Skip to content
This repository has been archived by the owner on Feb 5, 2024. It is now read-only.

Add single-node multi-GPU adjoint support #27

Merged
merged 54 commits into from
Jun 24, 2022
Merged

Add single-node multi-GPU adjoint support #27

merged 54 commits into from
Jun 24, 2022

Conversation

mlxd
Copy link
Member

@mlxd mlxd commented Jun 1, 2022 •
edited
Loading

Before submitting

Please complete the following checklist when submitting a PR:

  • All new features must include a unit test.
    If you've fixed a bug or added code that should be tested, add a test to the
    tests directory!

  • All new functions and code must be clearly commented and documented.
    If you do make documentation changes, make sure that the docs build and
    render correctly by running make docs.

  • Ensure that the test suite passes, by running make test.

  • Add a new entry to the .github/CHANGELOG.md file, summarizing the
    change, and including a link back to the PR.

  • Ensure that code is properly formatted by running make format.

When all the above are checked, delete everything above the dashed
line and fill in the pull request template.

Context: This WIP PR will add support for batching observables over multiple GPUs during adjoint differentiation.

Description of the Change:

Benefits:

Possible Drawbacks:

Related GitHub Issues:

@mlxd mlxd changed the title [WIP] Add singl-node multi-GPU adjoint support [WIP] Add single-node multi-GPU adjoint support Jun 1, 2022
@github-actions
Copy link

github-actions bot commented Jun 1, 2022

Hello. You may have forgotten to update the changelog!
Please edit .github/CHANGELOG.md with:

  • A one-to-two sentence description of the change. You may include a small working example for new features.
  • A link back to this PR.
  • Your name (or GitHub username) in the contributors section.

Lee O'Riordan and others added 26 commits June 1, 2022 14:36
Merge branch 'multigpu_adjoint' of github.com:PennyLaneAI/pennylane-l…
302c31e 
…ightning-gpu into multigpu_adjoint
@mlxd
Data buffer class to enacapsulate memory ops
9b4e5d1
@mlxd
Move CUDA memory ownership to buffer object
af3ba8f
@mlxd
Run linter
c69f12c
@mlxd
Format fix
f7ce879
@mlxd
Move data identifier to separate class for memory locality checking
b924f26
@mlxd
Restructure DataBuffer use
ee61195
@mlxd
Fix init issues
111a758
@mlxd
Merge remote-tracking branch 'origin/inplace_install' into multigpu_a…
08877cb 
…djoint
@mlxd
Add inter and intra device copy & move ops
900d677
@mlxd
Readd local thread ID for copy assignment op
d44d92e
@mlxd
Move the const ct'd gate calls to private init
d01e99c
@mlxd
Move buffer in SV classes to unqiue_ptr for simplified extensibility
afc32af
@mlxd
Create SV copy assignment ops for cross-GPU data move
7b59ccc
@mlxd
Enable full port to DataBuffer for gate ops; avoid [] to skip default…
6407cd8 
… ct'ed elements
Merge branch 'multigpu_adjoint' of github.com:PennyLaneAI/pennylane-l…
ebfcb73 
…ightning-gpu into multigpu_adjoint
Fix inplace gate map population
b8747a1
Fix error check for data size
fbdbbc4
Fix type comparison failure checks
65734fe
Remove cout from tests
dc3db41
@mlxd
Add data buffer test structures
1de79bd
@mlxd
Update standard of C++
d7d8951
@mlxd
Format fix
ba92921
@mlxd
Add data type entity to buffer public api
09ca5e4
@mlxd
More formatting
1ca8424
@mlxd
Add copy and move assignment tests to data buffer
230e4aa
@mlxd mlxd changed the title [WIP] Add single-node multi-GPU adjoint support Add single-node multi-GPU adjoint support Jun 17, 2022
@mlxd mlxd marked this pull request as ready for review June 17, 2022 20:21
maliasadi
maliasadi suggested changes Jun 22, 2022
View reviewed changes
Copy link
Member

@maliasadi maliasadi left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Great job @mlxd. I have a few questions and suggestions.

pennylane_lightning_gpu/lightning_gpu.py Outdated Show resolved Hide resolved
setup.py Outdated Show resolved Hide resolved
pennylane_lightning_gpu/src/util/DevTag.hpp Outdated Show resolved Hide resolved
pennylane_lightning_gpu/src/util/DataBuffer.hpp Outdated Show resolved Hide resolved
pennylane_lightning_gpu/src/util/DataBuffer.hpp Outdated Show resolved Hide resolved
pennylane_lightning_gpu/src/simulator/cuGateCache.hpp Outdated Show resolved Hide resolved
pennylane_lightning_gpu/src/tests/Test_StateVectorCudaManaged_Param.cpp Show resolved Hide resolved
pennylane_lightning_gpu/src/algorithms/AdjointDiffGPU.hpp Outdated Show resolved Hide resolved
pennylane_lightning_gpu/src/algorithms/AdjointDiffGPU.hpp Show resolved Hide resolved
pennylane_lightning_gpu/src/algorithms/AdjointDiffGPU.hpp Show resolved Hide resolved
@mlxd mlxd requested a review from maliasadi June 23, 2022 16:57
trevor-vincent
trevor-vincent approved these changes Jun 23, 2022
View reviewed changes
Copy link
Contributor

@trevor-vincent trevor-vincent left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Really awesome PR Lee. I presently see no issues.

maliasadi
maliasadi approved these changes Jun 23, 2022
View reviewed changes
Copy link
Member

@maliasadi maliasadi left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the updates. Great to see the multi-GPU adjoint support in Lightning-GPU 🚀

@mlxd
Copy link
Member Author

mlxd commented Jun 23, 2022

Cheers folks. I'll do a final pass and merge tomorrow.

@mlxd @AmintorDusko
Update pennylane_lightning_gpu/src/simulator/StateVectorCudaBase.hpp
2a1e274 
Co-authored-by: Amintor Dusko <87949283+AmintorDusko@users.noreply.github.com>
@mlxd
Copy link
Member Author

mlxd commented Jun 24, 2022

Thanks again all

@mlxd mlxd merged commit b4cb5c3 into main Jun 24, 2022
@vincentmr vincentmr deleted the multigpu_adjoint branch April 25, 2023 14:13
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Reviewers

@AmintorDusko AmintorDusko AmintorDusko approved these changes

@trevor-vincent trevor-vincent trevor-vincent approved these changes

@maliasadi maliasadi maliasadi approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@mlxd @trevor-vincent @maliasadi @AmintorDusko