Index: flang/lib/Optimizer/CodeGen/Target.cpp
===================================================================
--- flang/lib/Optimizer/CodeGen/Target.cpp
+++ flang/lib/Optimizer/CodeGen/Target.cpp
@@ -465,6 +465,46 @@
 };
 } // namespace
 
+//===----------------------------------------------------------------------===//
+// AMDGPU linux target specifics.
+//===----------------------------------------------------------------------===//
+
+namespace {
+struct TargetAMDGPU : public GenericTarget<TargetAMDGPU> {
+  using GenericTarget::GenericTarget;
+
+  static constexpr int defaultWidth = 64;
+
+  CodeGenSpecifics::Marshalling
+  complexArgumentType(mlir::Location loc, mlir::Type eleTy) const override {
+    CodeGenSpecifics::Marshalling marshal;
+    const auto *sem = &floatToSemantics(kindMap, eleTy);
+    if (sem == &llvm::APFloat::IEEEsingle() ||
+        sem == &llvm::APFloat::IEEEdouble()) {
+      // <2 x eleTy> vector of 2 eleTy
+      marshal.emplace_back(fir::VectorType::get(2, eleTy), AT{});
+    } else {
+      TODO(loc, "complex for this precision");
+    }
+    return marshal;
+  }
+
+  CodeGenSpecifics::Marshalling
+  complexReturnType(mlir::Location loc, mlir::Type eleTy) const override {
+    CodeGenSpecifics::Marshalling marshal;
+    const auto *sem = &floatToSemantics(kindMap, eleTy);
+    if (sem == &llvm::APFloat::IEEEsingle() ||
+        sem == &llvm::APFloat::IEEEdouble()) {
+      // <2 x eleTy>   vector of 2 eleTy
+      marshal.emplace_back(fir::VectorType::get(2, eleTy), AT{});
+    } else {
+      TODO(loc, "complex for this precision");
+    }
+    return marshal;
+  }
+};
+} // namespace
+
 // Instantiate the overloaded target instance based on the triple value.
 // TODO: Add other targets to this file as needed.
 std::unique_ptr<fir::CodeGenSpecifics>
@@ -497,6 +537,9 @@
   case llvm::Triple::ArchType::riscv64:
     return std::make_unique<TargetRISCV64>(ctx, std::move(trp),
                                            std::move(kindMap));
+  case llvm::Triple::ArchType::amdgcn:
+    return std::make_unique<TargetAMDGPU>(ctx, std::move(trp),
+                                          std::move(kindMap));
   }
   TODO(mlir::UnknownLoc::get(ctx), "target not implemented");
 }
Index: flang/test/Driver/target-gpu-features.f90
===================================================================
--- /dev/null
+++ flang/test/Driver/target-gpu-features.f90
@@ -0,0 +1,10 @@
+! REQUIRES: amdgpu-registered-target
+
+! Test that -mcpu are used and that the -target-cpu and -target-features
+! are also added to the fc1 command.
+
+! RUN: %flang --target=amdgcn-amd-amdhsa -mcpu=gfx902 -c %s -### 2>&1 \
+! RUN: | FileCheck %s -check-prefix=CHECK-AMDGCN
+
+! CHECK-AMDGCN: "-fc1" "-triple" "amdgcn-amd-amdhsa"
+! CHECK-AMDGCN-SAME: "-target-cpu" "gfx902"