pytorch
diff --git a/‎aten/src/ATen/native/transformers/cuda/attention.cu
Lines changed: 2 additions & 14 deletions b/‎aten/src/ATen/native/transformers/cuda/attention.cu
Lines changed: 2 additions & 14 deletions
diff --git a/‎aten/src/ATen/native/transformers/cuda/attention_backward.cu
Lines changed: 0 additions & 12 deletions b/‎aten/src/ATen/native/transformers/cuda/attention_backward.cu
Lines changed: 0 additions & 12 deletions
@@ -1113,10 +1113,8 @@ _flash_attention_forward(
   std::optional<Tensor> alibi_slopes = _alibi_slopes;
   const float softcap = 0.0;
 
-#ifndef USE_ROCM  // ROCM backend accepts std::optional for window_size_left/right directly.
-  const int non_null_window_left = window_size_left.value_or(-1);
-  const int non_null_window_right = window_size_right.value_or(-1);
-#endif
+  const int non_null_window_left = window_size_left.has_value() ? window_size_left.value() : -1;
+  const int non_null_window_right = window_size_right.has_value() ? window_size_right.value() : -1;
 
   // We are going to have two paths:
   // 1. The standard MHA path for dense tensors
@@ -1153,13 +1151,8 @@ _flash_attention_forward(
             softmax_scale,
             false /*zero_tensors*/,
             is_causal,
-#ifdef USE_ROCM
-            window_size_left,
-            window_size_right,
-#else
             non_null_window_left,
             non_null_window_right,
-#endif
             softcap,
             return_debug_mask,
             std::nullopt /*gen_*/);
@@ -1182,13 +1175,8 @@ _flash_attention_forward(
             dropout_p,
             softmax_scale,
             is_causal,
-#ifdef USE_ROCM
-            window_size_left,
-            window_size_right,
-#else
             non_null_window_left,
             non_null_window_right,
-#endif
             softcap,
             return_debug_mask, /*return_softmax (this is used for testing)*/
             std::nullopt);
 
@@ -87,10 +87,8 @@ std::tuple<Tensor, Tensor, Tensor> _flash_attention_backward(
   auto contiguous_grad_out = grad_out.contiguous();
   auto contiguous_out = out.contiguous();
 
-#ifndef USE_ROCM  // ROCM backend accepts std::optional for window_size_left/right directly.
   const int non_null_window_left = window_size_left.has_value() ? window_size_left.value() : -1;
   const int non_null_window_right = window_size_right.has_value() ? window_size_right.value() : -1;
-#endif
 
   std::optional<at::Tensor> dq{std::nullopt};
   std::optional<at::Tensor> dk{std::nullopt};
@@ -138,13 +136,8 @@ std::tuple<Tensor, Tensor, Tensor> _flash_attention_backward(
         softmax_scale,
         false /*zero_tensors*/,
         is_causal,
-#ifdef USE_ROCM
-        window_size_left,
-        window_size_right,
-#else
         non_null_window_left,
         non_null_window_right,
-#endif
         softcap,
         determinisitic,
         philox_seed,
@@ -166,13 +159,8 @@ std::tuple<Tensor, Tensor, Tensor> _flash_attention_backward(
         dropout_p,
         softmax_scale,
         is_causal,
-#ifdef USE_ROCM
-        window_size_left,
-        window_size_right,
-#else
         non_null_window_left,
         non_null_window_right,
-#endif
         softcap,
         determinisitic,
         philox_seed,