PyTorch wraps the C++ ATen tensor library that offers a wide range of operations implemented on GPU and CPU. Pytorch/XLA is a PyTorch extension; one of its purposes is to convert PyTorch operations to XLA operations. Lowering defines a process of converting a higher-level representation to a lower-level representation. In this document, I will refer to the process of converting PyTorch operation to XLA operation as the lowering. XLA Compiler will also lower XlaOp to HLO, but that’s beyond the scope of this documentation. We will forward operations that we haven’t provided an XLA lowering yet to CPU and call ATen implementations. Operations that are forwarded to the CPU will cause a significant slowdown. We must lower all operations used in the model to achieve the best performance.
You should follow the instructions in here to install required dependencies and build pytorch and pytorch/XLA from the source. You do not need access to TPU to implement the lowering. It is recommended to experiment on a workstation and configure it to use XLA:CPU. You can configure Pytorch/XLA to use XLA:CPU by running
export PJRT_DEVICE=CPU
You can find the definition of the C++ ATen operations in native_functions.yaml. After you build Pytorch/XLA from source, you will also find our default implementation (a boxed kernel which forwards calls to PyTorch native CPU) in xla/torch_xla/csrc/aten_cpu_fallback.h/cpp
. Pytorch operations can usually be mapped to PyTorch tensor api easily. If that is not the case searching the PyTorch native implementation under PyTorch repo is recommended. The goal is to lower the PyTorch operations into a sequence of XLA operations defined in here.
All file mentioned below lives under the xla/torch_xla/csrc
folder, with the exception of codegen/xla_native_functions.yaml
xla_native_functions.yaml
contains the list of all operators that are lowered. Each operator name must directly match a pytorch operator listed in native_functions.yaml. This file serves as the interface to adding new xla operators, and is an input to PyTorch's codegen machinery. It generates the below 3 files:XLANativeFunctions.h
,RegisterXLA.cpp
, andRegisterAutogradXLA.cpp
XLANativeFunctions.h
andaten_xla_type.cpp
are entry points of PyTorch to the pytorch_xla world, and contain the manually written lowerings to XLA for each operator.XLANativeFunctions.h
is auto-generated through a combination ofxla_native_functions.yaml
and the PyTorch corenative_functions.yaml
file, and contains declarations for kernels that need to be defined inaten_xla_type.cpp
. The kernels written here need to construct 'XLATensor' using the inputat::Tensor
and other parameters. The resultingXLATensor
needs to be converted back to theat::Tensor
before returning to the PyTorch world.RegisterXLA.cpp
andRegisterAutogradXLA.cpp
are auto-generated files that register all lowerings to the PyTorch Dispatcher. They also include auto-generated wrapper implementations ofout=
andinplace
operators.aten_cpu_fallback.h/.cpp
contain our boxed fallback implementation to CPU. The boxed fallback kernel will be used if a lowering is not explicitly defined inxla_native_functions.yaml
+aten_xla_type.cpp
, and the operator is not composite.tensor_methods.h
contains theXLATensor
declarations. These declarations are usually a one to one mapping of theat::Tensor
nodes we declared inXLANativeFunctions.h
tensor_methods.cpp
contains the implementation ofXLATensor node
defined intensor_methods.h
. We constructed the correspondingir::op
from the parameter’sir::Value
and wrapped it inside aXLATensor
. Ir stands for intermediate representation.ops/
directory contains allir::ops
declaration and definition. Smaller nodes can be put inops/ops.h/.cpp
. More complicated nodes can be put into a separate file. All ops inherit fromir::ops::Node
and provide a way to lower inputir::Value
to a sequence ofXlaOp
.
Our CI runs PyTorch native python tests for every change and every day. Those tests will use XLA implementation if we provide a lowering. We usually don’t need to add additional python tests for PyTorch/XLA unless we want to verify some xla behaviors(like dynamic shape) or we skipped the pytorch native test for some reason. The python test should be added to xla/test/test_operations.py
if it is required. We also need to add CPP tests in xla/test/cpp/test_aten_xla_tensor.cpp
. This test should call PyTorch c++ API and verify our implementation yields the same result as PyTorch native implementation. We also need to verify if the xla implementation is called when the tensor is a XLA tensor by checking the aten::op
and xla::op
counters.
The process of lowering is breaking down the PyTorch operations into a sequence of XlaOp. To provide a good lowering of the PyTorch operation, one needs to have a good grasp of what XLA is capable of. Reading the XlaOp document and looking into how similar ops is lowered is the best way to achieve that. You can find a minimal Op lowering example in this pr. You can also find a slightly more complicated example with backward lowering in this pr.
We have auto-generated wrapper implementations of out=
and inplace
operators for some operators in RegisterXLA.cpp
. We only need to lower the vanilla op in this case. An example would be lerp
operator which has 6 variants in native_functions.yaml
, they are
- lerp_.Scalar
- lerp_.Tensor
- lerp.Scalar_out
- lerp.Tensor_out
- lerp.Scalar
- lerp.Tensor
and will generate function prototypes
at::Tensor lerp(const at::Tensor & self, const at::Tensor & end, const at::Scalar & weight);
at::Tensor & lerp_(at::Tensor & self, const at::Tensor & end, const at::Scalar & weight);
at::Tensor lerp(const at::Tensor & self, const at::Tensor & end, const at::Tensor & weight);
at::Tensor & lerp_out(const at::Tensor & self, const at::Tensor & end, const at::Tensor & weight, at::Tensor & out);
at::Tensor & lerp_(at::Tensor & self, const at::Tensor & end, const at::Tensor & weight);
at::Tensor & lerp_out(const at::Tensor & self, const at::Tensor & end, const at::Scalar & weight, at::Tensor & out);
in XLANativeFunctions.h
if we add all of them to the xla_native_functions.yaml
. However if we only lower lerp.Scalar
and lerp.Tensor
and check RegisterXLA.cpp
, we will see
namespace {
at::Tensor wrapper_Scalar_lerp(const at::Tensor & self, const at::Tensor & end, const at::Scalar & weight) {
// No device check
// DeviceGuard omitted
return torch_xla::lerp(self, end, weight);
}
} // anonymous namespace
at::Tensor & wrapper_Scalar_lerp_(at::Tensor & self, const at::Tensor & end, const at::Scalar & weight) {
auto wrapper_Scalar_lerp__tmp = wrapper_Scalar_lerp(self, end, weight);
at::_copy_from(wrapper_Scalar_lerp__tmp, self);
return self;
}
...
m.impl("lerp_.Scalar",
TORCH_FN(wrapper_Scalar_lerp_));
The codegen will automatically generate lowerings for lerp_.Scalar
and lerp.Scalar_out
that use our lerp.Scalar
implementation, without us having to provide an explicit lowering.
In general, if there is an operator in pytorch core that has both an out-of-place and an out= variant, it's better to write a lowering for the out-of-place variant, since you'll get a code-generated out= lowering for free.
For each node we need to pass an ir::OpKind
. Here is an (example). You can find the OpKind
definition in interned_strings.h. If the aten symbol is missing, you can submit a PR like this.