Build the CUDA backend outside fbcode, with the LLM cache optional

Summary:
The CUDA backend's C++ runtime builds only in fbcode, and only with the LLM extension's KV cache linked in. So a C++ application that runs an exported vision or audio model on an NVIDIA GPU can't use it from any other Buck cell, cross-compiled aarch64 targets such as Jetson included. To run the same AOTInductor-compiled kernels it has to link LibTorch instead. XNNPACK, Vulkan and QNN already define their C++ targets in both cells; this does the same for CUDA.

- **Targets in both cells.** The `aoti/slim`, `extension/cuda` and `cuda/runtime` C++ targets are now defined outside fbcode too. There they carry `compatible_with = ["ovr_config//cuda:has_cuda"]`, so CPU-only builds skip them instead of failing to compile, and SlimTensor `storage` picks up its CUDA deps through a `select`. The Python AOTI targets stay fbcode-only.
- **Public headers.** Headers that dependents include (`slim_tensor.h`, `storage.h`, `runtime_api.h`, the shim and factory headers) move from `headers` to `exported_headers`, and `runtime_api`'s CUDA dependency becomes an exported external dep. Outside fbcode `headers` are private to the target, so nothing could compile against them. fbcode already exports both, so nothing changes there.
- **LLM cache optional, as in CMake.** CMake already builds `cuda_kv_cache.cpp` and `cuda_kv_pool.cpp` only with `EXECUTORCH_BUILD_EXTENSION_LLM`. Buck now does the same outside fbcode: the cache code in `cuda_backend.cpp` and `cuda_kv_cache.h` is behind that define, and the define is exported so that dependents see the same declarations as the backend. CMake now exports it too (`PUBLIC` instead of `PRIVATE` on `aoti_cuda_backend`), since the header's declarations depend on it; `test_cuda_kv_cache` and `cuda_batching` include that header. fbcode keeps the cache. A backend built without it still refuses a program whose KV cache lives off-graph, because `requires_offgraph_kv_storage` reads that from the program's constants and sits outside the gate.
- **Build flags.** The `nvcc` architecture flags stay fbcode-only. Outside fbcode the shims undefine `CPU_CAPABILITY_AVX2` and `CPU_CAPABILITY_AVX512` for CUDA sources: c10 exports them on x86, and `Half.h` then pulls host AVX2 `vec_half` code into nvcc compiles. `cuda-lazy` maps to `libcudart`, and the shims link `curand` and `thrust` for `rand.cu` and `sort.cu`.

fbcode builds keep the same sources and flags, and the CMake build is untouched.

Differential Revision: D129654471

Author: mattiasbrand

backends/aoti/slim/c10/core/targets.bzl

@@ -1,16 +1,12 @@
 load("@fbsource//xplat/executorch/build:runtime_wrapper.bzl", "runtime")
-load("@fbsource//tools/build_defs:fbsource_utils.bzl", "is_fbcode")
 
 def define_common_targets():
     """Define targets for SlimTensor c10 core module."""
 
-    if not is_fbcode():
-        return
-
     # Header-only library for DeviceType
     runtime.cxx_library(
         name = "device_type",
-        headers = [
+        exported_headers = [
             "DeviceType.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],
@@ -22,7 +18,7 @@ def define_common_targets():
     # Header-only library for Device
     runtime.cxx_library(
         name = "device",
-        headers = [
+        exported_headers = [
             "Device.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],
@@ -35,7 +31,7 @@ def define_common_targets():
     # Header-only library for ScalarType
     runtime.cxx_library(
         name = "scalar_type",
-        headers = [
+        exported_headers = [
             "ScalarType.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],
@@ -48,7 +44,7 @@ def define_common_targets():
     # Header-only library for SizesAndStrides
     runtime.cxx_library(
         name = "sizes_and_strides",
-        headers = [
+        exported_headers = [
             "SizesAndStrides.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],
@@ -62,7 +58,7 @@ def define_common_targets():
     # Header-only library for Contiguity
     runtime.cxx_library(
         name = "contiguity",
-        headers = [
+        exported_headers = [
             "Contiguity.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],
@@ -74,7 +70,7 @@ def define_common_targets():
     # Header-only library for WrapDimMinimal
     runtime.cxx_library(
         name = "wrap_dim_minimal",
-        headers = [
+        exported_headers = [
             "WrapDimMinimal.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],

backends/aoti/slim/c10/cuda/BUCK

@@ -1,7 +1,5 @@
-load("@fbcode_macros//build_defs:build_file_migration.bzl", "fbcode_target", "non_fbcode_target")
-load("@fbsource//xplat/executorch/build:runtime_wrapper.bzl", "runtime")
 load(":targets.bzl", "define_common_targets")
 
 oncall("executorch")
 
-fbcode_target(_kind = define_common_targets,)
+define_common_targets()

backends/aoti/slim/c10/macros/BUCK

@@ -1,6 +1,5 @@
-load("@fbcode_macros//build_defs:build_file_migration.bzl", "fbcode_target", "non_fbcode_target")
 load("targets.bzl", "define_common_targets")
 
 oncall("executorch")
 
-fbcode_target(_kind = define_common_targets,)
+define_common_targets()

backends/aoti/slim/c10/macros/targets.bzl

@@ -6,7 +6,7 @@ def define_common_targets():
     # Header-only library for Macros
     runtime.cxx_library(
         name = "macros",
-        headers = [
+        exported_headers = [
             "Macros.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],

backends/aoti/slim/core/BUCK

@@ -1,6 +1,5 @@
-load("@fbcode_macros//build_defs:build_file_migration.bzl", "fbcode_target", "non_fbcode_target")
 load("targets.bzl", "define_common_targets")
 
 oncall("executorch")
 
-fbcode_target(_kind = define_common_targets,)
+define_common_targets()

backends/aoti/slim/core/targets.bzl

@@ -1,3 +1,4 @@
+load("@fbsource//tools/build_defs:fbsource_utils.bzl", "is_fbcode")
 load("@fbsource//xplat/executorch/build:runtime_wrapper.bzl", "runtime")
 
 def define_common_targets():
@@ -6,7 +7,7 @@ def define_common_targets():
     # Header-only library for Storage
     runtime.cxx_library(
         name = "storage",
-        headers = [
+        exported_headers = [
             "storage.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],
@@ -17,16 +18,27 @@ def define_common_targets():
             "//executorch/backends/aoti/slim/util:shared_ptr",
             "//executorch/backends/aoti/slim/util:size_util",
             "//executorch/runtime/platform:platform",
-            "//executorch/backends/aoti/slim/c10/cuda:exception",
-            "//executorch/backends/aoti/slim/cuda:guard",
-            "//executorch/extension/cuda:caller_stream",
-            "//executorch/extension/cuda:cuda_allocator",
-        ],
+        ] + (
+            [
+                "//executorch/backends/aoti/slim/c10/cuda:exception",
+                "//executorch/backends/aoti/slim/cuda:guard",
+                "//executorch/extension/cuda:caller_stream",
+                "//executorch/extension/cuda:cuda_allocator",
+            ] if is_fbcode() or runtime.is_oss else select({
+                "DEFAULT": [],
+                "ovr_config//cuda:has_cuda": [
+                    "//executorch/backends/aoti/slim/c10/cuda:exception",
+                    "//executorch/backends/aoti/slim/cuda:guard",
+                    "//executorch/extension/cuda:caller_stream",
+                    "//executorch/extension/cuda:cuda_allocator",
+                ],
+            })
+        ),
     )
 
     runtime.cxx_library(
         name = "slimtensor",
-        headers = [
+        exported_headers = [
             "slim_tensor.h",
             "slim_tensor_view_incl.h",
         ],
@@ -41,6 +53,5 @@ def define_common_targets():
             "//executorch/backends/aoti/slim/util:array_ref_util",
             "//executorch/backends/aoti/slim/util:size_util",
             "//executorch/runtime/platform:platform",
-            "//executorch/backends/aoti/slim/c10/cuda:exception",
         ],
     )

backends/aoti/slim/cuda/BUCK

@@ -1,7 +1,5 @@
-load("@fbcode_macros//build_defs:build_file_migration.bzl", "fbcode_target", "non_fbcode_target")
-load("@fbsource//xplat/executorch/build:runtime_wrapper.bzl", "runtime")
 load(":targets.bzl", "define_common_targets")
 
 oncall("executorch")
 
-fbcode_target(_kind = define_common_targets,)
+define_common_targets()

backends/aoti/slim/cuda/targets.bzl

@@ -8,7 +8,7 @@ def define_common_targets():
         srcs = [
             "guard.cpp",
         ],
-        headers = [
+        exported_headers = [
             "guard.h",
         ],
         visibility = ["PUBLIC"],

backends/aoti/slim/factory/BUCK

@@ -1,6 +1,5 @@
-load("@fbcode_macros//build_defs:build_file_migration.bzl", "fbcode_target", "non_fbcode_target")
 load("targets.bzl", "define_common_targets")
 
 oncall("executorch")
 
-fbcode_target(_kind = define_common_targets,)
+define_common_targets()

backends/aoti/slim/factory/targets.bzl

@@ -6,7 +6,7 @@ def define_common_targets():
     # Header-only library for empty tensor factory functions
     runtime.cxx_library(
         name = "empty",
-        headers = [
+        exported_headers = [
             "empty.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],
@@ -19,7 +19,7 @@ def define_common_targets():
 
     runtime.cxx_library(
         name = "from_blob",
-        headers = [
+        exported_headers = [
             "from_blob.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],
@@ -32,7 +32,7 @@ def define_common_targets():
 
     runtime.cxx_library(
         name = "from_etensor",
-        headers = [
+        exported_headers = [
             "from_etensor.h",
         ],
         visibility = ["@EXECUTORCH_CLIENTS"],