ptr) { 77 packed_ptr[Layout::kCols * Layout::kRows * k + j] = packed_zero_point; 105 std::int8_t* packed_ptr, std::int32_t* sums_ptr, 257 reinterpret_cast<__m256i*>(packed_ptr + 0 * 16 * 4), r0_0); 259 reinterpret_cast<__m256i*>(packed_ptr + 2 * 16 * 4), r0_1); 261 reinterpret_cast<__m256i*>(packed_ptr + 4 * 16 * 4), r1_0); 263 reinterpret_cast<__m256i*>(packed_ptr + 6 * 16 * 4), r1_1); 265 reinterpret_cast<__m256i*>(packed_ptr + 1 * 16 * 4), r2_0); 267 reinterpret_cast<__m256i*>(packed_ptr + 3 * 16 * 4), r2_1); 269 reinterpret_cast<__m256i*>(packed_ptr + 5 * 16 * 4), r3_0); [all …]
|
D | pack_arm.cc | 38 std::int8_t* packed_ptr, std::int32_t* sums_ptr, in Pack8bitColMajorForNeon() argument 202 [packed_ptr] "+r"(packed_ptr), [sums_ptr] "+r"(sums_ptr) in Pack8bitColMajorForNeon() 239 static_assert(offsetof(Params, packed_ptr) == RUY_OFFSET_PACKED_PTR, ""); in CheckOffsetsInPackParams8bit() 262 const std::int8_t* packed_ptr = params.packed_ptr; in Pack8bitColMajorForNeon4Cols() local 467 [ packed_ptr ] "r"(packed_ptr), [ params ] "r"(¶ms) in Pack8bitColMajorForNeon4Cols() 483 const std::int8_t* packed_ptr = params.packed_ptr; in Pack8bitColMajorForNeon2Cols() local 606 [ packed_ptr ] "r"(packed_ptr), [ params ] "r"(¶ms) in Pack8bitColMajorForNeon2Cols() 633 int src_zero_point, std::int8_t* packed_ptr, in Pack8bitColMajorForNeonA55ish() argument 828 [ packed_ptr ] "+r"(packed_ptr), [ sums_ptr ] "+r"(sums_ptr) in Pack8bitColMajorForNeonA55ish() 843 int src_inc3, int src_rows, int src_zero_point, std::int8_t* packed_ptr, in Pack8bitColMajorForNeonDotprodA55ish() argument [all …]
|
D | pack_arm.h | 58 std::int8_t* packed_ptr, int packed_stride, 69 std::int8_t* packed_ptr, std::int32_t* sums_ptr, 75 int src_zero_point, std::int8_t* packed_ptr, 81 int src_zero_point, std::int8_t* packed_ptr, 86 int src_inc3, int src_rows, int src_zero_point, std::int8_t* packed_ptr, 92 int src_zero_point, std::int8_t* packed_ptr, 103 const std::int8_t* packed_ptr; 116 const std::int8_t* packed_ptr, int src_inc0, 125 params->packed_ptr = packed_ptr; 188 std::int8_t* packed_ptr = [all …]
|
D | pack_x86.h | 62 std::int8_t* packed_ptr, std::int32_t* sums_ptr); 95 std::int8_t* packed_ptr = 101 remaining_src_cols, src_matrix.layout.rows, packed_ptr, sums_ptr); 109 std::int8_t* packed_ptr, std::int32_t* sums_ptr); 142 std::int8_t* packed_ptr = 148 remaining_src_cols, src_matrix.layout.rows, packed_ptr, sums_ptr); 155 int src_rows, float* packed_ptr); 178 float* packed_ptr = 182 src_matrix.layout.rows, packed_ptr); 189 int src_rows, float* packed_ptr); [all …]
|
D | pack_avx2_fma.cc | 69 int src_rows, std::int8_t* packed_ptr, std::int32_t* sums_ptr, 272 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 0 * 8 * 4), 274 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 2 * 8 * 4), 276 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 4 * 8 * 4), 278 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 6 * 8 * 4), 280 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 1 * 8 * 4), 282 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 3 * 8 * 4), 284 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 5 * 8 * 4), 286 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 7 * 8 * 4), 343 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 0 * 8 * 4), [all …]
|
D | pack_avx.cc | 218 int src_rows, std::int8_t* packed_ptr, 414 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 0 * 8 * 4), 416 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 2 * 8 * 4), 418 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 4 * 8 * 4), 420 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 6 * 8 * 4), 422 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 1 * 8 * 4), 424 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 3 * 8 * 4), 426 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 5 * 8 * 4), 428 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 7 * 8 * 4), 485 _mm256_storeu_si256(reinterpret_cast<__m256i*>(packed_ptr + 0 * 8 * 4), [all …]
|
D | pack_common.h | 115 float* packed_ptr = packed_matrix->data + in Run() local 121 memcpy(packed_ptr, src_ptr, KernelCols * sizeof(float)); in Run() 122 packed_ptr += KernelCols * packed_matrix->layout.stride; in Run() 127 memcpy(packed_ptr, src_ptr, remaining_cols * sizeof(float)); in Run() 128 memset(packed_ptr + remaining_cols, 0, in Run()
|
/external/mesa3d/src/gallium/drivers/llvmpipe/ |
D | lp_test_format.c | 90 LLVMValueRef packed_ptr; in add_fetch_rgba_test() local 112 packed_ptr = LLVMGetParam(func, 1); in add_fetch_rgba_test() 124 packed_ptr, offset, i, j, cache); in add_fetch_rgba_test()
|