xref: /aosp_15_r20/external/tensorflow/tensorflow/compiler/xla/service/reduce_scatter_utils.h (revision b6fb3261f9314811a0f4371741dbb8839866f948)
1 /* Copyright 2021 The TensorFlow Authors. All Rights Reserved.
2 
3 Licensed under the Apache License, Version 2.0 (the "License");
4 you may not use this file except in compliance with the License.
5 You may obtain a copy of the License at
6 
7     http://www.apache.org/licenses/LICENSE-2.0
8 
9 Unless required by applicable law or agreed to in writing, software
10 distributed under the License is distributed on an "AS IS" BASIS,
11 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12 See the License for the specific language governing permissions and
13 limitations under the License.
14 ==============================================================================*/
15 
16 #ifndef TENSORFLOW_COMPILER_XLA_SERVICE_REDUCE_SCATTER_UTILS_H_
17 #define TENSORFLOW_COMPILER_XLA_SERVICE_REDUCE_SCATTER_UTILS_H_
18 
19 #include <functional>
20 
21 #include "tensorflow/compiler/xla/service/hlo_instructions.h"
22 
23 namespace xla {
24 
25 struct ReduceScatterSpec {
26   int64_t split_dim;
27   int64_t sharded_partitions = 1;
28   int64_t sharded_replicas = 1;
29   int64_t group_size;
30   std::vector<int64_t> original_split_dims;
31   HloInstruction* dynamic_slice;
32 };
33 
34 // Matches the given all-reduce operation to a reduce-scatter pattern.
35 std::optional<ReduceScatterSpec> MatchReduceScatter(
36     const HloAllReduceInstruction* ar, int64_t num_partitions,
37     int64_t num_replicas, bool allow_multiple_split_dims = false,
38     bool allow_intervening_reshape = false, int64_t min_rank = 1);
39 
40 // Matches the given all-reduce operation to a reduce-scatter pattern.
41 std::optional<ReduceScatterSpec> MatchReduceScatter(
42     const HloAllReduceInstruction* ar, int64_t num_partitions,
43     int64_t num_replicas, bool allow_multiple_split_dims,
44     bool allow_intervening_reshape, int64_t min_rank,
45     HloPredicate match_partition_id, HloPredicate match_replica_id);
46 
47 }  // namespace xla
48 
49 #endif  // TENSORFLOW_COMPILER_XLA_SERVICE_REDUCE_SCATTER_UTILS_H_
50