| /third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/gpu/cuda_impl/ |
| D | layer_norm_grad_grad_impl.cu | 193 … const T &epsilon, T *sum5, T *sum6, T *sum7, T *share_mem, in InputThreadReduceOuterMean() argument
222 sum6[0] += part2; in InputThreadReduceOuterMean()
231 … const half &epsilon, half *sum5, half *sum6, half *sum7, in InputThreadReduceOuterMean() argument
260 sum6[0] += part2; in InputThreadReduceOuterMean()
268 inline __device__ void InputWarpReduceOuterMean(T *sum5, T *sum6, T *sum7) { in InputWarpReduceOuterMean() argument
271 sum6[0] += __shfl_down_sync(0xffffffff, sum6[0], delta); in InputWarpReduceOuterMean()
277 inline __device__ void InputBlockReduceOuterMean(const int &col_dim, T *sum5, T *sum6, T *sum7, T *… in InputBlockReduceOuterMean() argument
284 share_mem[offset + 5] = sum6[0]; in InputBlockReduceOuterMean()
370 T sum6 = 0; in InputPropKernel() local
379 …InputThreadReduceOuterMean(row, col_dim, param_dim, epsilon, &sum5, &sum6, &sum7, share_mem.addr()… in InputPropKernel()
[all …]
|
| /third_party/cmsis/CMSIS/DSP/Source/MatrixFunctions/ |
| D | arm_mat_mult_f32.c | 549 float32_t sum0,sum1, sum2,sum3, sum4, sum5 , sum6, sum7; in arm_mat_mult_f32() local
598 sum6 = 0.0f; in arm_mat_mult_f32()
686 sum6 += vget_lane_f32(accum, 0) + vget_lane_f32(accum, 1); in arm_mat_mult_f32()
704 sum6 += *pIn1G++ * (*pIn2); in arm_mat_mult_f32()
719 *pxG++ = sum6; in arm_mat_mult_f32()
|
| /third_party/ffmpeg/libavfilter/ |
| D | vf_convolution.c | 192 int sum6 = c0[x] * 5 + c1[x] * -3 + c2[x] * -3 + in filter16_kirsch() local
202 sum6 = FFMAX(sum6, sum7); in filter16_kirsch()
204 sum4 = FFMAX(sum4, sum6); in filter16_kirsch()
295 int sum6 = c0[x] * 5 + c1[x] * -3 + c2[x] * -3 + in filter_kirsch() local
305 sum6 = FFMAX(sum6, sum7); in filter_kirsch()
307 sum4 = FFMAX(sum4, sum6); in filter_kirsch()
|
| /third_party/ffmpeg/libavcodec/mips/ |
| D | hpeldsp_msa.c | 670 v8u16 sum0, sum1, sum2, sum3, sum4, sum5, sum6, sum7; in common_hv_bil_no_rnd_8x8_msa() local
698 sum6 = add6 + add7 + 1; in common_hv_bil_no_rnd_8x8_msa()
702 SRA_4V(sum4, sum5, sum6, sum7, 2); in common_hv_bil_no_rnd_8x8_msa()
705 PCKEV_B2_SB(sum5, sum4, sum7, sum6, out0, out1); in common_hv_bil_no_rnd_8x8_msa()
|
| /third_party/mindspore/tests/st/ops/gpu/ |
| D | test_layer_norm_grad_grad_op.py | 91 sum6 = np.mean(part2, tuple(norm_axis), keepdims=True)
93 part4 = -(x - mean) * np.power(var + epsilon, -1.5) * (sum5 + sum6)
|
| /third_party/ffmpeg/libpostproc/ |
| D | postprocess_altivec_template.c | 658 vector signed int sum5, sum6, sum7, sum8, sum9; in dering_altivec() local
694 const vector signed int sump46 = vec_mergel(sum4, sum6); in dering_altivec()
|