Commit Graph

8 Commits

Author SHA1 Message Date
Animesh Jain
173f126068 [invoke_subgraph] Preserve node meta (#150782)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/150782
Approved by: https://github.com/bdhirsh
ghstack dependencies: #150666
2025-04-08 16:57:39 +00:00
Michael Lazos
0d44a8aea1 [Hierarchical Compile] Apply deduplication after output node creation (#150306)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/150306
Approved by: https://github.com/anijain2305
ghstack dependencies: #150303, #150304, #150305
2025-04-01 20:54:18 +00:00
Michael Lazos
a2300aff94 [Hierarchical Compile] Add cycle detection function for debug (#150304)
Remove print

Pull Request resolved: https://github.com/pytorch/pytorch/pull/150304
Approved by: https://github.com/anijain2305
ghstack dependencies: #150303
2025-04-01 20:54:10 +00:00
Animesh Jain
d43c6f0033 [invoke_subgraph] Run joint passes on the hop graphs (#139325)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/139325
Approved by: https://github.com/bdhirsh, https://github.com/zou3519
ghstack dependencies: #147559
2025-03-03 23:38:14 +00:00
Michael Lazos
5c3996cab2 [Dynamo] topologically sort duplicated graph regions (#143523)
Ensure regions are topologically sorted

Pull Request resolved: https://github.com/pytorch/pytorch/pull/143523
Approved by: https://github.com/williamwen42
2024-12-19 00:43:48 +00:00
Michael Lazos
4eafbe5288 [Dynamo] Flatten slices during graph deduplication (#143522)
I encountered this issue while debugging torchtune - overall we need to make sure to not miss nodes that are slice arguments.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/143522
Approved by: https://github.com/williamwen42
2024-12-18 23:12:34 +00:00
Tom Ritchford
d25e6e623f Fix unused Python variables in test/[a-d]* (#134665)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/134665
Approved by: https://github.com/albanD
2024-12-13 22:13:12 +00:00
Michael Lazos
c31543c7ae [Dynamo] Initial deduplication pass impl (#141383)
This PR implements the deduplication pass (blocked by config currently) for dynamo where identical regions from https://github.com/pytorch/pytorch/pull/141381 are replaced with a common subgraph.

The two phases of deduplication are explained below.

**Subgraph creation**:
Subgraph creation works by taking one representative region from each region group and creating a subgraph from it, which will then be used to replace all regions in the group. This is implemented by first copying all nodes of the region to the new subgraph and then finding all inputs which are not within the region and creating placeholders for them. For the outputs, all regions in a region group need to be scanned to ensure the largest set of outputs is found, and then an output node is created which returns a tuple of all outputs.

**Graph replacement**:
To replace each region with the extracted subgraph, the node index in the region and argument index within the node's flattened args and kwargs are recorded once during subgraph creation. This allows us to determine which (external to the region) nodes and in which order these nodes are passed as inputs. For the outputs, getitem nodes are created for each output, and all nodes in the region with external outputs are replaced by the proper getitem node. Finally, all original nodes are erased (there should be no uses of these left in the graph).

Pull Request resolved: https://github.com/pytorch/pytorch/pull/141383
Approved by: https://github.com/zou3519
ghstack dependencies: #141381, #141382
2024-12-11 02:22:21 +00:00