torch.use_deterministic_algorithms

torch.use_deterministic_algorithms(mode, *, warn_only=False)

Sets whether PyTorch operations must use “deterministic” algorithms. That is, algorithms which, given the same input, and when run on the same software and hardware, always produce the same output. When enabled, operations will use deterministic algorithms when available, and if only nondeterministic algorithms are available they will throw a RuntimeError when called.

Note

This setting alone is not always enough to make an application reproducible. Refer to Reproducibility for more information.

Note

torch.set_deterministic_debug_mode() offers an alternative interface for this feature.

The following normally-nondeterministic operations will act deterministically when mode=True:

• torch.nn.Conv1d when called on CUDA tensor

• torch.nn.Conv2d when called on CUDA tensor

• torch.nn.Conv3d when called on CUDA tensor

• torch.nn.ConvTranspose1d when called on CUDA tensor

• torch.nn.ConvTranspose2d when called on CUDA tensor

• torch.nn.ConvTranspose3d when called on CUDA tensor

• torch.nn.ReplicationPad1d when attempting to differentiate a CUDA tensor

• torch.nn.ReplicationPad2d when attempting to differentiate a CUDA tensor

• torch.nn.ReplicationPad3d when attempting to differentiate a CUDA tensor

• torch.bmm() when called on sparse-dense CUDA tensors

• torch.Tensor.__getitem__() when attempting to differentiate a CPU tensor and the index is a list of tensors

• torch.Tensor.index_put() with accumulate=False

• torch.Tensor.index_put() with accumulate=True when called on a CPU tensor

• torch.Tensor.put_() with accumulate=True when called on a CPU tensor

• torch.Tensor.scatter_add_() when called on a CUDA tensor

• torch.gather() when called on a CUDA tensor that requires grad

• torch.index_add() when called on CUDA tensor

• torch.index_select() when attempting to differentiate a CUDA tensor

• torch.repeat_interleave() when attempting to differentiate a CUDA tensor

• torch.Tensor.index_copy() when called on a CPU or CUDA tensor

• torch.Tensor.scatter() when src type is Tensor and called on CUDA tensor

• torch.Tensor.scatter_reduce() when reduce='sum' or reduce='mean' and called on CUDA tensor

The following normally-nondeterministic operations will throw a RuntimeError when mode=True:

• torch.nn.AvgPool3d when attempting to differentiate a CUDA tensor

• torch.nn.AdaptiveAvgPool2d when attempting to differentiate a CUDA tensor

• torch.nn.AdaptiveAvgPool3d when attempting to differentiate a CUDA tensor

• torch.nn.MaxPool3d when attempting to differentiate a CUDA tensor

• torch.nn.AdaptiveMaxPool2d when attempting to differentiate a CUDA tensor

• torch.nn.FractionalMaxPool2d when attempting to differentiate a CUDA tensor

• torch.nn.FractionalMaxPool3d when attempting to differentiate a CUDA tensor

• torch.nn.MaxUnpool1d

• torch.nn.MaxUnpool2d

• torch.nn.MaxUnpool3d

• torch.nn.functional.interpolate() when attempting to differentiate a CUDA tensor and one of the following modes is used:

  • linear

  • bilinear

  • bicubic

  • trilinear

• torch.nn.ReflectionPad1d when attempting to differentiate a CUDA tensor

• torch.nn.ReflectionPad2d when attempting to differentiate a CUDA tensor

• torch.nn.ReflectionPad3d when attempting to differentiate a CUDA tensor

• torch.nn.NLLLoss when called on a CUDA tensor

• torch.nn.CTCLoss when attempting to differentiate a CUDA tensor

• torch.nn.EmbeddingBag when attempting to differentiate a CUDA tensor when mode='max'

• torch.Tensor.put_() when accumulate=False

• torch.Tensor.put_() when accumulate=True and called on a CUDA tensor

• torch.histc() when called on a CUDA tensor

• torch.bincount() when called on a CUDA tensor and weights tensor is given

• torch.median() with indices output when called on a CUDA tensor

• torch.nn.functional.grid_sample() when attempting to differentiate a CUDA tensor

• torch.cumsum() when called on a CUDA tensor when dtype is floating point or complex

• torch.Tensor.scatter_reduce() when reduce='prod' and called on CUDA tensor

• torch.Tensor.resize_() when called with a quantized tensor

In addition, several operations fill uninitialized memory when this setting is turned on and when torch.utils.deterministic.fill_uninitialized_memory is turned on. See the documentation for that attribute for more information.

Note that deterministic operations tend to have worse performance than nondeterministic operations.

When this setting is turned on, the Inductor deterministic mode is also tuned on automatically. In deterministic mode, Inductor would avoid doing on device benchmarking that affect numerics. This includes:

• don’t pad matmul input shapes. Without enabling deterministic mode, Inductor would do benchmarking to check if padding matmul shape is beneficial.

• don’t autotune templates. Inductor has templates for kernels like matmul/conv/attention. Without enabling deterministic mode, Inductor would do autotuning to pick the best configs for those templates and adopt it if it’s faster than the kernel in eager mode. In deterministic mode, we pick the eager kernel.

• don’t autotune triton configs for reduction. Reduction numerics are very sensitive to triton configs. In deterministic mode, Inductor will use some heuristics to pick the most promising configs rather than do autotuning.

• Skip autotuning for reduction in coordinate descent tuning.

• Don’t benchmarking for the computation/communication reordering pass

• Disable the feature that dynamically scale down RBLOCK triton config for higher occupancy.

Note

This flag does not detect or prevent nondeterministic behavior caused by calling an inplace operation on a tensor with an internal memory overlap or by giving such a tensor as the out argument for an operation. In these cases, multiple writes of different data may target a single memory location, and the order of writes is not guaranteed.

Parameters

mode (bool) – If True, makes potentially nondeterministic operations switch to a deterministic algorithm or throw a runtime error. If False, allows nondeterministic operations.

Keyword Arguments

warn_only (bool, optional) – If True, operations that do not have a deterministic implementation will throw a warning instead of an error. Default: False

Example:

>>> torch.use_deterministic_algorithms(True)

# Backward mode nondeterministic error
>>> torch.nn.AvgPool3d(1)(torch.randn(3, 4, 5, 6, requires_grad=True).cuda()).sum().backward()
...
RuntimeError: avg_pool3d_backward_cuda does not have a deterministic implementation...