Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Temporarily disable test requring large shared memory. #1802

Merged
merged 2 commits into from
Jul 5, 2022
Merged

Temporarily disable test requring large shared memory. #1802

merged 2 commits into from
Jul 5, 2022

Conversation

shmsong
Copy link

@shmsong shmsong commented Jul 5, 2022

This PR temporarily disables FusionPersistentSoftmaxLocalSmem on devices with <70KB of shared memory.

@naoyam
Copy link
Collaborator

naoyam commented Jul 5, 2022

I'm seeing the same error with PersistentNormLocalShared:

[ RUN      ] NVFuserTest.FusionPersistentNormLocalShared_CUDA
unknown file: Failure
C++ exception with description "(dynamic_smem_size) < (available_dynamic_smem_without_reconfiguration + additional_dynamic_smem_available_through_reconfiguration) INTERNAL ASSERT FAILED at "../torch/csrc/jit/codegen/cuda/executor.cpp":615, please report a bug to PyTorch. The total shared memory allocation is larger than available memory. Dynamic size: 66048. Available size: 49136. Configured smem size: 49152. Device limit size: 65536

@shmsong
Copy link
Author

shmsong commented Jul 5, 2022

I'm seeing the same error with PersistentNormLocalShared:

[ RUN      ] NVFuserTest.FusionPersistentNormLocalShared_CUDA
unknown file: Failure
C++ exception with description "(dynamic_smem_size) < (available_dynamic_smem_without_reconfiguration + additional_dynamic_smem_available_through_reconfiguration) INTERNAL ASSERT FAILED at "../torch/csrc/jit/codegen/cuda/executor.cpp":615, please report a bug to PyTorch. The total shared memory allocation is larger than available memory. Dynamic size: 66048. Available size: 49136. Configured smem size: 49152. Device limit size: 65536

OK. We can guard that test as well.

@shmsong
Copy link
Author

shmsong commented Jul 5, 2022

These guards will be removed after #1792

naoyam
naoyam approved these changes Jul 5, 2022
View reviewed changes
Copy link
Collaborator

@naoyam naoyam left a comment

Choose a reason for hiding this comment

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

LGTM

@shmsong shmsong merged commit 37c579e into devel Jul 5, 2022
@shmsong shmsong deleted the disable_large_test branch July 5, 2022 16:32
naoyam added a commit that referenced this pull request Jul 11, 2022
@csarofeen @zasdfgbnm
@shmsong @IvanYashchuk @Sergei-Lebedev @jjsjann123 @kevinstephano @naoyam
[WIP] Playing around with grid persistence. (#1779)
1d22083 
* Refactor TransormPropagator to allow specifying a position and propagating to part of the DAG (#1775)

`MaxInfoPropagator` is renamed to `MaxInfoSpanningTree`, it now only does path-finding, and the propagation is in a separate class `MaxInfoSpanningTree::Propagator`. Same for `MaxRootDomainInfoPropagator`.

`MaxInfoSpanningTree` and `MaxRootDomainInfoSpanningTree`  now allow specifying a selector, which controls which subgraph should be included in path-finding.

`MaxRootDomainInfoSpanningTree` also gets a few new constructors for convenience to use.

`TransormPropagator` is now a subclass of `MaxInfoSpanningTree::Propagator`, so the way to use it has changed.

Now `MaxInfoSpanningTree` and `MaxRootDomainInfoSpanningTree` will store the path after generation so that the same path can be traversed multiple times. This will be useful to support use cases like new `computeAt`. Pseudo-code:
```C++
void TensorView::computeAt(TensorView tv, int pos) {
  auto ComputeAtSubgraphSelector selector(this, tv);
  MaxRootDomainInfoSpanningTree path(tv, pos, &selector);
  TransformPropagator propagator(tv, pos);
  path.traverse(&propagator);
  ComputeAtPosPropagator ca_propagator(tv, pos);
  path.traverse(&ca_propagator);
}
```

* Revert scheduling changes. Cleanup only.

* Start drafting grid persistent kernels.

* Extend mma dimension and layout checking to support strided batched matmul and tensor contractions (#1761)

Co-authored-by: Christian Sarofeen <csarofeen@nvidia.com>

* Fix FusionMaxRootDomainInfoSpanningTreePrintTwice_CUDA (#1781)

* Fix div(Val, TensorView) (#1778)

* Fix div(scalar, tensor)

* lintrunner: clang-format

* Adding sibling path for MaxInfoSpanningTree (#1776)

The sibling path is required to generate consistent replay for some cases where `MaxInfoSpanningTree` is used with a selector. For example, when the producer of a Welford is excluded from the propagation section. See test `FusionTransformPropagateSelectorSibling_CUDA` for a detailed example. Besides, since we know that siblings should be transformed exactly the same, the sibling path is a perfect next hop for preserving information.

If you want a spanning tree without a sibling path, you can override `allowSibling` as `return false` in your selector;

* Save.

* Disable register reuse across serial broadcast ops (#1787)

Disable memory aliasing for inner sharing across serial broadcast.

* Fix isIntegralType error msg (#1789)

* Transform propagator skip replay when possible (#1782)

This comment in the code describes what this PR is doing:

```C++
  // Note: [Using multiple TransformPropagators]
  // There are cases that we use multiple TransformPropagators along different
  // spanning trees with different references in the same fusion. Some of these
  // spanning trees could overlap. In cases when there are overlapping nodes,
  // TransformPropagator needs to respect the replay of others, because the
  // current TransformPropagator might not contain the most amount of
  // information on how to do the correct transformation. The logic below tells
  // TransformPropagator to skip the replay when not necessary.
```

* Output allocate patch (#1790)

Caching strides along with sizes. This is to support current expand, which introduces non-contiguous output tensor

* Add SpanningTreePrinter (#1786)

* New compute at interface (#1743)

Rewrite of the compute at pass to rely on the new propagation mechanisms.

* Fix TransformReplay::getMatchedLeafPosWithoutReplay* (#1791)

* Some further cleanup for the new computeAt interface (#1793)

Revert MaxProducerPosUpdater to old algo.

* Use TransformPropagatorWithCheck in many tests (#1795)

* validateDomain in TransformPropagator (#1796)

* InlinePropagator please don't replay (#1797)

This PR makes `InlinePropagator` just set compute-at positions. It will not replay any tensor. If you want to replay, please use `TransformPropagator` and friends to do so.

Currently, `InlinePropagator` is already asserting no replay for standard and best effort compute at. So this PR is mostly about making most inlined compute at works as well.

This PR also does a lot of cleanups to remove the word "replay" from comments and variable and function names from `InlinePropagator`.

I also cleaned up `recordReplayedPos` and `retrieveReplayedPos`, now the logic is much easier to understand.

* Coding style cleanups (#1798)

Per offline discussion with @csarofeen, this PR does many renaming for better coding style: For all propagation-related things, I am now using the names `P2C` and `C2P` instead of `CasP` and `PasC`. Because "A as B" somewhat implies we want to replay A the same as B, but "B to A" sounds more general and is a better word for this case. Also, I modified the order of function arguments to match the order in its name. For example `PasC` should have `(producer, consumer)` or `(to, from)`, but not `(consumer, producer)` or `(from, to)`, and `C2P` should have `(consumer, producer)` or `(from, to)`, but not `(producer, consumer)` or `(to, from)`.

* Add parsing support for `_to_copy` to handle AMP casts. (#1756)

1. Add support for _to_copy() to support AMP casts.
2. refactored cast, accept none for dtype
3. python tests

Co-authored-by: jjsjann123 <jiej@nvidia.com>

* MMA Rfactor support for cross-warp and cross-CTA split on K dimension (#1554)

* Indexing refactor stage 2 : Remove reference tensor in predicate indexing logic (#1784)

Co-authored-by: Christian Sarofeen <csarofeen@nvidia.com>

* More cleanup on InlinePropagator (#1800)

I just realized that `InlinePropagator` can be further simplified because it no longer replays.

Since `InlinePropagator` is no longer doing replay, it is more like a "for each" problem rather than a propagation problem:

For each tensor `tv`, if we already know what is the max position of `tv` that is mapped to the reference tensor's selected outer dimensions(stored in `mapped_reference_pos_` in the code), setting the CA position is a very local operation, and is as simple as checking `tv` itself and all its consumers to determine the inline position.

`InlinePropagator` is not completely a "for each" problem only because the computation of `mapped_reference_pos_` is a propagation problem.

This cleanup reorganizes the code of `InlinePropagator` so it is clear that `InlinePropagator` is nothing but a two-step process:
Step 1: Do a propagation to find the `mapped_reference_pos_` for all tensors.
Step 2: For each tensor, check itself and its consumers to determine the CA position.

Conceptually, I would like to split step 1 with step 2. Because this split makes these concepts decoupled. Especially, this PR makes `mapped_reference_pos_` only contain info about the reference tensor, and is independent of the CA position (Currently, this is not true for best effort and most inlined computeAt without this PR). Now, in my view, `InlinePropagator` is conceptually very simple and easy to understand.

In terms of implementation, step 1 and step 2 can be interleaved, because when we don't need to know the `mapped_reference_pos_` for `tv`'s consumer in order to compute the CA position of `tv`. So a one-pass traverse could do both step 1 and step 2 altogether.

* Temporarily disable test requring large shared memory. (#1802)

* Grouping grid allreduces across iterations (#1755)

* Extend the grouped grid reduction kernel

The kernel itself should work with an arbitrary number of inputs, but
the underlying data structure, Tuple, still explicitly needs to be
specialized for the number of values, which is currently limited to 8.

* Check siblings in getMaxPosAll (#1805)

* remove dead indexing code (#1806)

* Broadcast in dim with expand (#1794)

Fixes #1788

Added expand in broadcast_in_dim to support expanding to concrete size. Note that we are not supporting dynamic shape for concrete size at this moment.

* spam nvrtc options (#1783)

TORCH_WARN on nvrtc debug option impacting performance.

Co-authored-by: Gao, Xiang <qasdfgtyuiop@gmail.com>
Co-authored-by: S. Song <41357537+shmsong@users.noreply.github.com>
Co-authored-by: Ivan Yashchuk <IvanYashchuk@users.noreply.github.com>
Co-authored-by: Sergey Lebedev <sergeyle@nvidia.com>
Co-authored-by: jjsjann123 <jiej@nvidia.com>
Co-authored-by: Kevin Stephano <kevin.stephano@gmail.com>
Co-authored-by: Naoya Maruyama <naoyam@users.noreply.github.com>
@naoyam naoyam mentioned this pull request Jul 13, 2022
Merged
shmsong pushed a commit to shmsong/pytorch that referenced this pull request Jul 24, 2022
@jjsjann123
[NVFuser] Upstream push 0714
f686d82 
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/

Code changes includes:

- codegen improvements:
  1. Indexing refactor -> Remove reference tensor in predicate indexing logic
  2. MMA Rfactor support for cross-warp and cross-CTA split on K dimension
  3. Grouping grid allreduces across iterations
  4. Swizzle op formulation for non-affine swizzles
  5. Use scheduler_utils to cache inputs and outputs in schedulePointwise
- scheduler refactor
  1. New compute at interface
- transformation propagation refactor on MaxInfoSpanningTree
  1. Added sibling path that is required to generate consistent replay for some cases where `MaxInfoSpanningTree` is used with a selector.
  2. Optimization to skip Transform propagator
  3. SpanningTreePrinter for debugging
- parser update
  1. Fixes `div`
  2. Added `_to_copy`
  3. Broadcast in dim with expand to support expanding to concrete size
  4. Dropout prob extremal patch
- executor patch on caching strides for output allocation

Squashed commits to WAR github API
Commits that's actually in this PR from the devel branch:

```
3b87896 Fix allocation of work buffers and `fused_reduction::ParallelReduce` with unswitch (csarofeen#1818)
4cae122 schedulePointwise cleanup: - computeAt + InlinePropagator (csarofeen#1815)
3df9742 Use scheduler_utils to cache inputs and outputs in schedulePointwise (csarofeen#1811)
03180aa improve broadcast resolution (csarofeen#1792)
bee6c69 bug fix (csarofeen#1819)
4413c8f Support PYTORCH_NVFUSER_DUMP=transform_propagator (csarofeen#1812)
de6b7ca Fix negative position in InlinePropagator (csarofeen#1813)
10a996c Remove redundant check in schedulePointwise (csarofeen#1810)
acd5ed4 Swizzle op formulation for non-affine swizzles (csarofeen#1441)
3ed8330 Kernel args patch to show zero_init buffer (csarofeen#1809)
037a75a Dropout prob extremal patch (csarofeen#1804)
282c429 spam nvrtc options (csarofeen#1783)
3ba6a5f Broadcast in dim with expand (csarofeen#1794)
fd4be12 remove dead indexing code (csarofeen#1806)
fa4e6a4 Check siblings in getMaxPosAll (csarofeen#1805)
025c840 Grouping grid allreduces across iterations (csarofeen#1755)
37c579e Temporarily disable test requring large shared memory. (csarofeen#1802)
5f375d0 More cleanup on InlinePropagator (csarofeen#1800)
8d384da Indexing refactor stage 2 : Remove reference tensor in predicate indexing logic (csarofeen#1784)
f008140 MMA Rfactor support for cross-warp and cross-CTA split on K dimension (csarofeen#1554)
76b3cca Add parsing support for `_to_copy` to handle AMP casts. (csarofeen#1756)
ef04f6c Coding style cleanups (csarofeen#1798)
38c7f3c InlinePropagator please don't replay (csarofeen#1797)
3f2c263 validateDomain in TransformPropagator (csarofeen#1796)
c077085 Use TransformPropagatorWithCheck in many tests (csarofeen#1795)
d0d0908 Some further cleanup for the new computeAt interface (csarofeen#1793)
45f5203 Fix TransformReplay::getMatchedLeafPosWithoutReplay* (csarofeen#1791)
28cbaf9 New compute at interface (csarofeen#1743)
635ebfc Add SpanningTreePrinter (csarofeen#1786)
59f3c32 Output allocate patch (csarofeen#1790)
fe93bf5 Transform propagator skip replay when possible (csarofeen#1782)
ebf23a5 Fix isIntegralType error msg (csarofeen#1789)
0c82ecf Disable register reuse across serial broadcast ops (csarofeen#1787)
33a824d Adding sibling path for MaxInfoSpanningTree (csarofeen#1776)
86f46aa Fix div(Val, TensorView) (csarofeen#1778)
d3de227 Fix FusionMaxRootDomainInfoSpanningTreePrintTwice_CUDA (csarofeen#1781)
ecc7a87 Extend mma dimension and layout checking to support strided batched matmul and tensor contractions (csarofeen#1761)
```

[ghstack-poisoned]
shmsong pushed a commit to shmsong/pytorch that referenced this pull request Jul 24, 2022
@jjsjann123
Update on "[NVFuser] Upstream push 0714"
d720403 
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/

Code changes includes:

- codegen improvements:
  1. Indexing refactor -> Remove reference tensor in predicate indexing logic
  2. MMA Rfactor support for cross-warp and cross-CTA split on K dimension
  3. Grouping grid allreduces across iterations
  4. Swizzle op formulation for non-affine swizzles
  5. Use scheduler_utils to cache inputs and outputs in schedulePointwise
- scheduler refactor
  1. New compute at interface
- transformation propagation refactor on MaxInfoSpanningTree
  1. Added sibling path that is required to generate consistent replay for some cases where `MaxInfoSpanningTree` is used with a selector.
  2. Optimization to skip Transform propagator
  3. SpanningTreePrinter for debugging
- parser update
  1. Fixes `div`
  2. Added `_to_copy`
  3. Broadcast in dim with expand to support expanding to concrete size
  4. Dropout prob extremal patch
- executor patch on caching strides for output allocation

Squashed commits to WAR github API
Commits that's actually in this PR from the devel branch:

```
3b87896 Fix allocation of work buffers and `fused_reduction::ParallelReduce` with unswitch (csarofeen#1818)
4cae122 schedulePointwise cleanup: - computeAt + InlinePropagator (csarofeen#1815)
3df9742 Use scheduler_utils to cache inputs and outputs in schedulePointwise (csarofeen#1811)
03180aa improve broadcast resolution (csarofeen#1792)
bee6c69 bug fix (csarofeen#1819)
4413c8f Support PYTORCH_NVFUSER_DUMP=transform_propagator (csarofeen#1812)
de6b7ca Fix negative position in InlinePropagator (csarofeen#1813)
10a996c Remove redundant check in schedulePointwise (csarofeen#1810)
acd5ed4 Swizzle op formulation for non-affine swizzles (csarofeen#1441)
3ed8330 Kernel args patch to show zero_init buffer (csarofeen#1809)
037a75a Dropout prob extremal patch (csarofeen#1804)
282c429 spam nvrtc options (csarofeen#1783)
3ba6a5f Broadcast in dim with expand (csarofeen#1794)
fd4be12 remove dead indexing code (csarofeen#1806)
fa4e6a4 Check siblings in getMaxPosAll (csarofeen#1805)
025c840 Grouping grid allreduces across iterations (csarofeen#1755)
37c579e Temporarily disable test requring large shared memory. (csarofeen#1802)
5f375d0 More cleanup on InlinePropagator (csarofeen#1800)
8d384da Indexing refactor stage 2 : Remove reference tensor in predicate indexing logic (csarofeen#1784)
f008140 MMA Rfactor support for cross-warp and cross-CTA split on K dimension (csarofeen#1554)
76b3cca Add parsing support for `_to_copy` to handle AMP casts. (csarofeen#1756)
ef04f6c Coding style cleanups (csarofeen#1798)
38c7f3c InlinePropagator please don't replay (csarofeen#1797)
3f2c263 validateDomain in TransformPropagator (csarofeen#1796)
c077085 Use TransformPropagatorWithCheck in many tests (csarofeen#1795)
d0d0908 Some further cleanup for the new computeAt interface (csarofeen#1793)
45f5203 Fix TransformReplay::getMatchedLeafPosWithoutReplay* (csarofeen#1791)
28cbaf9 New compute at interface (csarofeen#1743)
635ebfc Add SpanningTreePrinter (csarofeen#1786)
59f3c32 Output allocate patch (csarofeen#1790)
fe93bf5 Transform propagator skip replay when possible (csarofeen#1782)
ebf23a5 Fix isIntegralType error msg (csarofeen#1789)
0c82ecf Disable register reuse across serial broadcast ops (csarofeen#1787)
33a824d Adding sibling path for MaxInfoSpanningTree (csarofeen#1776)
86f46aa Fix div(Val, TensorView) (csarofeen#1778)
d3de227 Fix FusionMaxRootDomainInfoSpanningTreePrintTwice_CUDA (csarofeen#1781)
ecc7a87 Extend mma dimension and layout checking to support strided batched matmul and tensor contractions (csarofeen#1761)
```

RUN_TORCHBENCH: nvfuser

Differential Revision: [D38043938](https://our.internmc.facebook.com/intern/diff/D38043938)

[ghstack-poisoned]
csarofeen pushed a commit that referenced this pull request Aug 4, 2022
@jjsjann123 @pytorchmergebot
[NVFuser] Upstream push 0714 (pytorch#81861)
4a000ff 
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/

Code changes includes:

- codegen improvements:
  1. Indexing refactor -> Remove reference tensor in predicate indexing logic
  2. MMA Rfactor support for cross-warp and cross-CTA split on K dimension
  3. Grouping grid allreduces across iterations
  4. Swizzle op formulation for non-affine swizzles
  5. Use scheduler_utils to cache inputs and outputs in schedulePointwise
- scheduler refactor
  1. New compute at interface
- transformation propagation refactor on MaxInfoSpanningTree
  1. Added sibling path that is required to generate consistent replay for some cases where `MaxInfoSpanningTree` is used with a selector.
  2. Optimization to skip Transform propagator
  3. SpanningTreePrinter for debugging
- parser update
  1. Fixes `div`
  2. Added `_to_copy`
  3. Broadcast in dim with expand to support expanding to concrete size
  4. Dropout prob extremal patch
- executor patch on caching strides for output allocation

Squashed commits to WAR github API
Commits that's actually in this PR from the devel branch:

```
3b87896 Fix allocation of work buffers and `fused_reduction::ParallelReduce` with unswitch (#1818)
4cae122 schedulePointwise cleanup: - computeAt + InlinePropagator (#1815)
3df9742 Use scheduler_utils to cache inputs and outputs in schedulePointwise (#1811)
03180aa improve broadcast resolution (#1792)
bee6c69 bug fix (#1819)
4413c8f Support PYTORCH_NVFUSER_DUMP=transform_propagator (#1812)
de6b7ca Fix negative position in InlinePropagator (#1813)
10a996c Remove redundant check in schedulePointwise (#1810)
acd5ed4 Swizzle op formulation for non-affine swizzles (#1441)
3ed8330 Kernel args patch to show zero_init buffer (#1809)
037a75a Dropout prob extremal patch (#1804)
282c429 spam nvrtc options (#1783)
3ba6a5f Broadcast in dim with expand (#1794)
fd4be12 remove dead indexing code (#1806)
fa4e6a4 Check siblings in getMaxPosAll (#1805)
025c840 Grouping grid allreduces across iterations (#1755)
37c579e Temporarily disable test requring large shared memory. (#1802)
5f375d0 More cleanup on InlinePropagator (#1800)
8d384da Indexing refactor stage 2 : Remove reference tensor in predicate indexing logic (#1784)
f008140 MMA Rfactor support for cross-warp and cross-CTA split on K dimension (#1554)
76b3cca Add parsing support for `_to_copy` to handle AMP casts. (#1756)
ef04f6c Coding style cleanups (#1798)
38c7f3c InlinePropagator please don't replay (#1797)
3f2c263 validateDomain in TransformPropagator (#1796)
c077085 Use TransformPropagatorWithCheck in many tests (#1795)
d0d0908 Some further cleanup for the new computeAt interface (#1793)
45f5203 Fix TransformReplay::getMatchedLeafPosWithoutReplay* (#1791)
28cbaf9 New compute at interface (#1743)
635ebfc Add SpanningTreePrinter (#1786)
59f3c32 Output allocate patch (#1790)
fe93bf5 Transform propagator skip replay when possible (#1782)
ebf23a5 Fix isIntegralType error msg (#1789)
0c82ecf Disable register reuse across serial broadcast ops (#1787)
33a824d Adding sibling path for MaxInfoSpanningTree (#1776)
86f46aa Fix div(Val, TensorView) (#1778)
d3de227 Fix FusionMaxRootDomainInfoSpanningTreePrintTwice_CUDA (#1781)
ecc7a87 Extend mma dimension and layout checking to support strided batched matmul and tensor contractions (#1761)
```

RUN_TORCHBENCH: nvfuser

Differential Revision: [D38043938](https://our.internmc.facebook.com/intern/diff/D38043938)
Pull Request resolved: pytorch#81861
Approved by: https://github.com/davidberard98
@naoyam naoyam mentioned this pull request Sep 16, 2022
Merged
1 task
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@naoyam naoyam naoyam 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.
2 participants
@shmsong @naoyam