Don't call sum() on a tensor that is not summable in layer_norm (pytorch#156600)

Don't call `sum()` on a tensor that is default constructed.

Previously we could call `sum()` on a tensor that was default-contructed. That would lead to an error like this:

```
Traceback (most recent call last):
  File "/home/ahmads/.conda/envs/pt3/lib/python3.12/unittest/case.py", line 58, in testPartExecutor
    yield
  File "/home/ahmads/.conda/envs/pt3/lib/python3.12/unittest/case.py", line 634, in run
    self._callTestMethod(testMethod)
  File "/home/ahmads/.conda/envs/pt3/lib/python3.12/unittest/case.py", line 589, in _callTestMethod
    if method() is not None:
       ^^^^^^^^
  File "/home/ahmads/personal/pytorch/torch/testing/_internal/common_utils.py", line 3191, in wrapper
    method(*args, **kwargs)
  File "/home/ahmads/personal/pytorch/test/test_nn.py", line 7235, in test_layer_norm_backwards_eps
    ln_out_cuda.backward(grad_output_cuda)
  File "/home/ahmads/personal/pytorch/torch/_tensor.py", line 647, in backward
    torch.autograd.backward(
  File "/home/ahmads/personal/pytorch/torch/autograd/__init__.py", line 354, in backward
    _engine_run_backward(
  File "/home/ahmads/personal/pytorch/torch/autograd/graph.py", line 829, in _engine_run_backward
    return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: tensor does not have a device
Exception raised from device_default at /home/ahmads/personal/pytorch/c10/core/TensorImpl.h:1265 (most recent call first):
C++ CapturedTraceback:
#4 std::_Function_handler<std::shared_ptr<c10::LazyValue<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > > const> (), c10::SetStackTraceFetcher(std::function<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > ()>)::{lambda()#1}>::_M_invoke(std::_Any_data const&) from Logging.cpp:0
#5 c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) from ??:0
pytorch#6 c10::detail::torchCheckFail(char const*, char const*, unsigned int, char const*) from ??:0
pytorch#7 at::TensorBase::options() const from :0
pytorch#8 at::meta::resize_reduction(at::impl::MetaBase&, at::Tensor const&, c10::OptionalArrayRef<long>, bool, c10::ScalarType, bool) from :0
pytorch#9 at::meta::structured_sum_dim_IntList::meta(at::Tensor const&, c10::OptionalArrayRef<long>, bool, std::optional<c10::ScalarType>) from ??:0
pytorch#10 at::(anonymous namespace)::wrapper_CompositeExplicitAutogradNonFunctional_sum_dim_IntList(at::Tensor const&, c10::OptionalArrayRef<long>, bool, std::optional<c10::ScalarType>) from RegisterCompositeExplicitAutogradNonFunctional_0.cpp:0
pytorch#11 c10::impl::wrap_kernel_functor_unboxed_<c10::impl::detail::WrapFunctionIntoFunctor_<c10::CompileTimeFunctionPointer<at::Tensor (at::Tensor const&, c10::OptionalArrayRef<long>, bool, std::optional<c10::ScalarType>), &at::(anonymous namespace)::wrapper_CompositeExplicitAutogradNonFunctional_sum_dim_IntList>, at::Tensor, c10::guts::typelist::typelist<at::Tensor const&, c10::OptionalArrayRef<long>, bool, std::optional<c10::ScalarType> > >, at::Tensor (at::Tensor const&, c10::OptionalArrayRef<long>, bool, std::optional<c10::ScalarType>)>::call(c10::OperatorKernel*, c10::DispatchKeySet, at::Tensor const&, c10::OptionalArrayRef<long>, bool, std::optional<c10::ScalarType>) from RegisterCompositeExplicitAutogradNonFunctional_0.cpp:0
pytorch#12 at::_ops::sum_dim_IntList::call(at::Tensor const&, c10::OptionalArrayRef<long>, bool, std::optional<c10::ScalarType>) from ??:0
pytorch#13 void at::native::(anonymous namespace)::LaunchGammaBetaBackwardCUDAKernel<float, float>(float const*, float const*, float const*, float const*, long, long, at::Tensor*, at::Tensor*, CUstream_st*) from ??:0
pytorch#14 void at::native::(anonymous namespace)::LayerNormBackwardKernelImplInternal<float>(at::Tensor const&, at::Tensor const&, at::Tensor const&, at::Tensor const&, at::Tensor const&, long, long, at::Tensor*, at::Tensor*, at::Tensor*) from ??:0
pytorch#15 at::native::(anonymous namespace)::LayerNormBackwardKernelImpl(at::Tensor const&, at::Tensor const&, at::Tensor const&, at::Tensor const&, at::Tensor const&, long, long, at::Tensor*, at::Tensor*, at::Tensor*) from ??:0
pytorch#16 at::native::layer_norm_backward_cuda(at::Tensor const&, at::Tensor const&, c10::ArrayRef<long>, at::Tensor const&, at::Tensor const&, std::optional<at::Tensor> const&, std::optional<at::Tensor> const&, std::array<bool, 3ul>) from ??:0
pytorch#17 at::(anonymous namespace)::(anonymous namespace)::wrapper_CUDA__native_layer_norm_backward(at::Tensor const&, at::Tensor const&, c10::ArrayRef<c10::SymInt>, at::Tensor const&, at::Tensor const&, std::optional<at::Tensor> const&, std::optional<at::Tensor> const&, std::array<bool, 3ul>) from RegisterCUDA_0.cpp:0

```

Now we only call `sum(0)` on tensors that are defined and properly guard the `sum(0)` and assignment.
Pull Request resolved: pytorch#156600
Approved by: https://github.com/eqy, https://github.com/ngimel

Author: ahmadsharif1

aten/src/ATen/native/cuda/layer_norm_kernel.cu

@@ -884,8 +884,12 @@ void LaunchGammaBetaBackwardCUDAKernel(
     LaunchAndCheckGammaBetaBackwardKernel<T, T_ACC, block_dim_x, block_dim_y, rows_per_block_y, true>(
       aligned_grid, blocks, threads, 0, cuda_stream, dY_data, X_data, mean_data, rstd_data, M, N, dgamma_blocks_ptr, dbeta_blocks_ptr);
 
-    *dgamma = dgamma_blocks.sum(0);
-    *dbeta = dbeta_blocks.sum(0);
+    if (dgamma_blocks.defined()) {
+      *dgamma = dgamma_blocks.sum(0);
+    }
+    if (dbeta_blocks.defined()) {
+      *dbeta = dbeta_blocks.sum(0);
+    }
   } else {
     // We are in the normal case where M is not that large.
     // We can change the tile shape (which is the last template parameter) in accordance with M.

test/test_nn.py

@@ -7212,25 +7212,32 @@ def test_layer_norm_eps(self):
         ln = torch.nn.LayerNorm(2, eps=1e-6, elementwise_affine=False)
         self.assertEqual(ln.forward(x), torch.zeros_like(x))
 
+
     @unittest.skipIf(not TEST_CUDA, "CUDA not available")
     def test_layer_norm_backwards_eps(self):
         dtype = torch.float
         m_x_n_list = [(3, 3), (5, 5), (11, 11), (55, 55),
                       (32, 32), (1024, 32), (1024, 1024),
-                      (33, 33), (1025, 33), (1025, 1025)]
-        for m, n in m_x_n_list:
-            x = torch.randn((m, n), dtype=dtype, requires_grad=True)
-            grad_output = torch.rand_like(x)
-            x_cuda = x.clone().detach().to("cuda").requires_grad_()
-            grad_output_cuda = grad_output.clone().detach().to("cuda")
-            ln = nn.LayerNorm(n, dtype=dtype)
-            ln_cuda = nn.LayerNorm(n, device="cuda", dtype=dtype)
-            ln_out = ln(x)
-            ln_out_cuda = ln_cuda(x_cuda)
-            ln_out.backward(grad_output)
-            ln_out_cuda.backward(grad_output_cuda)
-            self.assertEqual(ln.weight.grad, ln_cuda.weight.grad, f"weight grad failed: {m=} {n=}", rtol=1e-5, atol=1e-4)
-            self.assertEqual(ln.bias.grad, ln_cuda.bias.grad, f"bias grad failed: {m=} {n=}", rtol=1e-5, atol=1e-4)
+                      (33, 33), (1025, 33), (1025, 1025),
+                      (128 * 1024, 32), (32, 128 * 1024)]
+        boolean = [True, False]
+        combinations = itertools.product(boolean, repeat=2)
+        for elementwise_affine, bias in combinations:
+            for m, n in m_x_n_list:
+                x = torch.randn((m, n), dtype=dtype, requires_grad=True)
+                grad_output = torch.rand_like(x)
+                x_cuda = x.clone().detach().to("cuda").requires_grad_()
+                grad_output_cuda = grad_output.clone().detach().to("cuda")
+                ln = nn.LayerNorm(n, dtype=dtype, elementwise_affine=elementwise_affine, bias=bias)
+                ln_cuda = nn.LayerNorm(n, device="cuda", dtype=dtype, elementwise_affine=elementwise_affine, bias=bias)
+                ln_out = ln(x)
+                ln_out_cuda = ln_cuda(x_cuda)
+                ln_out.backward(grad_output)
+                ln_out_cuda.backward(grad_output_cuda)
+                if elementwise_affine:
+                    self.assertEqual(ln.weight.grad, ln_cuda.weight.grad, f"weight grad failed: {m=} {n=}", rtol=1e-4, atol=1e-4)
+                if bias and elementwise_affine:
+                    self.assertEqual(ln.bias.grad, ln_cuda.bias.grad, f"bias grad failed: {m=} {n=}", rtol=1e-5, atol=1e-4)
 
     @largeTensorTest("40GB", device="cuda")
     def test_layer_norm_large_tensor(self):