This refactor was prompted by challenges handling mixed int/float
operations in C++. A previous version of this patch
added overloads for each permutation of int/float and was unwieldy
https://github.com/pytorch/pytorch/pull/87722/ This PR takes a different
approach.
The general outline of the patch is to combine the C++ types SymIntNode
and SymFloatNode into a single type, SymNode. This is type erased; we
no longer know statically at C++ if we have an int/float and have to test
it with the is_int()/is_float() virtual methods. This has a number of
knock on effects.
- We no longer have C++ classes to bind to Python. Instead, we take an
entirely new approach to our Python API, where we have a SymInt/SymFloat
class defined entirely in Python, which hold a SymNode (which corresponds
to the C++ SymNode). However, SymNode is not pybind11-bound; instead,
it lives as-is in Python, and is wrapped into C++ SymNode using PythonSymNode
when it goes into C++. This implies a userland rename.
In principle, it is also possible for the canonical implementation of SymNode
to be written in C++, and then bound to Python with pybind11 (we have
this code, although it is commented out.) However, I did not implement
this as we currently have no C++ implementations of SymNode.
Because we do return SymInt/SymFloat from C++ bindings, the C++ binding
code needs to know how to find these classes. Currently, this is done
just by manually importing torch and getting the attributes.
- Because SymInt/SymFloat are easy Python wrappers, __sym_dispatch__ now
takes SymInt/SymFloat, rather than SymNode, bringing it in line with how
__torch_dispatch__ works.
Some miscellaneous improvements:
- SymInt now has a constructor that takes SymNode. Note that this
constructor is ambiguous if you pass in a subclass of SymNode,
so an explicit downcast is necessary. This means toSymFloat/toSymInt
are no more. This is a mild optimization as it means rvalue reference
works automatically.
- We uniformly use the caster for c10::SymInt/SymFloat, rather than
going the long way via the SymIntNode/SymFloatNode.
- Removed some unnecessary toSymInt/toSymFloat calls in normalize_*
functions, pretty sure this doesn't do anything.
- guard_int is now a free function, since to guard on an int you cannot
assume the method exists. A function can handle both int and SymInt
inputs.
- We clean up the magic method definition code for SymInt/SymFloat/SymNode.
ONLY the user classes (SymInt/SymFloat) get magic methods; SymNode gets
plain methods; this is to help avoid confusion between the two types.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
cc @jansel @mlazos @soumith @voznesenskym @yanboliang @penguinwu @anijain2305
Pull Request resolved: https://github.com/pytorch/pytorch/pull/87817
Approved by: https://github.com/albanD, https://github.com/anjali411
Fixes#83973 (This is a substitute PR for https://github.com/pytorch/pytorch/pull/85024)
First of all, thanks for your invaluable contributions to PyTorch everyone!
Given how extensively `torch.cuda.is_available` is used in the PyTorch ecosystem, IMHO it's worthwhile to provide downstream libraries/frameworks/users the ability to alter the default behavior of `torch.cuda.is_available` in the context of their PyTorch usage.
I'm confident there are many current and future such use cases which could benefit from leveraging a weakened, NVML-based `torch.cuda.is_available` assessment at a downstream framework's explicit direction (thanks @malfet 81da50a972 !). Though one could always patch out the `torch.cuda.is_available` function with another implementation in a downstream library, I think this environmental variable based configuration option is more convenient and the cost to including the option is quite low.
As discussed in https://github.com/pytorch/pytorch/pull/85024#issuecomment-1261542045, this PR gates new non-default NVML-based CUDA behavior with an environmental variable (PYTORCH_NVML_BASED_CUDA_CHK) that allows a user/framework to invoke non-default, NVML-based `is_available()` assessments if desired.
Thanks again for your work everyone!
@ngimel @malfet @awaelchli
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85951
Approved by: https://github.com/ngimel
`Sparsity` as a term doesn't reflect the tools that are developed by the AO. The `torch/ao/sparsity` also has utilities for structured pruning, which internally we always referred to as just "pruning". To avoid any confusion, we renamed `Sparsity` to `Prune`. We will not be introducing the backwards compatibility, as so far this toolset was kept under silent development.
This change will reflect the changes in the documentation as well.
**TODO:**
- [ ] Change the tutorials
- [ ] Confirm no bc-breakages
- [ ] Reflect the changes in the trackers and RFC docs
Fixes #ISSUE_NUMBER
Pull Request resolved: https://github.com/pytorch/pytorch/pull/84867
Approved by: https://github.com/supriyar
This achieves the same things as https://github.com/pytorch/pytorch/pull/85908 but using backends instead of kwargs (which breaks torchscript unfortunately). This also does mean we let go of numpy compatibility BUT the wins here are that users can control what opt einsum they wanna do!
The backend allows for..well you should just read the docs:
```
.. attribute:: torch.backends.opteinsum.enabled
A :class:`bool` that controls whether opt_einsum is enabled (on by default). If so,
torch.einsum will use opt_einsum (https://optimized-einsum.readthedocs.io/en/stable/path_finding.html)
to calculate an optimal path of contraction for faster performance.
.. attribute:: torch.backends.opteinsum.strategy
A :class:`str` that specifies which strategies to try when `torch.backends.opteinsum.enabled` is True.
By default, torch.einsum will try the "auto" strategy, but the "greedy" and "optimal" strategies are
also supported. Note that the "optimal" strategy is factorial on the number of inputs as it tries all
possible paths. See more details in opt_einsum's docs
(https://optimized-einsum.readthedocs.io/en/stable/path_finding.html).
```
In trying (and failing) to land 85908, I discovered that jit script does NOT actually pull from python's version of einsum (because it cannot support variadic args nor kwargs). Thus I learned that jitted einsum does not subscribe to the new opt_einsum path calculation. Overall, this is fine since jit script is getting deprecated, but where is the best place to document this?
## Test plan:
- added tests to CI
- locally tested that trying to set the strategy to something invalid will error properly
- locally tested that tests will pass even if you don't have opt-einsum
- locally tested that setting the strategy when opt-einsum is not there will also error properly
Pull Request resolved: https://github.com/pytorch/pytorch/pull/86219
Approved by: https://github.com/soulitzer, https://github.com/malfet
# Summary
- This code creates the runtime dispatch system for choosing a performant fused SDP kernel. The only choice of fused kernel is flash_attention. It also creates python flags and a context manager that can be used to turn off and on behavior for dispatch.
- This also adds support for flash_attention with dense tensors.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85984
Approved by: https://github.com/cpuhrsch
Summary:
Added an additional roundup knob( ``roundup_bypass_threshold_mb``) to bypass rounding the requested allocation size, for allocation requests larger than the threshold value (in MB). This can help reduce the memory footprint when making large allocations that are expected to be persistent or have a large lifetime.
Differential Revision: D39868104
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85940
Approved by: https://github.com/zdevito
### Deprecation reasons:
- For most users training is on one GPU per process so these APIs are rarely used
- They added one more API dimension
- They can be expressed in a composed manner
- They are not abstracted – specific to GPU
- They caused backend APIs and implementations to have nested `std::vector<std::vector<Tensor>>`, which is hard to read or maintain
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85961
Approved by: https://github.com/XilunWu, https://github.com/H-Huang
As per request from Vision team, adding `collate` function with an extra argument of `collate_fn_map` to dispatch custom collate functions for non-collection objects and specific objects.
If the type of batch element is not present in`collate_fn_map`, it will go through all keys in the insertion order to check if the type is a subclass of the key. If so, it will invoke the corresponding collate functions.
And, `default_collate` will utilize the `collate` function with a few by default collate function for `int`, `float`, `str` and `numpy object`.
Benefit:
- Domain teams can register their own `collate` function to handle their specific type of objects
- Easier for users to extend from the `collate` function.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85748
Approved by: https://github.com/NivekT, https://github.com/pmeier
Add unit tests and docstrings corresponding to PR https://github.com/pytorch/pytorch/pull/63289
UT:
1. `test_profiler_emit_itt` in `test/test_autograd.py`. This test is merely intended to catch if emit_itt breaks on construction.
2. Test `torch.profiler.itt` functions in `test/test_itt.py`
3. Only testing that emit_itt runs when `record_shapes` option is enabled in `test/test_profiler.py`.
Docstring:
1. add ITT related info into `docs/source/bottleneck.rst`
4. add `torch.profiler.itt` functions to `docs/source/profiler.rst`
5. add docstring to `torch.profiler.itt` functions in `torch/profiler/itt.py`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/84848
Approved by: https://github.com/malfet
### Description
- This PR renames `_all_gather_base` to `all_gather_into_tensor` so that it is clearer in meaning.
- The `all_gather_into_tensor` API differs from the `all_gather` API in the output it accepts -- a single, large tensor instead of a list of tensors.
- This PR also adds deprecation warning to `_all_gather_base`.
### Issue
`_all_gather_base` was implemented in https://github.com/pytorch/pytorch/pull/33924 to avoid unnecessary flattening. There was previous effort (#82639) to merge `_all_gather_base` with the existing `all_gather` API by detecting the parameter type passed in for the output.
There are, however, two "blockers" that make the merge difficult:
(i) The merge leads to backward compatibility break. We would need to change the parameter name `tensor_list` in `all_gather` to a general name `output` that can cover both tensor and tensor list.
(ii) Recently, the `all_gather` API has added uneven tensor support, utilizing the tensor boundaries implied by the list. We are, however, not sure to add such support to the `_all_gather_base` function, because that would require users to pass in additional tensor boundary information.
In view of the above, we decided to productize `_all_gather_base` as a separate function, but with a clearer name.
### Testing
Added tests:
- `test_all_gather_into_cat_tensor_cuda` -- output form as with `torch.cat`. For example:
```
>>> tensor_in
tensor([1, 2], device='cuda:0') # Rank 0
tensor([3, 4], device='cuda:1') # Rank 1
>>> tensor_out
tensor([1, 2, 3, 4], device='cuda:0') # Rank 0
tensor([1, 2, 3, 4], device='cuda:1') # Rank 1
```
- `test_all_gather_into_stack_tensor_cuda` -- output form as with `torch.stack`. For example:
```
>>> tensor_out2
tensor([[1, 2],
[3, 4]], device='cuda:0') # Rank 0
tensor([[1, 2],
[3, 4]], device='cuda:1') # Rank 1
```
The output form is determined by the shape of the output tensor passed by the user, no flag used.
Cc @rohan-varma @mrshenli @crcrpar @ptrblck @H-Huang
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85686
Approved by: https://github.com/rohan-varma, https://github.com/crcrpar
Small rework of how the error message is formatted, introduces a distinction between the arguments and the output of kernels. Verified manually on multiple examples that the message is printed as expected.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/85008
Approved by: https://github.com/lw
