diff --git a/projects/miopen/cmake/EnableCompilerWarnings.cmake b/projects/miopen/cmake/EnableCompilerWarnings.cmake
index 10a610f1c4a7..c08a3611e329 100644
--- a/projects/miopen/cmake/EnableCompilerWarnings.cmake
+++ b/projects/miopen/cmake/EnableCompilerWarnings.cmake
@@ -24,64 +24,29 @@
 #
 ################################################################################
 
-set(__default_cxx_compile_options
+set(__cxx_compile_options
+    -Werror
     -Wall
     -Wextra
-    -Wcomment
-    -Wendif-labels
-    -Wformat
-    -Winit-self
-    -Wreturn-type
-    -Wsequence-point
-    -Wswitch
-    -Wtrigraphs
+    # Additional warnings not included in -Wall/-Wextra
     -Wundef
-    -Wuninitialized
     -Wunreachable-code
-    -Wno-ignored-qualifiers
-    -Wno-sign-compare
+    -Wmissing-noreturn
+    -Wshadow
+    -Wsuggest-override
+    -Wold-style-cast
+    # TODO: Working to enable these warnings. Each requires code cleanup first.
+    # -Wconversion            # ~1000+ implicit narrowing/sign conversions to fix
+    # -Wdouble-promotion      # implicit float-to-double promotions
+    # Suppress specific warnings -- working to remove these by fixing the code
+    -Wno-c++11-narrowing        # ~40 instances: narrowing in brace init (batchnorm, ck_impl, addkernels)
+    -Wno-sign-compare           # ~1000+ instances: signed/unsigned comparisons throughout codebase
+    -Wno-deprecated-declarations # 2 deprecated MIOpen APIs still have callers
 )
 
 set(__clang_cxx_compile_options
-    -Weverything
-    -Wno-c++98-compat
-    -Wno-c++98-compat-pedantic
-    -Wno-conversion
-    -Wno-double-promotion
-    -Wno-exit-time-destructors
-    -Wno-extra-semi
-    -Wno-extra-semi-stmt
-    -Wno-missing-prototypes
-    -Wno-padded
     -Wno-unused-command-line-argument
-    -Wno-weak-vtables
-    -Wno-covered-switch-default
-    -Wno-unsafe-buffer-usage
-    -Wno-deprecated-declarations
-    -Wno-global-constructors
-    -Wno-reserved-identifier
-    -Wno-deprecated
-    -Wno-old-style-cast
-    -Wno-c++11-narrowing
-    -Wno-switch-enum
-    -Wno-suggest-override
-    -Wno-nonportable-system-include-path
-    -Wno-documentation
-    -Wmissing-noreturn)
-
-if(CMAKE_CXX_COMPILER_ID STREQUAL "Clang" AND CMAKE_CXX_COMPILER_VERSION VERSION_GREATER_EQUAL "19")
-    list(APPEND __clang_cxx_compile_options
-        -Wno-unique-object-duplication
-        -Wno-switch-default)
-endif()
-
-if(CMAKE_CXX_COMPILER_ID STREQUAL "Clang" AND CMAKE_CXX_COMPILER_VERSION VERSION_GREATER_EQUAL "23")
-    list(APPEND __clang_cxx_compile_options
-        -Wno-lifetime-safety
-        -Wno-lifetime-safety-suggestions
-        -Wno-lifetime-safety-intra-tu-suggestions
-        -Wno-lifetime-safety-cross-tu-suggestions)
-endif()
+)
 
 if(WIN32)
     list(APPEND __clang_cxx_compile_options
@@ -90,8 +55,9 @@ if(WIN32)
 endif()
 
 add_compile_options(
-    "$<$<AND:$<COMPILE_LANGUAGE:CXX>,$<CXX_COMPILER_ID:Clang>>:${__default_cxx_compile_options};${__clang_cxx_compile_options}>"
+    "$<$<COMPILE_LANGUAGE:CXX>:${__cxx_compile_options}>"
+    "$<$<AND:$<COMPILE_LANGUAGE:CXX>,$<CXX_COMPILER_ID:Clang>>:${__clang_cxx_compile_options}>"
 )
 
-unset(__default_cxx_compile_options)
+unset(__cxx_compile_options)
 unset(__clang_cxx_compile_options)
diff --git a/projects/miopen/driver/kthvalue_driver.hpp b/projects/miopen/driver/kthvalue_driver.hpp
index 75f7e5b535b2..44010c173107 100644
--- a/projects/miopen/driver/kthvalue_driver.hpp
+++ b/projects/miopen/driver/kthvalue_driver.hpp
@@ -294,7 +294,7 @@ int KthvalueDriver<TIO>::RunForwardGPU()
                               outputDesc,
                               output_dev->GetMem(),
                               indicesDesc,
-                              (size_t*)indices_dev->GetMem(),
+                              static_cast<size_t*>(indices_dev->GetMem()),
                               k,
                               dim,
                               keepDim);
diff --git a/projects/miopen/driver/rocrand_wrapper.cpp b/projects/miopen/driver/rocrand_wrapper.cpp
index 2a7d910c4be7..fe8ba5fe9687 100644
--- a/projects/miopen/driver/rocrand_wrapper.cpp
+++ b/projects/miopen/driver/rocrand_wrapper.cpp
@@ -26,7 +26,14 @@
 
 #include "rocrand_wrapper.hpp"
 
+// Suppress warnings from rocrand's own headers: in TheRock builds rocrand/stage/include
+// is injected globally as -I (not -isystem), so rocrand headers are not treated as system
+// includes and their internal issues trigger our warnings.
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Wold-style-cast"
+#pragma clang diagnostic ignored "-Wundef"
 #include <rocrand/rocrand.hpp>
+#pragma clang diagnostic pop
 
 namespace gpumemrand {
 
diff --git a/projects/miopen/src/CMakeLists.txt b/projects/miopen/src/CMakeLists.txt
index 9e6f401b7506..cd4ee6ab6dcb 100644
--- a/projects/miopen/src/CMakeLists.txt
+++ b/projects/miopen/src/CMakeLists.txt
@@ -1010,10 +1010,10 @@ endif()
 
 if(hipblaslt_FOUND)
     target_link_libraries( MIOpen PRIVATE roc::hipblaslt )
-    target_include_directories( MIOpen INTERFACE $<BUILD_INTERFACE:${HIPBLAS-COMMON_INCLUDE_DIR}> )
+    target_include_directories( MIOpen SYSTEM INTERFACE $<BUILD_INTERFACE:${HIPBLAS-COMMON_INCLUDE_DIR}> )
 endif()
 
-target_include_directories( MIOpen PRIVATE ${ROCRAND_INCLUDE_DIR} )
+target_include_directories( MIOpen SYSTEM PRIVATE ${ROCRAND_INCLUDE_DIR} )
 
 # For backward compatibility with ROCm 5.3
 # Build with library libMLIRMIOpen
diff --git a/projects/miopen/src/ck_impl/implicitgemm_ck_util.hpp b/projects/miopen/src/ck_impl/implicitgemm_ck_util.hpp
index bc3a840e35e6..2c2da8e56644 100644
--- a/projects/miopen/src/ck_impl/implicitgemm_ck_util.hpp
+++ b/projects/miopen/src/ck_impl/implicitgemm_ck_util.hpp
@@ -433,7 +433,8 @@ template <typename DeviceOpType,
           typename CKArgsType,
           typename ProblemDescriptionType = miopen::conv::ProblemDescription,
           bool CheckSplitK                = false>
-bool IsCKArgsSupported(const ProblemDescriptionType& problem, const std::string& kernel_id)
+bool IsCKArgsSupported([[maybe_unused]] const ProblemDescriptionType& problem,
+                       [[maybe_unused]] const std::string& kernel_id)
 {
 #if MIOPEN_BACKEND_HIP && MIOPEN_USE_COMPOSABLEKERNEL
     if(!kernel_id.empty())
@@ -687,7 +688,7 @@ ConvSolution InitAnyInvokerFactory(const ProblemDescriptionType& problem,
 }
 
 template <typename DataType, typename OutElemOp>
-OutElemOp GetOutElementOp(const miopen::fusion::ActivationOpInvokeParam& activationOp)
+OutElemOp GetOutElementOp([[maybe_unused]] const miopen::fusion::ActivationOpInvokeParam& activationOp)
 {
 #if MIOPEN_BACKEND_HIP && MIOPEN_USE_COMPOSABLEKERNEL
     auto activationMode = activationOp.activMode;
@@ -893,12 +894,12 @@ template <bool ZeroOutputs,
           typename Input1TposeOp,
           typename Input2TposeOp,
           typename OutputTposeOp>
-ConvSolution InitInvokerFactoryNCHW(const ExecutionContext& ctx,
-                                    const miopen::conv::ProblemDescription& problem,
-                                    const std::string& kernel_id,
-                                    const Input1TposeOp& input1_op,
-                                    const Input2TposeOp& input2_op,
-                                    const OutputTposeOp& output_op)
+ConvSolution InitInvokerFactoryNCHW([[maybe_unused]] const ExecutionContext& ctx,
+                                    [[maybe_unused]] const miopen::conv::ProblemDescription& problem,
+                                    [[maybe_unused]] const std::string& kernel_id,
+                                    [[maybe_unused]] const Input1TposeOp& input1_op,
+                                    [[maybe_unused]] const Input2TposeOp& input2_op,
+                                    [[maybe_unused]] const OutputTposeOp& output_op)
 {
     assert(problem.IsLayoutDefault());
 
@@ -1063,7 +1064,7 @@ template <bool ZeroOutputs,
           typename CastType,
           typename ProblemDescriptionType = miopen::conv::ProblemDescription>
 ConvSolution InitInvokerFactoryNHWC(const ExecutionContext&,
-                                    const ProblemDescriptionType& problem,
+                                    [[maybe_unused]] const ProblemDescriptionType& problem,
                                     const std::string& kernel_id)
 {
     ConvSolution result;
@@ -1270,10 +1271,10 @@ ConvSolution InitInvokerFactoryWrwNCHW(const ExecutionContext& ctx,
 
 template <typename InvokerFactoryMakerNCHW, typename InvokerFactoryMakerNHWC>
 ConvSolution
-MakeSolutionGroupConvImplicitGemmXdlops(const miopen::conv::ProblemDescription& problem,
-                                        InvokerFactoryMakerNCHW&& invoker_factory_maker_ncdhw,
-                                        InvokerFactoryMakerNHWC&& invoker_factory_maker_ndhwc,
-                                        const bool use_tf32 = false)
+MakeSolutionGroupConvImplicitGemmXdlops([[maybe_unused]] const miopen::conv::ProblemDescription& problem,
+                                        [[maybe_unused]] InvokerFactoryMakerNCHW&& invoker_factory_maker_ncdhw,
+                                        [[maybe_unused]] InvokerFactoryMakerNHWC&& invoker_factory_maker_ndhwc,
+                                        [[maybe_unused]] const bool use_tf32 = false)
 {
 
 #if MIOPEN_BACKEND_HIP && MIOPEN_USE_COMPOSABLEKERNEL
diff --git a/projects/miopen/src/env.cpp b/projects/miopen/src/env.cpp
index 24892ce7b3dd..5ad2a88f1eb5 100644
--- a/projects/miopen/src/env.cpp
+++ b/projects/miopen/src/env.cpp
@@ -35,7 +35,9 @@
 #include <string_view>
 
 #ifdef _WIN32
+#ifndef WIN32_LEAN_AND_MEAN
 #define WIN32_LEAN_AND_MEAN
+#endif
 #include <Windows.h>
 #endif
 
diff --git a/projects/miopen/src/include/miopen/batchnorm/common_spatial.hpp b/projects/miopen/src/include/miopen/batchnorm/common_spatial.hpp
index af563db96a41..e7e59436a279 100644
--- a/projects/miopen/src/include/miopen/batchnorm/common_spatial.hpp
+++ b/projects/miopen/src/include/miopen/batchnorm/common_spatial.hpp
@@ -170,7 +170,7 @@ inline int GetStashMethod(bool IsLayoutNHWC,
         (in_cstride) % ylocalsize == 0 ? ylocalsize : (in_cstride) % ylocalsize;
     unsigned int last_zlocalsize =
         n % (zlocalsize * nelements) == 0 ? (zlocalsize * nelements) : n % (zlocalsize * nelements);
-    if(last_ylocalsize < stash_values && last_zlocalsize >= (size_t)stash_values)
+    if(last_ylocalsize < stash_values && last_zlocalsize >= static_cast<size_t>(stash_values))
     {
         stash_method = 1;
     }
@@ -400,8 +400,8 @@ inline bool IsSpatialMultipleApplicable(const miopen::batchnorm::ProblemDescript
         //    be large enough
         //  - if C is not multiple of 2, intermediate results are stored in N dimension splitting
         //    float values in group of 2 bytes. N must be large enough
-        if((!bfp32parm && (c % 2 != 0 && last_zlocalsize < (size_t)stash_values)) ||
-           ((last_ylocalsize < stash_values) && (last_zlocalsize < (size_t)stash_values)))
+        if((!bfp32parm && (c % 2 != 0 && last_zlocalsize < static_cast<size_t>(stash_values))) ||
+           ((last_ylocalsize < stash_values) && (last_zlocalsize < static_cast<size_t>(stash_values))))
         {
             return false;
         }
@@ -425,7 +425,7 @@ inline bool IsSpatialMultipleApplicable(const miopen::batchnorm::ProblemDescript
         //  - if last block doesn't fit, intermediate results are stored in N dimension which must
         //    be large enough
         stash_values *= (problem.GetXDesc().GetType() == miopenFloat ? 1 : 2);
-        if(last_ylocalsize < stash_values && last_zlocalsize < (size_t)stash_values)
+        if(last_ylocalsize < stash_values && last_zlocalsize < static_cast<size_t>(stash_values))
         {
             return false;
         }
diff --git a/projects/miopen/src/include/miopen/conv/solvers.hpp b/projects/miopen/src/include/miopen/conv/solvers.hpp
index e27ea412219f..bc7079e17e51 100644
--- a/projects/miopen/src/include/miopen/conv/solvers.hpp
+++ b/projects/miopen/src/include/miopen/conv/solvers.hpp
@@ -4829,18 +4829,18 @@ struct ConvWinogradNHWCTransposingTunableSolver
 
 struct TransposedConvBinWinograd3x3U final : ConvWinogradNHWCTransposingSolver<ConvBinWinograd3x3U>
 {
-    const std::string& SolverDbId() const { return GetSolverDbId<TransposedConvBinWinograd3x3U>(); }
+    const std::string& SolverDbId() const override { return GetSolverDbId<TransposedConvBinWinograd3x3U>(); }
 };
 
 struct TransposedConvBinWinogradRxS final : ConvWinogradNHWCTransposingSolver<ConvBinWinogradRxS>
 {
-    const std::string& SolverDbId() const { return GetSolverDbId<TransposedConvBinWinogradRxS>(); }
+    const std::string& SolverDbId() const override { return GetSolverDbId<TransposedConvBinWinogradRxS>(); }
 };
 
 struct TransposedConvBinWinogradRxSf2x3g1 final
     : ConvWinogradNHWCTransposingSolver<ConvBinWinogradRxSf2x3g1>
 {
-    const std::string& SolverDbId() const
+    const std::string& SolverDbId() const override
     {
         return GetSolverDbId<TransposedConvBinWinogradRxSf2x3g1>();
     }
@@ -4850,7 +4850,7 @@ struct TransposedConvMPBidirectWinograd final
     : ConvWinogradNHWCTransposingSolver<
           ConvMPBidirectWinograd<WinoDataH, WinoFilterH, WinoDataW, WinoFilterW>>
 {
-    const std::string& SolverDbId() const
+    const std::string& SolverDbId() const override
     {
         return this->template GetSolverDbId<
             TransposedConvMPBidirectWinograd<WinoDataH, WinoFilterH, WinoDataW, WinoFilterW>>();
@@ -4862,7 +4862,7 @@ struct TransposedConvWinograd3x3MultipassWrW final
     : ConvWinogradNHWCTransposingSolver<
           ConvWinograd3x3MultipassWrW<WinoDataH, WinoFilterH, WinoDataW, WinoFilterW>>
 {
-    const std::string& SolverDbId() const
+    const std::string& SolverDbId() const override
     {
         return this->template GetSolverDbId<TransposedConvWinograd3x3MultipassWrW<WinoDataH,
                                                                                   WinoFilterH,
@@ -4875,7 +4875,7 @@ template <uint32_t Winodata, uint32_t Winofilter>
 struct TransposedConvWinoFuryRxS final
     : ConvWinogradNHWCTransposingSolver<ConvWinoFuryRxS<Winodata, Winofilter>>
 {
-    const std::string& SolverDbId() const
+    const std::string& SolverDbId() const override
     {
         return this->template GetSolverDbId<TransposedConvWinoFuryRxS<Winodata, Winofilter>>();
     }
@@ -4885,7 +4885,7 @@ template <uint32_t Winodata, uint32_t Winofilter>
 struct TransposedConvWinoRageRxS final
     : ConvWinogradNHWCTransposingSolver<ConvWinoRageRxS<Winodata, Winofilter>>
 {
-    const std::string& SolverDbId() const
+    const std::string& SolverDbId() const override
     {
         return this->template GetSolverDbId<TransposedConvWinoRageRxS<Winodata, Winofilter>>();
     }
@@ -4907,7 +4907,7 @@ struct TransposedConvMPBidirectWinograd_xdlops final
     : ConvWinogradNHWCTransposingTunableSolver<
           ConvMPBidirectWinograd_xdlops<WinoDataH, WinoFilterH, WinoDataW, WinoFilterW>>
 {
-    const std::string& SolverDbId() const
+    const std::string& SolverDbId() const override
     {
         return this->template GetSolverDbId<TransposedConvMPBidirectWinograd_xdlops<WinoDataH,
                                                                                     WinoFilterH,
diff --git a/projects/miopen/src/include/miopen/file_lock.hpp b/projects/miopen/src/include/miopen/file_lock.hpp
index ba27a179c5c2..6d83477ea230 100644
--- a/projects/miopen/src/include/miopen/file_lock.hpp
+++ b/projects/miopen/src/include/miopen/file_lock.hpp
@@ -13,7 +13,9 @@
 #ifndef NOMINMAX
 #define NOMINMAX
 #endif
+#ifndef WIN32_LEAN_AND_MEAN
 #define WIN32_LEAN_AND_MEAN
+#endif
 #include <windows.h>
 #include <io.h>
 #else
diff --git a/projects/miopen/src/include/miopen/pooling/solvers.hpp b/projects/miopen/src/include/miopen/pooling/solvers.hpp
index 38f9019eaa8b..2dd7c084d3c5 100644
--- a/projects/miopen/src/include/miopen/pooling/solvers.hpp
+++ b/projects/miopen/src/include/miopen/pooling/solvers.hpp
@@ -330,23 +330,23 @@ struct PoolingFwdNCHWTransposingSolver
 
 struct TransposedPoolingFwd2d final : PoolingFwdNCHWTransposingSolver<PoolingForward2d>
 {
-    const std::string& SolverDbId() const { return GetSolverDbId<TransposedPoolingFwd2d>(); }
+    const std::string& SolverDbId() const override { return GetSolverDbId<TransposedPoolingFwd2d>(); }
     PoolingForward2d::PerformanceConfigType
     GetDefaultPerformanceConfig(const ExecutionContext& ctx,
-                                const miopen::pooling::ProblemDescription& problem) const
+                                const miopen::pooling::ProblemDescription& problem) const override
     {
         return PoolingForward2d{}.GetDefaultPerformanceConfig(ctx, problem);
     }
     bool IsValidPerformanceConfig(const ExecutionContext& ctx,
                                   const miopen::pooling::ProblemDescription& problem,
-                                  const PoolingForward2d::PerformanceConfigType& config) const
+                                  const PoolingForward2d::PerformanceConfigType& config) const override
     {
         return PoolingForward2d{}.IsValidPerformanceConfig(ctx, problem, config);
     }
     PoolingForward2d::PerformanceConfigType
     Search(const ExecutionContext& context,
            const miopen::pooling::ProblemDescription& problem,
-           const AnyInvokeParams& invoke_context) const
+           const AnyInvokeParams& invoke_context) const override
     {
         return PoolingForward2d{}.Search(context, problem, invoke_context);
     }
@@ -354,23 +354,23 @@ struct TransposedPoolingFwd2d final : PoolingFwdNCHWTransposingSolver<PoolingFor
 
 struct TransposedPoolingFwdNd final : PoolingFwdNCHWTransposingSolver<PoolingForwardNd>
 {
-    const std::string& SolverDbId() const { return GetSolverDbId<TransposedPoolingFwdNd>(); }
+    const std::string& SolverDbId() const override { return GetSolverDbId<TransposedPoolingFwdNd>(); }
     PoolingForwardNd::PerformanceConfigType
     GetDefaultPerformanceConfig(const ExecutionContext& ctx,
-                                const miopen::pooling::ProblemDescription& problem) const
+                                const miopen::pooling::ProblemDescription& problem) const override
     {
         return PoolingForwardNd{}.GetDefaultPerformanceConfig(ctx, problem);
     }
     bool IsValidPerformanceConfig(const ExecutionContext& ctx,
                                   const miopen::pooling::ProblemDescription& problem,
-                                  const PoolingForwardNd::PerformanceConfigType& config) const
+                                  const PoolingForwardNd::PerformanceConfigType& config) const override
     {
         return PoolingForwardNd{}.IsValidPerformanceConfig(ctx, problem, config);
     }
     PoolingForwardNd::PerformanceConfigType
     Search(const ExecutionContext& context,
            const miopen::pooling::ProblemDescription& problem,
-           const AnyInvokeParams& invoke_context) const
+           const AnyInvokeParams& invoke_context) const override
     {
         return PoolingForwardNd{}.Search(context, problem, invoke_context);
     }
@@ -474,23 +474,23 @@ struct PoolingBwdNCHWTransposingSolver
 
 struct TransposedPoolingBwd2d final : PoolingBwdNCHWTransposingSolver<PoolingBackward2d>
 {
-    const std::string& SolverDbId() const { return GetSolverDbId<TransposedPoolingBwd2d>(); }
+    const std::string& SolverDbId() const override { return GetSolverDbId<TransposedPoolingBwd2d>(); }
     PoolingBackward2d::PerformanceConfigType
     GetDefaultPerformanceConfig(const ExecutionContext& ctx,
-                                const miopen::pooling::ProblemDescription& problem) const
+                                const miopen::pooling::ProblemDescription& problem) const override
     {
         return PoolingBackward2d{}.GetDefaultPerformanceConfig(ctx, problem);
     }
     bool IsValidPerformanceConfig(const ExecutionContext& ctx,
                                   const miopen::pooling::ProblemDescription& problem,
-                                  const PoolingBackward2d::PerformanceConfigType& config) const
+                                  const PoolingBackward2d::PerformanceConfigType& config) const override
     {
         return PoolingBackward2d{}.IsValidPerformanceConfig(ctx, problem, config);
     }
     PoolingBackward2d::PerformanceConfigType
     Search(const ExecutionContext& context,
            const miopen::pooling::ProblemDescription& problem,
-           const AnyInvokeParams& invoke_context) const
+           const AnyInvokeParams& invoke_context) const override
     {
         return PoolingBackward2d{}.Search(context, problem, invoke_context);
     }
@@ -498,23 +498,23 @@ struct TransposedPoolingBwd2d final : PoolingBwdNCHWTransposingSolver<PoolingBac
 
 struct TransposedPoolingBwdNd final : PoolingBwdNCHWTransposingSolver<PoolingBackwardNd>
 {
-    const std::string& SolverDbId() const { return GetSolverDbId<TransposedPoolingBwdNd>(); }
+    const std::string& SolverDbId() const override { return GetSolverDbId<TransposedPoolingBwdNd>(); }
     PoolingBackwardNd::PerformanceConfigType
     GetDefaultPerformanceConfig(const ExecutionContext& ctx,
-                                const miopen::pooling::ProblemDescription& problem) const
+                                const miopen::pooling::ProblemDescription& problem) const override
     {
         return PoolingBackwardNd{}.GetDefaultPerformanceConfig(ctx, problem);
     }
     bool IsValidPerformanceConfig(const ExecutionContext& ctx,
                                   const miopen::pooling::ProblemDescription& problem,
-                                  const PoolingBackwardNd::PerformanceConfigType& config) const
+                                  const PoolingBackwardNd::PerformanceConfigType& config) const override
     {
         return PoolingBackwardNd{}.IsValidPerformanceConfig(ctx, problem, config);
     }
     PoolingBackwardNd::PerformanceConfigType
     Search(const ExecutionContext& context,
            const miopen::pooling::ProblemDescription& problem,
-           const AnyInvokeParams& invoke_context) const
+           const AnyInvokeParams& invoke_context) const override
     {
         return PoolingBackwardNd{}.Search(context, problem, invoke_context);
     }
diff --git a/projects/miopen/src/include/miopen/solver/implicitgemm_util.hpp b/projects/miopen/src/include/miopen/solver/implicitgemm_util.hpp
index 3f4e3ecde5d7..e6d5cb6cb5e8 100644
--- a/projects/miopen/src/include/miopen/solver/implicitgemm_util.hpp
+++ b/projects/miopen/src/include/miopen/solver/implicitgemm_util.hpp
@@ -177,7 +177,7 @@ inline T igemm_get_max_gks(T gemm_k, T gemm_k_per_block, T max_log2_splits)
         return 0;
     T rem      = gemm_k / gemm_k_per_block;
     T rem_pow2 = rem & (~(rem - 1));
-    T gks      = (T)log2(rem_pow2);
+    T gks      = static_cast<T>(log2(rem_pow2));
 
     if(gks > max_log2_splits)
         gks = max_log2_splits;
diff --git a/projects/miopen/src/kernels/hip_float8.hpp b/projects/miopen/src/kernels/hip_float8.hpp
index 352cd97bb9b0..7d837be2a219 100644
--- a/projects/miopen/src/kernels/hip_float8.hpp
+++ b/projects/miopen/src/kernels/hip_float8.hpp
@@ -46,6 +46,13 @@
 
 #define USE_SIMPLER_HIP_F8x8 0
 
+// Pull FP8 config from config.h when available (host compilation).
+// During kernel compilation config.h is not in the include path;
+// the macros are passed via -D flags from the solver instead.
+#if __has_include(<miopen/config.h>)
+#include <miopen/config.h>
+#endif
+
 #ifndef MIOPEN_FP8_CLIPPING
 #define MIOPEN_FP8_CLIPPING 1
 #endif
diff --git a/projects/miopen/src/kernels/miopen_rocrand.hpp b/projects/miopen/src/kernels/miopen_rocrand.hpp
index 3effb18802e2..019e14193be7 100644
--- a/projects/miopen/src/kernels/miopen_rocrand.hpp
+++ b/projects/miopen/src/kernels/miopen_rocrand.hpp
@@ -44,12 +44,19 @@
 #define ROCRAND_DETAIL_XORWOW_BM_NOT_IN_STATE
 
 // Use inlined rocrand header for runtime compilation to avoid external dependency
-// For host code compilation, use the regular rocrand header
+// For host code compilation, use the regular rocrand header.
+// Suppress warnings from rocrand's own headers: in TheRock builds rocrand/stage/include
+// is injected globally as -I (not -isystem), so rocrand headers are not treated as system
+// includes and their internal issues trigger our warnings.
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Wold-style-cast"
+#pragma clang diagnostic ignored "-Wundef"
 #ifdef MIOPEN_HIP_RUNTIME_COMPILE
 #include "rocrand_xorwow_inlined.h"
 #else
 #include <rocrand/rocrand_xorwow.h>
 #endif
+#pragma clang diagnostic pop
 
 namespace prng {
 
diff --git a/projects/miopen/src/solver/conv/conv_asm_implicit_gemm_gtc_bwd_nhwc.cpp b/projects/miopen/src/solver/conv/conv_asm_implicit_gemm_gtc_bwd_nhwc.cpp
index d44ab5ba59fa..4a30062d10c2 100644
--- a/projects/miopen/src/solver/conv/conv_asm_implicit_gemm_gtc_bwd_nhwc.cpp
+++ b/projects/miopen/src/solver/conv/conv_asm_implicit_gemm_gtc_bwd_nhwc.cpp
@@ -873,7 +873,7 @@ bool PerformanceConfigAsmImplicitGemmGTCBwdXdlopsNHWC::IsValid(
                     n = n | (n >> 16);
                     return n - (n >> 1);
                 }(padded_k_num);
-                int k_pow2 = (int)log2(prev_pow2);
+                int k_pow2 = static_cast<int>(log2(prev_pow2));
 
                 return std::min(k_pow2, BWD_MAX_GEMM_K_SPLITS);
             }
diff --git a/projects/miopen/test/conv_common.hpp b/projects/miopen/test/conv_common.hpp
index 5e141908bab7..87f116690296 100644
--- a/projects/miopen/test/conv_common.hpp
+++ b/projects/miopen/test/conv_common.hpp
@@ -2272,14 +2272,14 @@ struct conv_driver : test_driver
             {
                 auto output = get_output_tensor<T, Tout>(filter, input, weights, out_layout);
 
-                auto gen_positive_value = [=](auto...) {
+                auto gen_positive_value = [=, this](auto...) {
                     auto data_type = input.desc.GetType();
                     int v_max      = is_int8 ? 16 : (data_type == miopenHalf) ? 4 : 17;
                     return gen_float ? prng::gen_canonical<double>()
                                      : static_cast<double>(prng::gen_A_to_B(1, v_max));
                 };
 
-                auto gen_sign_value = [=](auto... is) {
+                auto gen_sign_value = [=, this](auto... is) {
                     auto data_type = input.desc.GetType();
                     int v_max      = is_int8 ? 16 : (data_type == miopenHalf) ? 4 : 17;
                     return gen_float ? prng::gen_A_to_B(-1.0, 1.0)
diff --git a/projects/miopen/test/driver.hpp b/projects/miopen/test/driver.hpp
index 13097aaca388..53b08f24492e 100644
--- a/projects/miopen/test/driver.hpp
+++ b/projects/miopen/test/driver.hpp
@@ -112,7 +112,7 @@ struct test_driver
         template <class Source, class T>
         void add_source(Source src, T& x)
         {
-            data_sources.push_back([=, &x](std::function<void()> callback) {
+            data_sources.push_back([=, this, &x](std::function<void()> callback) {
                 for(auto y : src()) // NOLINT
                 {
                     x = T(y);
@@ -332,7 +332,7 @@ struct test_driver
     template <class X, class G>
     generate_tensor_t<X, G> generate_tensor(std::set<X> dims, X single, G g)
     {
-        return {[=]() -> std::set<X> {
+        return {[=, this]() -> std::set<X> {
                     if(full_set)
                         return dims;
                     else
@@ -351,7 +351,7 @@ struct test_driver
     template <class F, class G>
     auto lazy_generate_tensor(F f, G g) -> generate_tensor_t<miopen::range_value<decltype(f())>, G>
     {
-        return {[=]() -> decltype(f()) {
+        return {[=, this]() -> decltype(f()) {
                     if(full_set)
                         return f();
                     else
@@ -363,7 +363,7 @@ struct test_driver
     template <class F, class X, class G>
     generate_tensor_t<X, G> lazy_generate_tensor(F f, X single, G g)
     {
-        return {[=]() -> std::set<X> {
+        return {[=, this]() -> std::set<X> {
                     if(full_set)
                         return f();
                     else
@@ -382,7 +382,7 @@ struct test_driver
     template <class F, class G>
     generate_tensor_t<std::vector<int>, G> get_tensor(F gen_shapes, G gen_value)
     {
-        return lazy_generate_tensor([=] { return gen_shapes(batch_factor); }, gen_value);
+        return lazy_generate_tensor([=, this] { return gen_shapes(batch_factor); }, gen_value);
     }
 
     template <class G = tensor_elem_gen_integer>
@@ -390,7 +390,7 @@ struct test_driver
     get_bn_spatial_input_tensor(G tensor_elem_gen = tensor_elem_gen_integer{})
     {
         return lazy_generate_tensor(
-            [=] { return get_bn_spatial_inputs(batch_factor); }, {4, 64, 28, 28}, tensor_elem_gen);
+            [=, this] { return get_bn_spatial_inputs(batch_factor); }, {4, 64, 28, 28}, tensor_elem_gen);
     }
 
     template <class G = tensor_elem_gen_integer>
@@ -398,7 +398,7 @@ struct test_driver
     get_bn_peract_input_tensor(G tensor_elem_gen = tensor_elem_gen_integer{})
     {
         return lazy_generate_tensor(
-            [=] { return get_bn_peract_inputs(batch_factor); }, {16, 32, 8, 8}, tensor_elem_gen);
+            [=, this] { return get_bn_peract_inputs(batch_factor); }, {16, 32, 8, 8}, tensor_elem_gen);
     }
 
     template <class G = tensor_elem_gen_integer>
@@ -406,14 +406,14 @@ struct test_driver
     get_input_tensor(G tensor_elem_gen = tensor_elem_gen_integer{})
     {
         return lazy_generate_tensor(
-            [=] { return get_inputs(batch_factor); }, {16, 32, 8, 8}, tensor_elem_gen);
+            [=, this] { return get_inputs(batch_factor); }, {16, 32, 8, 8}, tensor_elem_gen);
     }
 
     template <class G = tensor_elem_gen_integer>
     generate_tensor_t<std::vector<int>, G>
     get_3d_bn_spatial_input_tensor(G tensor_elem_gen = tensor_elem_gen_integer{})
     {
-        return lazy_generate_tensor([=] { return get_3d_bn_spatial_inputs(batch_factor); },
+        return lazy_generate_tensor([=, this] { return get_3d_bn_spatial_inputs(batch_factor); },
                                     {16, 32, 8, 8, 8},
                                     tensor_elem_gen);
     }
@@ -422,7 +422,7 @@ struct test_driver
     generate_tensor_t<std::vector<int>, G>
     get_3d_bn_peract_input_tensor(G tensor_elem_gen = tensor_elem_gen_integer{})
     {
-        return lazy_generate_tensor([=] { return get_3d_bn_peract_inputs(batch_factor); },
+        return lazy_generate_tensor([=, this] { return get_3d_bn_peract_inputs(batch_factor); },
                                     {16, 32, 8, 8, 8},
                                     tensor_elem_gen);
     }
@@ -432,7 +432,7 @@ struct test_driver
     get_weights_tensor(G tensor_elem_gen = tensor_elem_gen_integer{})
     {
         return lazy_generate_tensor(
-            [=] { return get_weights(batch_factor); }, {64, 32, 5, 5}, tensor_elem_gen);
+            [=, this] { return get_weights(batch_factor); }, {64, 32, 5, 5}, tensor_elem_gen);
     }
 
     template <class X>
@@ -456,7 +456,7 @@ struct test_driver
     template <class T>
     generate_data_t<std::vector<T>> generate_data(std::vector<T> dims, T single)
     {
-        return {[=]() -> std::vector<T> {
+        return {[=, this]() -> std::vector<T> {
             if(full_set)
                 return dims;
             else
@@ -468,7 +468,7 @@ struct test_driver
     generate_data_t<std::vector<T>>
     generate_data_limited(std::vector<T> dims, int limit_multiplier, T single)
     {
-        return {[=]() -> std::vector<T> {
+        return {[=, this]() -> std::vector<T> {
             if(full_set)
             {
                 if(limit_set > 0)
@@ -504,7 +504,7 @@ struct test_driver
     template <class T>
     generate_data_t<std::vector<T>> generate_data(std::vector<T> dims)
     {
-        return {[=]() -> std::vector<T> {
+        return {[=, this]() -> std::vector<T> {
             if(full_set)
                 return dims;
             else
@@ -515,13 +515,13 @@ struct test_driver
     template <class T>
     generate_data_t<std::vector<T>> generate_multi_data(std::vector<std::vector<T>> multi_dims)
     {
-        return {[=]() -> std::vector<T> { return generate_data(multi_dims.at(dataset_id))(T{}); }};
+        return {[=, this]() -> std::vector<T> { return generate_data(multi_dims.at(dataset_id))(T{}); }};
     }
 
     template <class T>
     generate_data_t<std::vector<T>> generate_data_limited(std::vector<T> dims, int limit_multiplier)
     {
-        return {[=]() -> std::vector<T> {
+        return {[=, this]() -> std::vector<T> {
             if(full_set)
             {
                 if(limit_set > 0)
@@ -547,7 +547,7 @@ struct test_driver
     generate_data_t<std::vector<T>>
     generate_multi_data_limited(std::vector<std::vector<T>> multi_dims, int limit_multiplier)
     {
-        return {[=]() -> std::vector<T> {
+        return {[=, this]() -> std::vector<T> {
             return generate_data_limited(multi_dims.at(dataset_id), limit_multiplier)(T{});
         }};
     }
@@ -555,7 +555,7 @@ struct test_driver
     template <class F, class T>
     auto lazy_generate_data(F f, T single) -> generate_data_t<decltype(f())>
     {
-        return {[=]() -> decltype(f()) {
+        return {[=, this]() -> decltype(f()) {
             if(full_set)
                 return f();
             else
@@ -566,7 +566,7 @@ struct test_driver
     template <class F>
     auto lazy_generate_data(F f) -> generate_data_t<decltype(f())>
     {
-        return {[=]() -> decltype(f()) {
+        return {[=, this]() -> decltype(f()) {
             if(full_set)
                 return f();
             else
@@ -577,7 +577,7 @@ struct test_driver
     template <class T>
     generate_data_t<std::vector<T>> generate_single(T single)
     {
-        return {[=]() -> std::vector<T> { return {single}; }};
+        return {[=, this]() -> std::vector<T> { return {single}; }};
     }
 
     template <class X>
@@ -613,7 +613,7 @@ struct test_driver
 
     auto verify_reporter()
     {
-        return [=](bool pass,
+        return [=, this](bool pass,
                    std::vector<double> error,
                    const auto& out_cpu,
                    const auto& out_gpu,
diff --git a/projects/miopen/test/gtest/kernel_tuning_net.cpp b/projects/miopen/test/gtest/kernel_tuning_net.cpp
index 304adb9800d4..ab103db57853 100644
--- a/projects/miopen/test/gtest/kernel_tuning_net.cpp
+++ b/projects/miopen/test/gtest/kernel_tuning_net.cpp
@@ -197,7 +197,7 @@ template <typename PerfConfig>
 class KernelTuningNetTest : public ::testing::TestWithParam<KernelTuningNetTestCase>
 {
 protected:
-    void TestParameterPredictionModel(std::string solver_nm)
+    void TestParameterPredictionModel([[maybe_unused]] std::string solver_nm)
     {
 #if MIOPEN_ENABLE_AI_KERNEL_TUNING && MIOPEN_USE_COMPOSABLEKERNEL
         auto test_case = GetParam();
diff --git a/projects/miopen/test/gtest/kthvalue.hpp b/projects/miopen/test/gtest/kthvalue.hpp
index 2aa7e6fd41d1..62d5ebb5905a 100644
--- a/projects/miopen/test/gtest/kthvalue.hpp
+++ b/projects/miopen/test/gtest/kthvalue.hpp
@@ -156,7 +156,7 @@ struct KthvalueFwdTest : public ::testing::TestWithParam<KthvalueTestCase>
                                          output.desc,
                                          output_dev.get(),
                                          indicesDesc,
-                                         (size_t*)indices_dev.get(),
+                                         static_cast<size_t*>(indices_dev.get()),
                                          config.k,
                                          config.dim,
                                          config.keepDim);
diff --git a/projects/miopen/test/gtest/na_inference_find2.cpp b/projects/miopen/test/gtest/na_inference_find2.cpp
index 0313097c2242..fa45d8eddcc8 100644
--- a/projects/miopen/test/gtest/na_inference_find2.cpp
+++ b/projects/miopen/test/gtest/na_inference_find2.cpp
@@ -152,7 +152,7 @@ struct verify_inference_batchnorm_activ
             miopenTensorArgument_t{miopenTensorBatchnormBias, nullptr, bnbias_dev.get()},
             miopenTensorArgument_t{miopenTensorBatchnormEstimatedMean, nullptr, estMean_dev.get()},
             miopenTensorArgument_t{miopenTensorBatchnormEstimatedVariance, nullptr, estVariance_dev.get()},
-            miopenTensorArgument_t{miopenScalarBatchnormEpsilon, nullptr, (void*)&mut_eps},
+            miopenTensorArgument_t{miopenScalarBatchnormEpsilon, nullptr, static_cast<void*>(&mut_eps)},
             // clang-format on
         };
 
diff --git a/projects/miopen/test/gtest/rnn_seq_api.hpp b/projects/miopen/test/gtest/rnn_seq_api.hpp
index 4f7cded37687..6699a726bd69 100644
--- a/projects/miopen/test/gtest/rnn_seq_api.hpp
+++ b/projects/miopen/test/gtest/rnn_seq_api.hpp
@@ -1823,7 +1823,7 @@ struct RNNSeqApiCommon : public ::testing::TestWithParam<RNNSeqApiParam>
         seqTensor<T> dy(output);
 
         const auto num_hidden_layers = numLayers * ((dirMode != 0) ? 2 : 1);
-        tensor<T> hx                 = [=]() {
+        tensor<T> hx                 = [=, this]() {
             if(pytorchTensorDescriptorFormat)
                 return tensor<T>(std::vector{num_hidden_layers, batchSize, hiddenSize, 1, 1});
             else