Support per-tensor device mesh at op level #18025
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
wschin
force-pushed
the
wechi/per-tensor-mesh
branch
2 times, most recently
from
db85449 to
10bb2fd
Compare
souptc
approved these changes
Oct 24, 2023
wschin
force-pushed
the
wechi/per-tensor-mesh
branch
from
10bb2fd to
1641874
Compare
wschin
force-pushed
the
wechi/per-tensor-mesh
branch
from
1641874 to
d358409
Compare
wschin
changed the title
wschin
force-pushed
the
wechi/per-tensor-mesh
branch
2 times, most recently
from
ab6424e to
85e3087
Compare
Since Reshape may change device mesh from, e.g., [0, 1] to [0, 1, 0, 1], we can't assume since device mesh per op. Lint resolve style warnings
wschin
force-pushed
the
wechi/per-tensor-mesh
branch
from
85e3087 to
40f985e
Compare
souptc
pushed a commit
that referenced
this pull request
Oct 27, 2023
This DistributedReshape aims at supporting all sharding patterns encountered in llama 2. All patterns found are tested in `TestDistributedReshape` in `onnxruntime_test_distributed.py`. This PR implements algorithms to compute the categories below. - All inputs and outputs are replica, so it's computed like a normal Reshape. - Two-axis fusion (if any of the inputs and outputs are sharded). This category convers, e.g., `[batch, seq, hidden] -> [batch x seq, hidden]`. - Two-axis decomposition (if any of the inputs and outputs are sharded). This category convers, e.g., `[batch x seq, hidden] -> [batch, seq, hidden]`. Review guideline: - Ignore the changes in sharding_spec.h and sharding_spec.cc since they come from another PR #18025. - First, read onnxruntime_test_distributed.py to get familiar with the input/output of DistributedReshape. - Second, check the new APIs in reshape.h/reshape.cc to expose CUDA Reshape kernel to DistributedReshape. - For DistributedReshape, check its `ComputeInternal` for the 3 categories mentioned above.
kleiti
pushed a commit
to kleiti/onnxruntime
that referenced
this pull request
Mar 22, 2024
Since Reshape may change device mesh from, e.g., [0, 1] to [0, 1, 0, 1], we can't assume same device mesh per op. At each operator, we replace a single operator-level device mesh - `device_mesh_shapes` - `device_mesh_elements` with per-tensor device meshes - `input_device_mesh_shapes` (input_device_mesh_shapes[i] is the device mesh's shape for the i-th input, e.g., "[3]" for 1-D mesh with 3 devices) - `input_device_mesh_elements` (input_device_mesh_elements[i] is the flattened device mesh elements for the i-th input; e.g., "[0, 1, 2]" if you have 3 devices in that mesh) - `output_device_mesh_shapes` - `output_device_mesh_elements` Check out the change in onnxruntime_test_distributed.py for examples. It's also heavily used in microsoft#18068's `onnxruntime_test_distributed.py` change.
kleiti
pushed a commit
to kleiti/onnxruntime
that referenced
this pull request
Mar 22, 2024
This DistributedReshape aims at supporting all sharding patterns encountered in llama 2. All patterns found are tested in `TestDistributedReshape` in `onnxruntime_test_distributed.py`. This PR implements algorithms to compute the categories below. - All inputs and outputs are replica, so it's computed like a normal Reshape. - Two-axis fusion (if any of the inputs and outputs are sharded). This category convers, e.g., `[batch, seq, hidden] -> [batch x seq, hidden]`. - Two-axis decomposition (if any of the inputs and outputs are sharded). This category convers, e.g., `[batch x seq, hidden] -> [batch, seq, hidden]`. Review guideline: - Ignore the changes in sharding_spec.h and sharding_spec.cc since they come from another PR microsoft#18025. - First, read onnxruntime_test_distributed.py to get familiar with the input/output of DistributedReshape. - Second, check the new APIs in reshape.h/reshape.cc to expose CUDA Reshape kernel to DistributedReshape. - For DistributedReshape, check its `ComputeInternal` for the 3 categories mentioned above.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Since Reshape may change device mesh from, e.g., [0, 1] to [0, 1, 0, 1], we can't assume same device mesh per op. At each operator, we replace a single operator-level device mesh
device_mesh_shapesdevice_mesh_elementswith per-tensor device meshes
input_device_mesh_shapes(input_device_mesh_shapes[i] is the device mesh's shape for the i-th input, e.g., "[3]" for 1-D mesh with 3 devices)input_device_mesh_elements(input_device_mesh_elements[i] is the flattened device mesh elements for the i-th input; e.g., "[0, 1, 2]" if you have 3 devices in that mesh)output_device_mesh_shapesoutput_device_mesh_elementsCheck out the change in onnxruntime_test_distributed.py for examples. It's also heavily used in #18068's
onnxruntime_test_distributed.pychange.