Add conversion code for Float16 that was missed in #8174 (#8178)

* Add conversion code for Float16 that was missed in #8174

* Don't sniff for _Float16 when building ASAN

* Update HalideRuntime.h

src/runtime/HalideRuntime.h

@@ -86,6 +86,20 @@ extern "C" {
 
 #ifndef COMPILING_HALIDE_RUNTIME
 
+// ASAN builds can cause linker errors for Float16, so sniff for that and
+// don't enable it by default.
+#if defined(__has_feature)
+#if __has_feature(address_sanitizer)
+#define HALIDE_RUNTIME_ASAN_DETECTED
+#endif
+#endif
+
+#if defined(__SANITIZE_ADDRESS__) && !defined(HALIDE_RUNTIME_ASAN_DETECTED)
+#define HALIDE_RUNTIME_ASAN_DETECTED
+#endif
+
+#if !defined(HALIDE_RUNTIME_ASAN_DETECTED)
+
 // clang had _Float16 added as a reserved name in clang 8, but
 // doesn't actually support it on most platforms until clang 15.
 // Ideally there would be a better way to detect if the type
@@ -108,6 +122,8 @@ extern "C" {
 #endif
 #endif
 
+#endif  // !HALIDE_RUNTIME_ASAN_DETECTED
+
 #endif  // !COMPILING_HALIDE_RUNTIME
 
 /** \file

tools/halide_image_io.h

@@ -2227,6 +2227,10 @@ struct ImageTypeConversion {
 
         const halide_type_t src_type = src.type();
         switch (src_type.element_of().as_u32()) {
+#ifdef HALIDE_CPP_COMPILER_HAS_FLOAT16
+        case halide_type_t(halide_type_float, 16).as_u32():
+            return convert_image<DstElemType>(src.template as<_Float16, AnyDims>());
+#endif
         case halide_type_t(halide_type_float, 32).as_u32():
             return convert_image<DstElemType>(src.template as<float, AnyDims>());
         case halide_type_t(halide_type_float, 64).as_u32():
@@ -2272,6 +2276,10 @@ struct ImageTypeConversion {
         // Call the appropriate static-to-static conversion routine
         // based on the desired dst type.
         switch (dst_type.element_of().as_u32()) {
+#ifdef HALIDE_CPP_COMPILER_HAS_FLOAT16
+        case halide_type_t(halide_type_float, 16).as_u32():
+            return convert_image<_Float16>(src);
+#endif
         case halide_type_t(halide_type_float, 32).as_u32():
             return convert_image<float>(src);
         case halide_type_t(halide_type_float, 64).as_u32():