| /external/pytorch/test/distributed/_tensor/ |
| D | test_attention.py | 150 c10d_functional.all_to_all_single: self.world_size * 3
246 c10d_functional.all_to_all_single: (self.world_size - 1) * num_layers,
255 c10d_functional.all_to_all_single: (self.world_size * 2 - 1)
295 c10d_functional.all_to_all_single: (self.world_size - 1)
305 c10d_functional.all_to_all_single: (self.world_size * 2 - 1)
|
| /external/pytorch/torch/distributed/algorithms/_quantization/ |
| D | quantization.py | 133 elif func == dist.all_to_all_single:
140 dist.all_to_all_single(
|
| /external/pytorch/torch/distributed/ |
| D | _functional_collectives.py | 444 def all_to_all_single( function
483 tensor = torch.ops._c10d_functional.all_to_all_single( # type: ignore[attr-defined]
520 tensor = torch.ops._c10d_functional_autograd.all_to_all_single( # type: ignore[attr-defined]
557 return all_to_all_single(self, output_split_sizes, input_split_sizes, group, tag)
1088 all_to_all_single(
1135 all_to_all_single as legacy_all_to_all_single,
|
| D | _functional_collectives_impl.py | 108 return torch.ops._c10d_functional.all_to_all_single(
|
| D | distributed_c10d.py | 3880 def all_to_all_single( function
4964 all_to_all_single,
|
| /external/pytorch/torch/distributed/_shard/sharded_tensor/ |
| D | reshard.py | 14 from torch.distributed.nn.functional import all_to_all, all_to_all_single
140 local_shard = all_to_all_single(
|
| /external/pytorch/torch/distributed/_shard/sharding_spec/chunk_sharding_spec_ops/ |
| D | _common.py | 17 all_to_all_single,
197 output = all_to_all_single(
|
| /external/pytorch/torch/distributed/tensor/debug/ |
| D | _comm_mode.py | 36 funcol_native.all_to_all_single: funcol_py.all_to_all_single,
41 funcol_autograd.all_to_all_single: funcol_py.all_to_all_single,
|
| /external/pytorch/test/distributed/ |
| D | test_multi_threaded_pg.py | 111 dist.all_to_all_single(out, send, sizes, sizes)
122 dist.all_to_all_single(out, send, sizes, sizes)
132 dist.all_to_all_single(out, send)
|
| D | test_c10d_functional_native.py | 16 all_to_all_single,
359 output = torch.ops._c10d_functional.all_to_all_single(
375 output = all_to_all_single(
783 output = funcol.all_to_all_single(
|
| D | test_fake_pg.py | 148 dist.all_to_all_single(out_tensor, in_tensor, output_split, input_split)
|
| D | test_functional_api.py | 524 y = ft_c.all_to_all_single(
542 y = ft_c.all_to_all_single(
558 y = ft_c.all_to_all_single(
|
| D | test_inductor_collectives.py | 413 a2a = torch.ops.c10d_functional.all_to_all_single(
469 a2a = torch.ops.c10d_functional.all_to_all_single(
821 torch.distributed.all_to_all_single(output, input, group=pg)
|
| D | test_c10d_spawn.py | 241 y = torch.distributed.nn.all_to_all_single(
|
| D | test_c10d_common.py | 1320 dist.all_to_all_single(tensor_h, tensor)
1924 dist.all_to_all_single(output_tensor, input_tensor)
|
| /external/pytorch/torch/csrc/distributed/c10d/ |
| D | Functional.cpp | 257 at::Tensor all_to_all_single( in all_to_all_single() function
375 c10::DispatchKey::CompositeExplicitAutograd, ::all_to_all_single), in TORCH_LIBRARY()
415 .typed<decltype(all_to_all_single)>() in forward()
434 .typed<decltype(all_to_all_single)>() in backward()
|
| /external/pytorch/torch/distributed/nn/ |
| D | functional.py | 174 def all_to_all_single( function
416 dist.all_to_all_single(
|
| /external/pytorch/test/distributed/_tensor/debug/ |
| D | test_comm_mode.py | 218 dist.all_to_all_single(inp, inp)
|
| /external/pytorch/test/distributed/algorithms/quantization/ |
| D | test_quantization.py | 300 dist.all_to_all_single, qtype, quant_loss=None
|
| /external/pytorch/docs/source/ |
| D | distributed.rst | 496 .. autofunction:: all_to_all_single
|
| /external/pytorch/torch/_inductor/ |
| D | lowering.py | 6411 @register_lowering(_c10d_functional.all_to_all_single)
6415 _c10d_functional.all_to_all_single.default,
|
| /external/pytorch/torch/testing/_internal/distributed/ |
| D | distributed_test.py | 3776 dist.all_to_all_single,
3801 dist.all_to_all_single(
|