Merge branch 'develop' into mem_color_ordering_fix

src/targets/gpu/include/migraphx/gpu/device/asin.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_ASIN_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_ASIN_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void asin(hipStream_t stream, const argument& result, const argument& arg);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/atan.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_ATAN_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_ATAN_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void atan(hipStream_t stream, const argument& result, const argument& arg);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/concat.hpp

-#ifndef MIGRAPH_GUARD_RTGLIB_DEVICE_CONCAT_HPP
-#define MIGRAPH_GUARD_RTGLIB_DEVICE_CONCAT_HPP
+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_CONCAT_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_CONCAT_HPP
 
 #include <migraphx/argument.hpp>
 #include <migraphx/config.hpp>
 #include <hip/hip_runtime_api.h>
 
 namespace migraphx {
-inline namespace MIGRAPH_INLINE_NS {
+inline namespace MIGRAPHX_INLINE_NS {
 namespace gpu {
 namespace device {
 
@@ -17,7 +17,7 @@ argument concat(hipStream_t stream,
 
 } // namespace device
 } // namespace gpu
-} // namespace MIGRAPH_INLINE_NS
+} // namespace MIGRAPHX_INLINE_NS
 } // namespace migraphx
 
 #endif

src/targets/gpu/include/migraphx/gpu/device/contiguous.hpp

-#ifndef MIGRAPH_GUARD_MIGRAPHLIB_KERNELS_HPP
-#define MIGRAPH_GUARD_MIGRAPHLIB_KERNELS_HPP
+#ifndef MIGRAPHX_GUARD_MIGRAPHLIB_KERNELS_HPP
+#define MIGRAPHX_GUARD_MIGRAPHLIB_KERNELS_HPP
 
 #include <migraphx/argument.hpp>
 #include <migraphx/config.hpp>
 #include <hip/hip_runtime_api.h>
 
 namespace migraphx {
-inline namespace MIGRAPH_INLINE_NS {
+inline namespace MIGRAPHX_INLINE_NS {
 namespace gpu {
 namespace device {
 
@@ -14,7 +14,7 @@ void contiguous(hipStream_t stream, argument result, argument arg);
 
 } // namespace device
 } // namespace gpu
-} // namespace MIGRAPH_INLINE_NS
+} // namespace MIGRAPHX_INLINE_NS
 } // namespace migraphx
 
 #endif

src/targets/gpu/include/migraphx/gpu/device/cos.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_COS_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_COS_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void cos(hipStream_t stream, const argument& result, const argument& arg);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/cosh.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_COSH_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_COSH_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void cosh(hipStream_t stream, const argument& result, const argument& arg);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/exp.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_EXP_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_EXP_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void exp(hipStream_t stream, const argument& result, const argument& arg);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/gather.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_GATHER_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_GATHER_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+argument gather(hipStream_t stream,
+                const migraphx::shape& output_shape,
+                std::vector<migraphx::argument> args,
+                int axis);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/log.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_LOG_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_LOG_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void log(hipStream_t stream, const argument& result, const argument& arg);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/max.hpp

+
+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_MAX_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_MAX_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void max(hipStream_t stream, const argument& result, const argument& arg1, const argument& arg2);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/min.hpp

+
+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_MIN_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_MIN_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void min(hipStream_t stream, const argument& result, const argument& arg1, const argument& arg2);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/mul.hpp

 
-#ifndef MIGRAPH_GUARD_RTGLIB_DEVICE_MUL_HPP
-#define MIGRAPH_GUARD_RTGLIB_DEVICE_MUL_HPP
+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_MUL_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_MUL_HPP
 
 #include <migraphx/argument.hpp>
 #include <migraphx/config.hpp>
 #include <hip/hip_runtime_api.h>
 
 namespace migraphx {
-inline namespace MIGRAPH_INLINE_NS {
+inline namespace MIGRAPHX_INLINE_NS {
 namespace gpu {
 namespace device {
 
@@ -21,7 +21,7 @@ void mul(hipStream_t stream,
 
 } // namespace device
 } // namespace gpu
-} // namespace MIGRAPH_INLINE_NS
+} // namespace MIGRAPHX_INLINE_NS
 } // namespace migraphx
 
 #endif

src/targets/gpu/include/migraphx/gpu/device/pad.hpp

+
+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_PAD_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_PAD_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+argument pad(hipStream_t stream,
+             argument result,
+             argument arg1,
+             float value,
+             std::vector<std::int64_t> pads);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/sin.hpp

-
-#ifndef MIGRAPH_GUARD_RTGLIB_DEVICE_ADD_HPP
-#define MIGRAPH_GUARD_RTGLIB_DEVICE_ADD_HPP
+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_SIN_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_SIN_HPP
 
 #include <migraphx/argument.hpp>
 #include <migraphx/config.hpp>
 #include <hip/hip_runtime_api.h>
 
 namespace migraphx {
-inline namespace MIGRAPH_INLINE_NS {
+inline namespace MIGRAPHX_INLINE_NS {
 namespace gpu {
 namespace device {
 
@@ -15,7 +14,7 @@ void sin(hipStream_t stream, const argument& result, const argument& arg);
 
 } // namespace device
 } // namespace gpu
-} // namespace MIGRAPH_INLINE_NS
+} // namespace MIGRAPHX_INLINE_NS
 } // namespace migraphx
 
 #endif

src/targets/gpu/include/migraphx/gpu/device/sinh.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_SINH_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_SINH_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void sinh(hipStream_t stream, const argument& result, const argument& arg);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/sub.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_SUB_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_SUB_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void sub(hipStream_t stream, const argument& result, const argument& arg1, const argument& arg2);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/device/tan.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_DEVICE_TAN_HPP
+#define MIGRAPHX_GUARD_RTGLIB_DEVICE_TAN_HPP
+
+#include <migraphx/argument.hpp>
+#include <migraphx/config.hpp>
+#include <hip/hip_runtime_api.h>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+namespace device {
+
+void tan(hipStream_t stream, const argument& result, const argument& arg);
+
+} // namespace device
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/eliminate_workspace.hpp

-#ifndef MIGRAPH_GUARD_RTGLIB_ELIMINATE_WORKSPACE_HPP
-#define MIGRAPH_GUARD_RTGLIB_ELIMINATE_WORKSPACE_HPP
+#ifndef MIGRAPHX_GUARD_RTGLIB_ELIMINATE_WORKSPACE_HPP
+#define MIGRAPHX_GUARD_RTGLIB_ELIMINATE_WORKSPACE_HPP
 
 #include <string>
 #include <migraphx/instruction_ref.hpp>
 #include <migraphx/config.hpp>
 
 namespace migraphx {
-inline namespace MIGRAPH_INLINE_NS {
+inline namespace MIGRAPHX_INLINE_NS {
 struct program;
 
 namespace gpu {
@@ -17,7 +17,7 @@ struct eliminate_workspace
     void apply(program& p) const;
 };
 } // namespace gpu
-} // namespace MIGRAPH_INLINE_NS
+} // namespace MIGRAPHX_INLINE_NS
 } // namespace migraphx
 
 #endif

src/targets/gpu/include/migraphx/gpu/elu.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_ELU_HPP
+#define MIGRAPHX_GUARD_RTGLIB_ELU_HPP
+
+#include <migraphx/gpu/lowering.hpp>
+#include <migraphx/manage_ptr.hpp>
+#include <migraphx/instruction.hpp>
+#include <migraphx/operators.hpp>
+#include <migraphx/generate.hpp>
+#include <migraphx/shape_for_each.hpp>
+#include <migraphx/config.hpp>
+#include <migraphx/gpu/miopen.hpp>
+#include <migraphx/gpu/hip.hpp>
+#include <migraphx/dfor.hpp>
+#include <migraphx/gpu/device/contiguous.hpp>
+#include <migraphx/gpu/device/add.hpp>
+#include <migraphx/iterator_for.hpp>
+#include <migraphx/gpu/rocblas.hpp>
+#include <migraphx/gpu/context.hpp>
+#include <utility>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+
+struct miopen_elu
+{
+    shared<activation_descriptor> ad;
+    std::string name() const { return "gpu::elu"; }
+    shape compute_shape(const std::vector<shape>& inputs) const;
+    argument
+    compute(context& ctx, const shape& output_shape, const std::vector<argument>& args) const;
+    int output_alias(const std::vector<shape>& shapes) const { return shapes.size() - 1; }
+};
+
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif

src/targets/gpu/include/migraphx/gpu/exp.hpp

+#ifndef MIGRAPHX_GUARD_RTGLIB_EXP_HPP
+#define MIGRAPHX_GUARD_RTGLIB_EXP_HPP
+
+#include <migraphx/gpu/lowering.hpp>
+#include <migraphx/gpu/oper.hpp>
+#include <migraphx/manage_ptr.hpp>
+#include <migraphx/instruction.hpp>
+#include <migraphx/operators.hpp>
+#include <migraphx/generate.hpp>
+#include <migraphx/shape_for_each.hpp>
+#include <migraphx/gpu/miopen.hpp>
+#include <migraphx/gpu/hip.hpp>
+#include <migraphx/dfor.hpp>
+#include <migraphx/gpu/device/contiguous.hpp>
+#include <migraphx/gpu/device/exp.hpp>
+#include <migraphx/iterator_for.hpp>
+#include <migraphx/gpu/rocblas.hpp>
+#include <migraphx/gpu/context.hpp>
+#include <migraphx/config.hpp>
+#include <utility>
+
+namespace migraphx {
+inline namespace MIGRAPHX_INLINE_NS {
+namespace gpu {
+
+struct hip_exp : unary_device<hip_exp, device::exp>
+{
+};
+
+} // namespace gpu
+} // namespace MIGRAPHX_INLINE_NS
+} // namespace migraphx
+
+#endif