Merge branch 'develop' into lwpck-976

include/ck/utility/inner_product.hpp

@@ -192,6 +192,8 @@ inner_product<int8x4_t, int8x4_t, int32_t>(const int8x4_t& a, const int8x4_t& b,
 #else
     c = __builtin_amdgcn_sdot4(bit_cast<int32_t>(a), bit_cast<int32_t>(b), c, false);
 #endif
+#elif defined(CK_USE_AMD_V_DOT4_I32_I8_GFX11)
+    c = __builtin_amdgcn_sudot4(true, bit_cast<int32_t>(a), true, bit_cast<int32_t>(b), c, false);
 #else
     const vector_type<int8_t, 4> a_vector{a};
     const vector_type<int8_t, 4> b_vector{b};

include/ck/utility/statically_indexed_array_multi_index.hpp

@@ -5,6 +5,7 @@
 #define CK_STATICALLY_INDEXED_ARRAY_MULTI_INDEX_HPP
 
 #include "common_header.hpp"
+#include "ck/utility/math_v2.hpp"
 
 namespace ck {
 

library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_dl_instance.hpp

@@ -6,8 +6,6 @@
 #include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_dl.hpp"
 #include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
 
-#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
-
 namespace ck {
 namespace tensor_operation {
 namespace device {