| /third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/gpu/cuda_impl/ |
| D | batchnorm_fold2_impl.cu | 87 __global__ void BatchNormFold2GradMul(const T *dout, const T *x, T *tmp_x, size_t NCHW) { in BatchNormFold2GradMul() argument
89 tmp_x[i] = dout[i] * x[i]; in BatchNormFold2GradMul()
118 …2GradReduce(const T *dout, const T *x, T *d_beta, T *tmp, T *reduce_x, T *tmp2, T *tmp_x, size_t N, in BatchNormFold2GradReduce() argument
122 …BatchNormFold2GradMul<<<GET_BLOCKS(num_count), GET_THREADS, 0, cuda_stream>>>(dout, x, tmp_x, num_… in BatchNormFold2GradReduce()
123 …BatchNormFold2GradReduce1<<<GET_BLOCKS(N * C), GET_THREADS, 0, cuda_stream>>>(dout, tmp, tmp_x, tm… in BatchNormFold2GradReduce()
128 … float *reduce_x, float *tmp2, float *tmp_x, size_t N, size_t C, size_t H,
|
| D | batchnorm_fold2_impl.cuh | 34 …2GradReduce(const T *dout, const T *x, T *d_beta, T *tmp, T *reduce_x, T *tmp2, T *tmp_x, size_t N,
|
| /third_party/mesa3d/src/gallium/frontends/nine/ |
| D | nine_ff.c | 323 struct ureg_dst tmp_x = ureg_writemask(tmp, TGSI_WRITEMASK_X); in ureg_normalize3() local
325 ureg_DP3(ureg, tmp_x, src, src); in ureg_normalize3()
326 ureg_RSQ(ureg, tmp_x, _X(tmp)); in ureg_normalize3()
589 struct ureg_dst tmp_x = ureg_writemask(tmp, TGSI_WRITEMASK_X); in nine_ff_build_vs() local
605 ureg_DP3(ureg, tmp_x, vs->aVtx, vs->aVtx); in nine_ff_build_vs()
609 ureg_MAD(ureg, tmp_x, _Y(tmp), _YYYY(cPsz2), _XXXX(cPsz2)); in nine_ff_build_vs()
610 ureg_MAD(ureg, tmp_x, _Y(tmp), _X(tmp), _WWWW(cPsz1)); in nine_ff_build_vs()
611 ureg_RSQ(ureg, tmp_x, _X(tmp)); in nine_ff_build_vs()
612 ureg_MUL(ureg, tmp_x, _X(tmp), _Z(tmp)); in nine_ff_build_vs()
613 ureg_MUL(ureg, tmp_x, _X(tmp), _WWWW(_CONST(100))); in nine_ff_build_vs()
[all …]
|
| /third_party/pixman/pixman/ |
| D | pixman-gradient-walker.c | 102 int32_t tmp_x; in gradient_walker_reset() local
104 tmp_x = 0x10000 - right_x; in gradient_walker_reset()
106 left_x = tmp_x; in gradient_walker_reset()
|
| /third_party/mesa3d/src/intel/compiler/ |
| D | brw_vec4_cse.cpp | 110 src_reg tmp_x = xs[0]; in operands_match() local
122 tmp_x.ud &= mask; in operands_match()
125 return tmp_x.equals(tmp_y); in operands_match()
|
| /third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/gpu/quant/ |
| D | batchnorm_fold2_grad_gpu_kernel.h | 68 auto *tmp_x = GetDeviceAddress<T>(workspace, 3); in Launch() local
82 …BatchNormFold2GradReduce(dout, x, d_beta, tmp, reduce_x, tmp2, tmp_x, batch_size_, channel_, heigh… in Launch()
|
| /third_party/mindspore/tests/st/auto_monad/ |
| D | test_auto_monad_gpu.py | 327 tmp_x = self.bn1(tmp_c1)
329 tmp_w = self.add(tmp_x, tmp_y)
|