diff --git a/doc/make_constants.py b/doc/make_constants.py
index cef8c8c99c6981d0c02a34d7ffd9cb02161d9422..6416ca0f73f9041514cd3fe554b0a8a483d6b29d 100644
--- a/doc/make_constants.py
+++ b/doc/make_constants.py
@@ -27,6 +27,7 @@ nv_devattr = ("cl_nv_device_attribute_query", "0.92")
 gl_sharing = ("cl_khr_gl_sharing", "0.92")
 cl_11 = ("CL_1.1", "0.92")
 cl_12 = ("CL_1.2", "2011.2")
+amd_devattr = ("cl_amd_device_attribute_query", "2013.2")
 
 
 def get_extra_lines(tup):
@@ -95,8 +96,22 @@ const_ext_lookup = {
             "HALF_FP_CONFIG":
             ("cl_khr_fp16", "2011.1"),
 
-            "PROFILING_TIMER_OFFSET_AMD":
-            ("cl_amd_device_attribute_query", "2011.1"),
+            "PROFILING_TIMER_OFFSET_AMD": amd_devattr,
+            "TOPOLOGY_AMD": amd_devattr,
+            "BOARD_NAME_AMD": amd_devattr,
+            "GLOBAL_FREE_MEMORY_AMD": amd_devattr,
+            "SIMD_PER_COMPUTE_UNIT_AMD": amd_devattr,
+            "SIMD_WIDTH_AMD": amd_devattr,
+            "SIMD_INSTRUCTION_WIDTH_AMD": amd_devattr,
+            "WAVEFRONT_WIDTH_AMD": amd_devattr,
+            "GLOBAL_MEM_CHANNELS_AMD": amd_devattr,
+            "GLOBAL_MEM_CHANNEL_BANKS_AMD": amd_devattr,
+            "GLOBAL_MEM_CHANNEL_BANK_WIDTH_AMD": amd_devattr,
+            "LOCAL_MEM_SIZE_PER_COMPUTE_UNIT_AMD": amd_devattr,
+            "LOCAL_MEM_BANKS_AMD": amd_devattr,
+
+            "MAX_ATOMIC_COUNTERS_EXT":
+            ("cl_ext_atomic_counters_64", "2013.2"),
 
             "PARENT_DEVICE_EXT":
             fission,
diff --git a/pyopencl/array.py b/pyopencl/array.py
index 68f28e5d160e6ff3ca25aec4a2d50213aff2979a..ea99b1115049079357ea212665f6337c9cd8a858 100644
--- a/pyopencl/array.py
+++ b/pyopencl/array.py
@@ -232,7 +232,7 @@ class Array(object):
     as in :mod:`numpy`.  Arithmetic methods in :class:`Array` support the
     broadcasting of scalars. (e.g. `array+5`)
 
-    *cqa* must be a :class:`pyopencl.CommandQueue` or a:class:`Context`.
+    *cqa* must be a :class:`pyopencl.CommandQueue` or a :class:`pyopencl.Context`.
 
     If it is a queue, *cqa* specifies the queue in which the array carries out
     its computations by default. If a default queue (and thereby overloaded
@@ -240,7 +240,8 @@ class Array(object):
     :class:`Context`.
 
     *cqa* will at some point be renamed *cq*, so it should be considered
-    'positional-only'.
+    'positional-only'. Arguments starting from 'order' should be considered
+    keyword-only.
 
     *allocator* may be `None` or a callable that, upon being called with an
     argument of the number of bytes to be allocated, returns an
diff --git a/src/wrapper/wrap_cl.hpp b/src/wrapper/wrap_cl.hpp
index d246f15062baf3e87f9d35967374562231ddf9af..b85cb6f0fd04786384183a05e9663eacd8aa5ce9 100644
--- a/src/wrapper/wrap_cl.hpp
+++ b/src/wrapper/wrap_cl.hpp
@@ -720,6 +720,62 @@ namespace pyopencl
           case CL_DEVICE_PREFERRED_INTEROP_USER_SYNC: DEV_GET_INT_INF(cl_bool);
           case CL_DEVICE_PRINTF_BUFFER_SIZE: DEV_GET_INT_INF(cl_bool);
 #endif
+// {{{ AMD dev attrs
+//
+// types of AMD dev attrs divined from
+// https://www.khronos.org/registry/cl/api/1.2/cl.hpp
+#ifdef CL_DEVICE_PROFILING_TIMER_OFFSET_AMD
+          case CL_DEVICE_PROFILING_TIMER_OFFSET_AMD: DEV_GET_INT_INF(cl_ulong);
+#endif
+/* FIXME
+#ifdef CL_DEVICE_TOPOLOGY_AMD
+          case CL_DEVICE_TOPOLOGY_AMD:
+#endif
+*/
+#ifdef CL_DEVICE_BOARD_NAME_AMD
+          case CL_DEVICE_BOARD_NAME_AMD: ;
+            PYOPENCL_GET_STR_INFO(Device, m_device, param_name);
+#endif
+#ifdef CL_DEVICE_GLOBAL_FREE_MEMORY_AMD
+          case CL_DEVICE_GLOBAL_FREE_MEMORY_AMD:
+            {
+              std::vector<size_t> result;
+              PYOPENCL_GET_VEC_INFO(Device, m_device, param_name, result);
+              PYOPENCL_RETURN_VECTOR(size_t, result);
+            }
+#endif
+#ifdef CL_DEVICE_SIMD_PER_COMPUTE_UNIT_AMD
+          case CL_DEVICE_SIMD_PER_COMPUTE_UNIT_AMD: DEV_GET_INT_INF(cl_uint);
+#endif
+#ifdef CL_DEVICE_SIMD_WIDTH_AMD
+          case CL_DEVICE_SIMD_WIDTH_AMD: DEV_GET_INT_INF(cl_uint);
+#endif
+#ifdef CL_DEVICE_SIMD_INSTRUCTION_WIDTH_AMD
+          case CL_DEVICE_SIMD_INSTRUCTION_WIDTH_AMD: DEV_GET_INT_INF(cl_uint);
+#endif
+#ifdef CL_DEVICE_WAVEFRONT_WIDTH_AMD
+          case CL_DEVICE_WAVEFRONT_WIDTH_AMD: DEV_GET_INT_INF(cl_uint);
+#endif
+#ifdef CL_DEVICE_GLOBAL_MEM_CHANNELS_AMD
+          case CL_DEVICE_GLOBAL_MEM_CHANNELS_AMD: DEV_GET_INT_INF(cl_uint);
+#endif
+#ifdef CL_DEVICE_GLOBAL_MEM_CHANNEL_BANKS_AMD
+          case CL_DEVICE_GLOBAL_MEM_CHANNEL_BANKS_AMD: DEV_GET_INT_INF(cl_uint);
+#endif
+#ifdef CL_DEVICE_GLOBAL_MEM_CHANNEL_BANK_WIDTH_AMD
+          case CL_DEVICE_GLOBAL_MEM_CHANNEL_BANK_WIDTH_AMD: DEV_GET_INT_INF(cl_uint);
+#endif
+#ifdef CL_DEVICE_LOCAL_MEM_SIZE_PER_COMPUTE_UNIT_AMD
+          case CL_DEVICE_LOCAL_MEM_SIZE_PER_COMPUTE_UNIT_AMD: DEV_GET_INT_INF(cl_uint);
+#endif
+#ifdef CL_DEVICE_LOCAL_MEM_BANKS_AMD
+          case CL_DEVICE_LOCAL_MEM_BANKS_AMD: DEV_GET_INT_INF(cl_uint);
+#endif
+// }}}
+
+#ifdef CL_DEVICE_MAX_ATOMIC_COUNTERS_EXT
+          case CL_DEVICE_MAX_ATOMIC_COUNTERS_EXT: DEV_GET_INT_INF(cl_uint);
+#endif
 
           default:
             throw error("Device.get_info", CL_INVALID_VALUE);
diff --git a/src/wrapper/wrap_constants.cpp b/src/wrapper/wrap_constants.cpp
index d6e2246cf19073f59d055cf97f154b2a682ee912..64511d01656c5a5c0f629242884fdab4944f56b3 100644
--- a/src/wrapper/wrap_constants.cpp
+++ b/src/wrapper/wrap_constants.cpp
@@ -312,10 +312,50 @@ void pyopencl_expose_constants()
     ADD_ATTR(DEVICE_, KERNEL_EXEC_TIMEOUT_NV);
     ADD_ATTR(DEVICE_, INTEGRATED_MEMORY_NV);
 #endif
-// cl_amd_device_attribute_query
+// {{{ cl_amd_device_attribute_query
 #ifdef CL_DEVICE_PROFILING_TIMER_OFFSET_AMD
     ADD_ATTR(DEVICE_, PROFILING_TIMER_OFFSET_AMD);
 #endif
+#ifdef CL_DEVICE_TOPOLOGY_AMD
+    ADD_ATTR(DEVICE_, TOPOLOGY_AMD);
+#endif
+#ifdef CL_DEVICE_BOARD_NAME_AMD
+    ADD_ATTR(DEVICE_, BOARD_NAME_AMD);
+#endif
+#ifdef CL_DEVICE_GLOBAL_FREE_MEMORY_AMD
+    ADD_ATTR(DEVICE_, GLOBAL_FREE_MEMORY_AMD);
+#endif
+#ifdef CL_DEVICE_SIMD_PER_COMPUTE_UNIT_AMD
+    ADD_ATTR(DEVICE_, SIMD_PER_COMPUTE_UNIT_AMD);
+#endif
+#ifdef CL_DEVICE_SIMD_WIDTH_AMD
+    ADD_ATTR(DEVICE_, SIMD_WIDTH_AMD);
+#endif
+#ifdef CL_DEVICE_SIMD_INSTRUCTION_WIDTH_AMD
+    ADD_ATTR(DEVICE_, SIMD_INSTRUCTION_WIDTH_AMD);
+#endif
+#ifdef CL_DEVICE_WAVEFRONT_WIDTH_AMD
+    ADD_ATTR(DEVICE_, WAVEFRONT_WIDTH_AMD);
+#endif
+#ifdef CL_DEVICE_GLOBAL_MEM_CHANNELS_AMD
+    ADD_ATTR(DEVICE_, GLOBAL_MEM_CHANNELS_AMD);
+#endif
+#ifdef CL_DEVICE_GLOBAL_MEM_CHANNEL_BANKS_AMD
+    ADD_ATTR(DEVICE_, GLOBAL_MEM_CHANNEL_BANKS_AMD);
+#endif
+#ifdef CL_DEVICE_GLOBAL_MEM_CHANNEL_BANK_WIDTH_AMD
+    ADD_ATTR(DEVICE_, GLOBAL_MEM_CHANNEL_BANK_WIDTH_AMD);
+#endif
+#ifdef CL_DEVICE_LOCAL_MEM_SIZE_PER_COMPUTE_UNIT_AMD
+    ADD_ATTR(DEVICE_, LOCAL_MEM_SIZE_PER_COMPUTE_UNIT_AMD);
+#endif
+#ifdef CL_DEVICE_LOCAL_MEM_BANKS_AMD
+    ADD_ATTR(DEVICE_, LOCAL_MEM_BANKS_AMD);
+#endif
+// }}}
+#ifdef CL_DEVICE_MAX_ATOMIC_COUNTERS_EXT
+    ADD_ATTR(DEVICE_, MAX_ATOMIC_COUNTERS_EXT);
+#endif
 #if defined(cl_ext_device_fission) && defined(PYOPENCL_USE_DEVICE_FISSION)
     ADD_ATTR(DEVICE_, PARENT_DEVICE_EXT);
     ADD_ATTR(DEVICE_, PARTITION_TYPES_EXT);
@@ -336,6 +376,10 @@ void pyopencl_expose_constants()
     ADD_ATTR(DEVICE_, REFERENCE_COUNT);
     ADD_ATTR(DEVICE_, PREFERRED_INTEROP_USER_SYNC);
     ADD_ATTR(DEVICE_, PRINTF_BUFFER_SIZE);
+#endif
+#ifdef cl_khr_image2d_from_buffer
+    ADD_ATTR(DEVICE_, IMAGE_PITCH_ALIGNMENT);
+    ADD_ATTR(DEVICE_, IMAGE_BASE_ADDRESS_ALIGNMENT);
 #endif
   }