Migrate backends/vulkan to the new namespace (#5876)

Summary:
Pull Request resolved: #5876

Move the vulkan backend out of the `torch::` namespace, and update to avoid using the `torch::` or `exec_aten::` namespaces.

Reviewed By: jorgep31415

Differential Revision: D63903939

fbshipit-source-id: 42ff0ba2b560eb3e2bee02c014e39f44d7b0ec94
(cherry picked from commit b9e9479)

Author: dbort

backends/vulkan/runtime/VulkanBackend.cpp

@@ -30,11 +30,23 @@
 #include <type_traits>
 #include <vector>
 
-namespace torch {
-namespace executor {
+namespace executorch {
+namespace backends {
 namespace vulkan {
 namespace {
 
+using executorch::runtime::ArrayRef;
+using executorch::runtime::Backend;
+using executorch::runtime::BackendExecutionContext;
+using executorch::runtime::BackendInitContext;
+using executorch::runtime::CompileSpec;
+using executorch::runtime::DelegateHandle;
+using executorch::runtime::Error;
+using executorch::runtime::EValue;
+using executorch::runtime::FreeableBuffer;
+using executorch::runtime::kTensorDimensionLimit;
+using executorch::runtime::Result;
+
 using namespace vkcompute;
 
 // Flatbuffer types
@@ -357,7 +369,7 @@ class GraphBuilder {
 bool maybe_resize_input(
     ComputeGraph* graph,
     const size_t input_i,
-    exec_aten::Tensor& et_tensor) {
+    executorch::aten::Tensor& et_tensor) {
   ValueRef in_tensor_ref = graph->inputs()[input_i].value;
   vTensorPtr in_tensor = graph->get_tensor(in_tensor_ref);
 
@@ -392,17 +404,18 @@ bool maybe_resize_input(
 void maybe_resize_output(
     ComputeGraph* graph,
     const size_t output_i,
-    exec_aten::Tensor& et_tensor) {
+    executorch::aten::Tensor& et_tensor) {
   ValueRef out_tensor_ref = graph->outputs()[output_i].value;
   vTensorPtr out_tensor = graph->get_tensor(out_tensor_ref);
 
-  exec_aten::SizesType new_output_size[kTensorDimensionLimit];
+  executorch::aten::SizesType new_output_size[kTensorDimensionLimit];
   size_t ndim = out_tensor->sizes().size();
   for (int i = 0; i < ndim; ++i) {
     new_output_size[i] = out_tensor->sizes()[i];
   }
 
-  exec_aten::ArrayRef<exec_aten::SizesType> output_size{new_output_size, ndim};
+  executorch::aten::ArrayRef<executorch::aten::SizesType> output_size{
+      new_output_size, ndim};
   Error err = resize_tensor(et_tensor, output_size);
 
   ET_CHECK_MSG(err == Error::Ok, "Failed to resize output tensor.");
@@ -555,5 +568,5 @@ static auto success_with_compiler = register_backend(backend);
 
 } // namespace
 } // namespace vulkan
-} // namespace executor
-} // namespace torch
+} // namespace backends
+} // namespace executorch

backends/vulkan/runtime/VulkanDelegateHeader.cpp

@@ -15,10 +15,13 @@
 
 #pragma clang diagnostic ignored "-Wdeprecated"
 
-namespace torch {
-namespace executor {
+namespace executorch {
+namespace backends {
 namespace vulkan {
 
+using executorch::runtime::Error;
+using executorch::runtime::Result;
+
 namespace {
 
 struct ByteSlice {
@@ -101,5 +104,5 @@ Result<VulkanDelegateHeader> VulkanDelegateHeader::parse(const void* data) {
 }
 
 } // namespace vulkan
-} // namespace executor
-} // namespace torch
+} // namespace backends
+} // namespace executorch

backends/vulkan/runtime/VulkanDelegateHeader.h

@@ -10,8 +10,8 @@
 
 #include <executorch/runtime/core/result.h>
 
-namespace torch {
-namespace executor {
+namespace executorch {
+namespace backends {
 namespace vulkan {
 
 // Byte decoding utilities
@@ -22,7 +22,8 @@ uint32_t getUInt16LE(const uint8_t* data);
 struct VulkanDelegateHeader {
   bool is_valid() const;
 
-  static Result<VulkanDelegateHeader> parse(const void* data);
+  static executorch::runtime::Result<VulkanDelegateHeader> parse(
+      const void* data);
 
   uint32_t header_size;
   uint32_t flatbuffer_offset;
@@ -32,5 +33,5 @@ struct VulkanDelegateHeader {
 };
 
 } // namespace vulkan
-} // namespace executor
-} // namespace torch
+} // namespace backends
+} // namespace executorch

backends/vulkan/test/utils/test_utils.cpp

@@ -322,7 +322,7 @@ void record_reference_matmul(
   _(uint8_t, Byte)                \
   _(int8_t, Char)                 \
   _(int32_t, Int)                 \
-  _(exec_aten::Half, Half)        \
+  _(executorch::aten::Half, Half) \
   _(float, Float)                 \
   _(int8_t, QInt8)
 

backends/vulkan/test/vulkan_compute_api_test.cpp

@@ -485,7 +485,7 @@ TEST_F(VulkanComputeAPITest, test_buffer_float16) {
   if (!context()->adapter_ptr()->has_full_float16_buffers_support()) {
     GTEST_SKIP();
   }
-  test_storage_buffer_type<exec_aten::Half, vkapi::kHalf>(16);
+  test_storage_buffer_type<executorch::aten::Half, vkapi::kHalf>(16);
 }
 
 TEST_F(VulkanComputeAPITest, test_buffer_int8) {
@@ -567,7 +567,7 @@ TEST_F(VulkanComputeAPITest, buffer_tensor_sanity_check) {
             run_buffer_tensor_sanity_check<float>(a);
             break;
           case vkapi::kHalf:
-            run_buffer_tensor_sanity_check<exec_aten::Half>(a);
+            run_buffer_tensor_sanity_check<executorch::aten::Half>(a);
             break;
           case vkapi::kChar:
             run_buffer_tensor_sanity_check<int8_t>(a);
@@ -2399,7 +2399,7 @@ TEST(VulkanToFromGPUShaderTest, round_trip_tests) {
 
   for (auto& sizes : to_test) {
     RUN_TESTS(float, vkapi::kFloat)
-    RUN_TESTS(exec_aten::Half, vkapi::kHalf)
+    RUN_TESTS(executorch::aten::Half, vkapi::kHalf)
   }
 
   for (auto& sizes : to_test_int8) {