diff --git a/sycl/cmake/modules/AddSYCLUnitTest.cmake b/sycl/cmake/modules/AddSYCLUnitTest.cmake
index b4b01b4fb45ea..9f1fa77774361 100644
--- a/sycl/cmake/modules/AddSYCLUnitTest.cmake
+++ b/sycl/cmake/modules/AddSYCLUnitTest.cmake
@@ -1,6 +1,6 @@
 # Internal function to create SYCL unit tests with code reuse
-# add_sycl_unittest_internal(test_dirname SHARED|OBJECT is_preview file1.cpp, file2.cpp ...)
-function(add_sycl_unittest_internal test_dirname link_variant is_preview)
+# add_sycl_unittest_internal(test_dirname SHARED|OBJECT is_preview is_no_cgh file1.cpp, file2.cpp ...)
+function(add_sycl_unittest_internal test_dirname link_variant is_preview is_no_cgh)
   # Enable exception handling for these unit tests
   set(LLVM_REQUIRES_EH ON)
   set(LLVM_REQUIRES_RTTI ON)
@@ -34,7 +34,11 @@ function(add_sycl_unittest_internal test_dirname link_variant is_preview)
   # Chaning CMAKE_CURRENT_BINARY_DIR should not affect this variable in its
   # parent scope.
   if (${is_preview})
-    set(CMAKE_CURRENT_BINARY_DIR "${CMAKE_CURRENT_BINARY_DIR}/Preview")
+      set(CMAKE_CURRENT_BINARY_DIR "${CMAKE_CURRENT_BINARY_DIR}/Preview")
+  endif()
+
+  if (${is_no_cgh})
+    set(CMAKE_CURRENT_BINARY_DIR "${CMAKE_CURRENT_BINARY_DIR}/NoCGH")
   endif()
 
   if ("${link_variant}" MATCHES "SHARED")
@@ -65,6 +69,18 @@ function(add_sycl_unittest_internal test_dirname link_variant is_preview)
     set(sycl_cache_suffix "_preview")
   endif()
 
+  if (${is_no_cgh})
+    set(sycl_cache_suffix "_no_cgh")
+  endif()
+
+  if (${is_no_cgh})
+    target_compile_definitions(
+      ${test_dirname}
+      PRIVATE
+        __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
+    )
+  endif()
+
   if (SYCL_ENABLE_XPTI_TRACING)
     target_compile_definitions(${test_dirname}
       PRIVATE XPTI_ENABLE_INSTRUMENTATION XPTI_STATIC_LIBRARY)
@@ -150,7 +166,6 @@ function(add_sycl_unittest_internal test_dirname link_variant is_preview)
         -Wno-inconsistent-missing-override
     )
   endif()
-  
   target_compile_definitions(${test_dirname} PRIVATE SYCL_DISABLE_FSYCL_SYCLHPP_WARNING)
 endfunction()
 
@@ -160,6 +175,7 @@ endfunction()
 # the SYCL preview features enabled.
 # Produces two binaries, named `basename(test_name_prefix_non_preview)` and `basename(test_name_prefix_preview)`
 macro(add_sycl_unittest test_name_prefix link_variant)
-  add_sycl_unittest_internal(${test_name_prefix}_non_preview ${link_variant} FALSE ${ARGN})
-  add_sycl_unittest_internal(${test_name_prefix}_preview ${link_variant} TRUE ${ARGN})
+  add_sycl_unittest_internal(${test_name_prefix}_non_preview ${link_variant} FALSE FALSE ${ARGN})
+  add_sycl_unittest_internal(${test_name_prefix}_no_cgh ${link_variant} FALSE TRUE ${ARGN})
+  add_sycl_unittest_internal(${test_name_prefix}_preview ${link_variant} TRUE FALSE ${ARGN})
 endmacro()
diff --git a/sycl/include/sycl/ext/oneapi/experimental/enqueue_functions.hpp b/sycl/include/sycl/ext/oneapi/experimental/enqueue_functions.hpp
index adc48ef08a674..3c0b4606154d7 100644
--- a/sycl/include/sycl/ext/oneapi/experimental/enqueue_functions.hpp
+++ b/sycl/include/sycl/ext/oneapi/experimental/enqueue_functions.hpp
@@ -109,6 +109,26 @@ event submit_with_event_impl(const queue &Q, PropertiesT Props,
                              const sycl::detail::code_location &CodeLoc) {
   return Q.submit_with_event(Props, detail::type_erased_cgfo_ty{CGF}, CodeLoc);
 }
+
+template <typename KernelName, typename PropertiesT, typename KernelType,
+          int Dims>
+void submit_kernel_direct_impl(const queue &Q, PropertiesT Props,
+                               nd_range<Dims> Range,
+                               const KernelType &KernelFunc,
+                               const sycl::detail::code_location &CodeLoc) {
+  Q.submit_kernel_direct_without_event<KernelName, PropertiesT, KernelType,
+                                       Dims>(Props, Range, KernelFunc, CodeLoc);
+}
+
+template <typename KernelName, typename PropertiesT, typename KernelType,
+          int Dims>
+event submit_kernel_direct_with_event_impl(
+    const queue &Q, PropertiesT Props, nd_range<Dims> Range,
+    const KernelType &KernelFunc, const sycl::detail::code_location &CodeLoc) {
+  return Q.submit_kernel_direct_with_event<KernelName, PropertiesT, KernelType,
+                                           Dims>(Props, Range, KernelFunc,
+                                                 CodeLoc);
+}
 } // namespace detail
 
 template <typename CommandGroupFunc, typename PropertiesT>
@@ -126,6 +146,17 @@ void submit(const queue &Q, CommandGroupFunc &&CGF,
   submit(Q, empty_properties_t{}, std::forward<CommandGroupFunc>(CGF), CodeLoc);
 }
 
+template <typename KernelName = sycl::detail::auto_name, typename PropertiesT,
+          typename KernelType, int Dims>
+void submit(const queue &Q, PropertiesT Props, nd_range<Dims> Range,
+            const KernelType &KernelFunc,
+            const sycl::detail::code_location &CodeLoc =
+                sycl::detail::code_location::current()) {
+  sycl::ext::oneapi::experimental::detail::submit_kernel_direct_impl<
+      KernelName, PropertiesT, KernelType, Dims>(Q, Props, Range, KernelFunc,
+                                                 CodeLoc);
+}
+
 template <typename CommandGroupFunc, typename PropertiesT>
 event submit_with_event(const queue &Q, PropertiesT Props,
                         CommandGroupFunc &&CGF,
@@ -143,6 +174,18 @@ event submit_with_event(const queue &Q, CommandGroupFunc &&CGF,
                            std::forward<CommandGroupFunc>(CGF), CodeLoc);
 }
 
+template <typename KernelName = sycl::detail::auto_name, typename PropertiesT,
+          typename KernelType, int Dims>
+event submit_with_event(const queue &Q, PropertiesT Props, nd_range<Dims> Range,
+                        const KernelType &KernelFunc,
+                        const sycl::detail::code_location &CodeLoc =
+                            sycl::detail::code_location::current()) {
+  return sycl::ext::oneapi::experimental::detail::
+      submit_kernel_direct_with_event_impl<KernelName, PropertiesT, KernelType,
+                                           Dims>(Q, Props, Range, KernelFunc,
+                                                 CodeLoc);
+}
+
 template <typename KernelName = sycl::detail::auto_name, typename KernelType>
 void single_task(handler &CGH, const KernelType &KernelObj) {
   CGH.single_task<KernelName>(KernelObj);
@@ -259,10 +302,21 @@ template <typename KernelName = sycl::detail::auto_name, int Dimensions,
           typename KernelType, typename... ReductionsT>
 void nd_launch(queue Q, nd_range<Dimensions> Range, const KernelType &KernelObj,
                ReductionsT &&...Reductions) {
+#ifdef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
+  if constexpr (sizeof...(ReductionsT) == 0) {
+    submit<KernelName>(std::move(Q), empty_properties_t{}, Range, KernelObj);
+  } else {
+    submit(std::move(Q), [&](handler &CGH) {
+      nd_launch<KernelName>(CGH, Range, KernelObj,
+                            std::forward<ReductionsT>(Reductions)...);
+    });
+  }
+#else
   submit(std::move(Q), [&](handler &CGH) {
     nd_launch<KernelName>(CGH, Range, KernelObj,
                           std::forward<ReductionsT>(Reductions)...);
   });
+#endif
 }
 
 template <typename KernelName = sycl::detail::auto_name, int Dimensions,
@@ -283,10 +337,25 @@ template <typename KernelName = sycl::detail::auto_name, int Dimensions,
           typename Properties, typename KernelType, typename... ReductionsT>
 void nd_launch(queue Q, launch_config<nd_range<Dimensions>, Properties> Config,
                const KernelType &KernelObj, ReductionsT &&...Reductions) {
+#ifdef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
+  if constexpr (sizeof...(ReductionsT) == 0) {
+    ext::oneapi::experimental::detail::LaunchConfigAccess<nd_range<Dimensions>,
+                                                          Properties>
+        ConfigAccess(Config);
+    submit<KernelName>(std::move(Q), ConfigAccess.getProperties(),
+                       ConfigAccess.getRange(), KernelObj);
+  } else {
+    submit(std::move(Q), [&](handler &CGH) {
+      nd_launch<KernelName>(CGH, Config, KernelObj,
+                            std::forward<ReductionsT>(Reductions)...);
+    });
+  }
+#else
   submit(std::move(Q), [&](handler &CGH) {
     nd_launch<KernelName>(CGH, Config, KernelObj,
                           std::forward<ReductionsT>(Reductions)...);
   });
+#endif
 }
 
 template <int Dimensions, typename... ArgsT>
diff --git a/sycl/include/sycl/khr/free_function_commands.hpp b/sycl/include/sycl/khr/free_function_commands.hpp
index 4138edd5821e6..4dd7c067126a1 100644
--- a/sycl/include/sycl/khr/free_function_commands.hpp
+++ b/sycl/include/sycl/khr/free_function_commands.hpp
@@ -153,27 +153,42 @@ void launch_grouped(const queue &q, range<1> r, range<1> size,
                     const KernelType &k,
                     const sycl::detail::code_location &codeLoc =
                         sycl::detail::code_location::current()) {
+#ifdef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
+  submit(q, ext::oneapi::experimental::empty_properties_t{},
+         nd_range<1>(r, size), k);
+#else
   submit(
       q, [&](handler &h) { launch_grouped<KernelType>(h, r, size, k); },
       codeLoc);
+#endif
 }
 template <typename KernelType>
 void launch_grouped(const queue &q, range<2> r, range<2> size,
                     const KernelType &k,
                     const sycl::detail::code_location &codeLoc =
                         sycl::detail::code_location::current()) {
+#ifdef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
+  submit(q, ext::oneapi::experimental::empty_properties_t{},
+         nd_range<2>(r, size), k);
+#else
   submit(
       q, [&](handler &h) { launch_grouped<KernelType>(h, r, size, k); },
       codeLoc);
+#endif
 }
 template <typename KernelType>
 void launch_grouped(const queue &q, range<3> r, range<3> size,
                     const KernelType &k,
                     const sycl::detail::code_location &codeLoc =
                         sycl::detail::code_location::current()) {
+#ifdef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
+  submit(q, ext::oneapi::experimental::empty_properties_t{},
+         nd_range<3>(r, size), k);
+#else
   submit(
       q, [&](handler &h) { launch_grouped<KernelType>(h, r, size, k); },
       codeLoc);
+#endif
 }
 
 template <typename... Args>
diff --git a/sycl/include/sycl/queue.hpp b/sycl/include/sycl/queue.hpp
index 3eb710f32ab60..1d9b490e8f9c5 100644
--- a/sycl/include/sycl/queue.hpp
+++ b/sycl/include/sycl/queue.hpp
@@ -140,6 +140,40 @@ class __SYCL_EXPORT SubmissionInfo {
       ext::oneapi::experimental::event_mode_enum::none;
 };
 
+// This class is intended to store the kernel runtime information,
+// extracted from the compile time kernel structures.
+class __SYCL_EXPORT KernelDataDesc {
+public:
+  KernelDataDesc() {}
+
+  KernelDataDesc(const KernelDataDesc &rhs) = delete;
+
+  KernelDataDesc(KernelDataDesc &&rhs) = delete;
+
+  KernelDataDesc &operator=(const KernelDataDesc &rhs) = delete;
+
+  KernelDataDesc &operator=(KernelDataDesc &&rhs) = delete;
+
+  std::shared_ptr<detail::HostKernelBase> &HostKernel() { return MHostKernel; }
+  const std::shared_ptr<detail::HostKernelBase> &HostKernel() const {
+    return MHostKernel;
+  }
+
+  char *GetKernelFuncPtr() { return (*MHostKernel).getPtr(); }
+  char *GetKernelFuncPtr() const { return (*MHostKernel).getPtr(); }
+
+  detail::DeviceKernelInfo *&DeviceKernelInfoPtr() {
+    return MDeviceKernelInfoPtr;
+  }
+  detail::DeviceKernelInfo *DeviceKernelInfoPtr() const {
+    return MDeviceKernelInfoPtr;
+  }
+
+private:
+  std::shared_ptr<detail::HostKernelBase> MHostKernel;
+  detail::DeviceKernelInfo *MDeviceKernelInfoPtr = nullptr;
+};
+
 } // namespace v1
 } // namespace detail
 
@@ -158,6 +192,19 @@ template <typename CommandGroupFunc, typename PropertiesT>
 event submit_with_event_impl(const queue &Q, PropertiesT Props,
                              CommandGroupFunc &&CGF,
                              const sycl::detail::code_location &CodeLoc);
+
+template <typename KernelName, typename PropertiesT, typename KernelType,
+          int Dims>
+void submit_kernel_direct_impl(const queue &Q, PropertiesT Props,
+                               nd_range<Dims> Range,
+                               const KernelType &KernelFunc,
+                               const sycl::detail::code_location &CodeLoc);
+
+template <typename KernelName, typename PropertiesT, typename KernelType,
+          int Dims>
+event submit_kernel_direct_with_event_impl(
+    const queue &Q, PropertiesT Props, nd_range<Dims> Range,
+    const KernelType &KernelFunc, const sycl::detail::code_location &CodeLoc);
 } // namespace detail
 } // namespace ext::oneapi::experimental
 
@@ -2650,12 +2697,21 @@ class __SYCL_EXPORT queue : public detail::OwnerLessBase<queue> {
         "Use queue.submit() instead");
 
     detail::tls_code_loc_t TlsCodeLocCapture(CodeLoc);
+
+#ifdef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
+    return submit_kernel_direct_with_event<KernelName, PropertiesT, KernelType,
+                                           1, detail::WrapAs::single_task,
+                                           void>(Properties, nd_range<1>{1, 1},
+                                                 KernelFunc);
+
+#else
     return submit(
         [&](handler &CGH) {
           CGH.template single_task<KernelName, KernelType, PropertiesT>(
               Properties, KernelFunc);
         },
         TlsCodeLocCapture.query());
+#endif
   }
 
   /// single_task version with a kernel represented as a lambda.
@@ -3203,11 +3259,24 @@ class __SYCL_EXPORT queue : public detail::OwnerLessBase<queue> {
   parallel_for(nd_range<Dims> Range, RestT &&...Rest) {
     constexpr detail::code_location CodeLoc = getCodeLocation<KernelName>();
     detail::tls_code_loc_t TlsCodeLocCapture(CodeLoc);
+#ifdef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
+    if constexpr (sizeof...(RestT) == 1) {
+      return submit_kernel_direct_with_event<KernelName>(
+          ext::oneapi::experimental::empty_properties_t{}, Range, Rest...);
+    } else {
+      return submit(
+          [&](handler &CGH) {
+            CGH.template parallel_for<KernelName>(Range, Rest...);
+          },
+          TlsCodeLocCapture.query());
+    }
+#else
     return submit(
         [&](handler &CGH) {
           CGH.template parallel_for<KernelName>(Range, Rest...);
         },
         TlsCodeLocCapture.query());
+#endif
   }
 
   /// parallel_for version with a kernel represented as a lambda + nd_range that
@@ -3586,6 +3655,19 @@ class __SYCL_EXPORT queue : public detail::OwnerLessBase<queue> {
       const queue &Q, PropertiesT Props, CommandGroupFunc &&CGF,
       const sycl::detail::code_location &CodeLoc);
 
+  template <typename KernelName, typename PropertiesT, typename KernelType,
+            int Dims>
+  friend void ext::oneapi::experimental::detail::submit_kernel_direct_impl(
+      const queue &Q, PropertiesT Props, nd_range<Dims> Range,
+      const KernelType &KernelFunc, const sycl::detail::code_location &CodeLoc);
+
+  template <typename KernelName, typename PropertiesT, typename KernelType,
+            int Dims>
+  friend event
+  ext::oneapi::experimental::detail::submit_kernel_direct_with_event_impl(
+      const queue &Q, PropertiesT Props, nd_range<Dims> Range,
+      const KernelType &KernelFunc, const sycl::detail::code_location &CodeLoc);
+
   template <typename PropertiesT>
   void ProcessSubmitProperties(PropertiesT Props,
                                detail::v1::SubmissionInfo &SI) const {
@@ -3599,6 +3681,33 @@ class __SYCL_EXPORT queue : public detail::OwnerLessBase<queue> {
     }
   }
 
+  template <int Dims, typename LambdaArgType> struct TransformUserItemType {
+    using type = std::conditional_t<
+        std::is_convertible_v<nd_item<Dims>, LambdaArgType>, nd_item<Dims>,
+        std::conditional_t<std::is_convertible_v<item<Dims>, LambdaArgType>,
+                           item<Dims>, LambdaArgType>>;
+  };
+
+  template <typename KernelName, typename KernelType, int Dims,
+            detail::WrapAs WrapAsVal>
+  void ProcessKernelDataDesc(const KernelType &KernelFunc,
+                             detail::v1::KernelDataDesc &KDDesc) const {
+
+    using LambdaArgType = sycl::detail::lambda_arg_type<KernelType, item<Dims>>;
+    using TransformedArgType = std::conditional_t<
+        WrapAsVal == detail::WrapAs::parallel_for,
+        std::conditional_t<
+            std::is_integral<LambdaArgType>::value && Dims == 1, item<Dims>,
+            typename TransformUserItemType<Dims, LambdaArgType>::type>,
+        void>;
+
+    KDDesc.HostKernel().reset(
+        new detail::HostKernel<KernelType, TransformedArgType, Dims>(
+            KernelFunc));
+
+    KDDesc.DeviceKernelInfoPtr() = &detail::getDeviceKernelInfo<KernelName>();
+  }
+
 #ifndef __INTEL_PREVIEW_BREAKING_CHANGES
   /// TODO: Unused. Remove these when ABI-break window is open.
   /// Not using `type_erased_cgfo_ty` on purpose.
@@ -3670,6 +3779,30 @@ class __SYCL_EXPORT queue : public detail::OwnerLessBase<queue> {
                                const detail::code_location &CodeLoc,
                                bool IsTopCodeLoc) const;
 
+  event submit_kernel_direct_with_event_impl(
+      nd_range<1> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) const;
+
+  event submit_kernel_direct_with_event_impl(
+      nd_range<2> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) const;
+
+  event submit_kernel_direct_with_event_impl(
+      nd_range<3> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) const;
+
+  void submit_kernel_direct_without_event_impl(
+      nd_range<1> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) const;
+
+  void submit_kernel_direct_without_event_impl(
+      nd_range<2> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) const;
+
+  void submit_kernel_direct_without_event_impl(
+      nd_range<3> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) const;
+
   /// A template-free version of submit_without_event as const member function.
   void submit_without_event_impl(const detail::type_erased_cgfo_ty &CGH,
                                  const detail::v1::SubmissionInfo &SubmitInfo,
@@ -3699,6 +3832,57 @@ class __SYCL_EXPORT queue : public detail::OwnerLessBase<queue> {
                                   TlsCodeLocCapture.isToplevel());
   }
 
+  template <typename KernelName = detail::auto_name, typename PropertiesT,
+            typename KernelType, int Dims,
+            detail::WrapAs WrapAsVal = detail::WrapAs::parallel_for,
+            typename ElementType = sycl::nd_item<Dims>>
+  event
+  submit_kernel_direct_with_event(PropertiesT Props, nd_range<Dims> Range,
+                                  const KernelType &KernelFunc,
+                                  const detail::code_location &CodeLoc =
+                                      detail::code_location::current()) const {
+    (void)Props;
+    detail::tls_code_loc_t TlsCodeLocCapture(CodeLoc);
+    detail::v1::KernelDataDesc KDDesc{};
+
+    using NameT =
+        typename detail::get_kernel_name_t<KernelName, KernelType>::name;
+
+    ProcessKernelDataDesc<NameT, KernelType, Dims, WrapAsVal>(KernelFunc,
+                                                              KDDesc);
+
+    detail::KernelWrapper<WrapAsVal, NameT, KernelType, ElementType,
+                          PropertiesT>::wrap(KernelFunc);
+
+    return submit_kernel_direct_with_event_impl(Range, KDDesc,
+                                                TlsCodeLocCapture.query(),
+                                                TlsCodeLocCapture.isToplevel());
+  }
+
+  template <typename KernelName = detail::auto_name, typename PropertiesT,
+            typename KernelType, int Dims>
+  void submit_kernel_direct_without_event(
+      PropertiesT Props, nd_range<Dims> Range, const KernelType &KernelFunc,
+      const detail::code_location &CodeLoc =
+          detail::code_location::current()) const {
+    (void)Props;
+    detail::tls_code_loc_t TlsCodeLocCapture(CodeLoc);
+    detail::v1::KernelDataDesc KDDesc{};
+
+    using NameT =
+        typename detail::get_kernel_name_t<KernelName, KernelType>::name;
+
+    ProcessKernelDataDesc<NameT, KernelType, Dims,
+                          detail::WrapAs::parallel_for>(KernelFunc, KDDesc);
+
+    detail::KernelWrapper<detail::WrapAs::parallel_for, NameT, KernelType,
+                          sycl::nd_item<Dims>, PropertiesT>::wrap(KernelFunc);
+
+    submit_kernel_direct_without_event_impl(Range, KDDesc,
+                                            TlsCodeLocCapture.query(),
+                                            TlsCodeLocCapture.isToplevel());
+  }
+
   /// Submits a command group function object to the queue, in order to be
   /// scheduled for execution on the device.
   ///
diff --git a/sycl/source/detail/queue_impl.cpp b/sycl/source/detail/queue_impl.cpp
index c59b5eaa20387..6b0f98fd91b4e 100644
--- a/sycl/source/detail/queue_impl.cpp
+++ b/sycl/source/detail/queue_impl.cpp
@@ -417,6 +417,135 @@ queue_impl::submit_impl(const detail::type_erased_cgfo_ty &CGF,
   return EventImpl;
 }
 
+std::vector<ArgDesc> queue_impl::extractArgsAndReqsFromLambda(
+    char *LambdaPtr, detail::kernel_param_desc_t (*ParamDescGetter)(int),
+    size_t NumKernelParams) {
+
+  size_t IndexShift = 0;
+  std::vector<ArgDesc> Args;
+
+  Args.reserve(NumKernelParams);
+
+  for (size_t I = 0; I < NumKernelParams; ++I) {
+    detail::kernel_param_desc_t ParamDesc = ParamDescGetter(I);
+    void *Ptr = LambdaPtr + ParamDesc.offset;
+    const detail::kernel_param_kind_t &Kind = ParamDesc.kind;
+    const int &Size = ParamDesc.info;
+
+    Args.emplace_back(Kind, Ptr, Size, I + IndexShift);
+  }
+
+  return Args;
+}
+
+detail::EventImplPtr queue_impl::submit_kernel_direct_impl(
+    const NDRDescT &NDRDesc, const v1::KernelDataDesc &KDDesc,
+    bool CallerNeedsEvent, const detail::code_location &CodeLoc,
+    bool IsTopCodeLoc) {
+
+  KernelData KData;
+
+  KData.setDeviceKernelInfoPtr(KDDesc.DeviceKernelInfoPtr());
+  KData.setKernelFunc(KDDesc.GetKernelFuncPtr());
+  KData.setNDRDesc(NDRDesc);
+
+  return submit_kernel_direct_impl(KData, KDDesc.HostKernel(), CallerNeedsEvent,
+                                   CodeLoc, IsTopCodeLoc);
+}
+
+detail::EventImplPtr queue_impl::submit_kernel_direct_impl(
+    KernelData &KData, std::shared_ptr<detail::HostKernelBase> HostKernel,
+    bool CallerNeedsEvent, const detail::code_location &CodeLoc,
+    bool IsTopCodeLoc) {
+
+  auto SubmitKernelFunc =
+      [&](detail::CG::StorageInitHelper &CGData) -> EventImplPtr {
+    std::unique_ptr<detail::CG> CommandGroup;
+    std::vector<std::shared_ptr<detail::stream_impl>> StreamStorage;
+    std::vector<std::shared_ptr<const void>> AuxiliaryResources;
+
+    KData.extractArgsAndReqsFromLambda();
+
+    CommandGroup.reset(new detail::CGExecKernel(
+        KData.getNDRDesc(), HostKernel,
+        nullptr, // Kernel
+        nullptr, // KernelBundle
+        std::move(CGData), std::move(KData).getArgs(),
+        *KData.getDeviceKernelInfoPtr(), std::move(StreamStorage),
+        std::move(AuxiliaryResources), detail::CGType::Kernel,
+        UR_KERNEL_CACHE_CONFIG_DEFAULT,
+        false, // KernelIsCooperative
+        false, // KernelUsesClusterLaunch
+        0,     // KernelWorkGroupMemorySize
+        CodeLoc));
+    CommandGroup->MIsTopCodeLoc = IsTopCodeLoc;
+
+    return detail::Scheduler::getInstance().addCG(std::move(CommandGroup),
+                                                  *this, true);
+  };
+
+  return submit_direct(CallerNeedsEvent, SubmitKernelFunc);
+}
+
+template <typename SubmitCommandFuncType>
+detail::EventImplPtr
+queue_impl::submit_direct(bool CallerNeedsEvent,
+                          SubmitCommandFuncType &SubmitCommandFunc) {
+  detail::CG::StorageInitHelper CGData;
+  std::unique_lock<std::mutex> Lock(MMutex);
+
+  // Graphs are not supported yet for the no-handler path
+  assert(!hasCommandGraph());
+
+  // Set the No Last Event Mode to false, since the no-handler path
+  // does not support it yet.
+  MNoLastEventMode.store(false, std::memory_order_relaxed);
+
+  // Used by queue_empty() and getLastEvent()
+  MEmpty.store(false, std::memory_order_release);
+
+  // Sync with an external event
+  std::optional<event> ExternalEvent = popExternalEvent();
+  if (ExternalEvent) {
+    CGData.MEvents.push_back(getSyclObjImpl(*ExternalEvent));
+  }
+
+  // Sync with the last event for in order queue
+  EventImplPtr &LastEvent = MDefaultGraphDeps.LastEventPtr;
+  if (isInOrder() && LastEvent) {
+    CGData.MEvents.push_back(LastEvent);
+  }
+
+  // Barrier and un-enqueued commands synchronization for out or order queue
+  if (!isInOrder()) {
+    MMissedCleanupRequests.unset(
+        [&](MissedCleanupRequestsType &MissedCleanupRequests) {
+          for (auto &UpdatedGraph : MissedCleanupRequests)
+            doUnenqueuedCommandCleanup(UpdatedGraph);
+          MissedCleanupRequests.clear();
+        });
+
+    if (MDefaultGraphDeps.LastBarrier &&
+        !MDefaultGraphDeps.LastBarrier->isEnqueued()) {
+      CGData.MEvents.push_back(MDefaultGraphDeps.LastBarrier);
+    }
+  }
+
+  EventImplPtr EventImpl = SubmitCommandFunc(CGData);
+
+  // Sync with the last event for in order queue
+  if (isInOrder() && !EventImpl->isDiscarded()) {
+    LastEvent = EventImpl;
+  }
+
+  // Barrier and un-enqueued commands synchronization for out or order queue
+  if (!isInOrder() && !EventImpl->isEnqueued()) {
+    MDefaultGraphDeps.UnenqueuedCmdEvents.push_back(EventImpl);
+  }
+
+  return CallerNeedsEvent ? EventImpl : nullptr;
+}
+
 template <typename HandlerFuncT>
 event queue_impl::submitWithHandler(const std::vector<event> &DepEvents,
                                     bool CallerNeedsEvent,
diff --git a/sycl/source/detail/queue_impl.hpp b/sycl/source/detail/queue_impl.hpp
index 2b1d8f637b05d..dc5ea2e72f35e 100644
--- a/sycl/source/detail/queue_impl.hpp
+++ b/sycl/source/detail/queue_impl.hpp
@@ -359,6 +359,51 @@ class queue_impl : public std::enable_shared_from_this<queue_impl> {
     return createSyclObjFromImpl<event>(ResEvent);
   }
 
+  event submit_kernel_direct_with_event(
+      nd_range<1> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) {
+    detail::EventImplPtr EventImpl = submit_kernel_direct_impl(
+        NDRDescT{Range}, KDDesc, true, CodeLoc, IsTopCodeLoc);
+    return createSyclObjFromImpl<event>(EventImpl);
+  }
+
+  event submit_kernel_direct_with_event(
+      nd_range<2> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) {
+    detail::EventImplPtr EventImpl = submit_kernel_direct_impl(
+        NDRDescT{Range}, KDDesc, true, CodeLoc, IsTopCodeLoc);
+    return createSyclObjFromImpl<event>(EventImpl);
+  }
+
+  event submit_kernel_direct_with_event(
+      nd_range<3> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) {
+    detail::EventImplPtr EventImpl = submit_kernel_direct_impl(
+        NDRDescT{Range}, KDDesc, true, CodeLoc, IsTopCodeLoc);
+    return createSyclObjFromImpl<event>(EventImpl);
+  }
+
+  void submit_kernel_direct_without_event(
+      nd_range<1> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) {
+    submit_kernel_direct_impl(NDRDescT{Range}, KDDesc, false, CodeLoc,
+                              IsTopCodeLoc);
+  }
+
+  void submit_kernel_direct_without_event(
+      nd_range<2> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) {
+    submit_kernel_direct_impl(NDRDescT{Range}, KDDesc, false, CodeLoc,
+                              IsTopCodeLoc);
+  }
+
+  void submit_kernel_direct_without_event(
+      nd_range<3> Range, const detail::v1::KernelDataDesc &KDDesc,
+      const detail::code_location &CodeLoc, bool IsTopCodeLoc) {
+    submit_kernel_direct_impl(NDRDescT{Range}, KDDesc, false, CodeLoc,
+                              IsTopCodeLoc);
+  }
+
   void submit_without_event(const detail::type_erased_cgfo_ty &CGF,
                             const v1::SubmissionInfo &SubmitInfo,
                             const detail::code_location &Loc,
@@ -870,6 +915,46 @@ class queue_impl : public std::enable_shared_from_this<queue_impl> {
                                    bool IsTopCodeLoc,
                                    const v1::SubmissionInfo &SubmitInfo);
 
+  std::vector<ArgDesc> extractArgsAndReqsFromLambda(
+      char *LambdaPtr, detail::kernel_param_desc_t (*ParamDescGetter)(int),
+      size_t NumKernelParams);
+
+  /// Performs kernel submission to the queue.
+  ///
+  /// \param NDRDesc is an NDRange descriptor
+  /// \param KDDesc is a descriptor of the kernel
+  /// \param CallerNeedsEvent is a boolean indicating whether the event is
+  ///        required by the user after the call.
+  /// \param CodeLoc is the code location of the submit call
+  /// \param IsTopCodeLoc Used to determine if the object is in a local
+  ///        scope or in the top level scope.
+  ///
+  /// \return a SYCL event representing submitted command group or nullptr.
+  detail::EventImplPtr submit_kernel_direct_impl(
+      const NDRDescT &NDRDesc, const v1::KernelDataDesc &KDDesc,
+      bool CallerNeedsEvent, const detail::code_location &CodeLoc,
+      bool IsTopCodeLoc);
+
+  /// Performs kernel submission to the queue.
+  ///
+  /// \param KData contains aggregated data related to the kernel
+  /// \param HostKernel stores the kernel lambda instance
+  /// \param CallerNeedsEvent is a boolean indicating whether the event is
+  ///        required by the user after the call.
+  /// \param CodeLoc is the code location of the submit call
+  /// \param IsTopCodeLoc Used to determine if the object is in a local
+  ///        scope or in the top level scope.
+  ///
+  /// \return a SYCL event representing submitted command group or nullptr.
+  detail::EventImplPtr submit_kernel_direct_impl(
+      KernelData &KData, std::shared_ptr<detail::HostKernelBase> HostKernel,
+      bool CallerNeedsEvent, const detail::code_location &CodeLoc,
+      bool IsTopCodeLoc);
+
+  template <typename SubmitCommandFuncType>
+  detail::EventImplPtr submit_direct(bool CallerNeedsEvent,
+                                     SubmitCommandFuncType &SubmitCommandFunc);
+
   /// Helper function for submitting a memory operation with a handler.
   /// \param DepEvents is a vector of dependencies of the operation.
   /// \param HandlerFunc is a function that submits the operation with a
diff --git a/sycl/source/queue.cpp b/sycl/source/queue.cpp
index b73c24091d4ed..cb7da34228098 100644
--- a/sycl/source/queue.cpp
+++ b/sycl/source/queue.cpp
@@ -319,6 +319,48 @@ event queue::submit_with_event_impl(
   return impl->submit_with_event(CGH, SubmitInfo, CodeLoc, IsTopCodeLoc);
 }
 
+event queue::submit_kernel_direct_with_event_impl(
+    nd_range<1> Range, const detail::v1::KernelDataDesc &KDDesc,
+    const detail::code_location &CodeLoc, bool IsTopCodeLoc) const {
+  return impl->submit_kernel_direct_with_event(Range, KDDesc, CodeLoc,
+                                               IsTopCodeLoc);
+}
+
+event queue::submit_kernel_direct_with_event_impl(
+    nd_range<2> Range, const detail::v1::KernelDataDesc &KDDesc,
+    const detail::code_location &CodeLoc, bool IsTopCodeLoc) const {
+  return impl->submit_kernel_direct_with_event(Range, KDDesc, CodeLoc,
+                                               IsTopCodeLoc);
+}
+
+event queue::submit_kernel_direct_with_event_impl(
+    nd_range<3> Range, const detail::v1::KernelDataDesc &KDDesc,
+    const detail::code_location &CodeLoc, bool IsTopCodeLoc) const {
+  return impl->submit_kernel_direct_with_event(Range, KDDesc, CodeLoc,
+                                               IsTopCodeLoc);
+}
+
+void queue::submit_kernel_direct_without_event_impl(
+    nd_range<1> Range, const detail::v1::KernelDataDesc &KDDesc,
+    const detail::code_location &CodeLoc, bool IsTopCodeLoc) const {
+  impl->submit_kernel_direct_without_event(Range, KDDesc, CodeLoc,
+                                           IsTopCodeLoc);
+}
+
+void queue::submit_kernel_direct_without_event_impl(
+    nd_range<2> Range, const detail::v1::KernelDataDesc &KDDesc,
+    const detail::code_location &CodeLoc, bool IsTopCodeLoc) const {
+  impl->submit_kernel_direct_without_event(Range, KDDesc, CodeLoc,
+                                           IsTopCodeLoc);
+}
+
+void queue::submit_kernel_direct_without_event_impl(
+    nd_range<3> Range, const detail::v1::KernelDataDesc &KDDesc,
+    const detail::code_location &CodeLoc, bool IsTopCodeLoc) const {
+  impl->submit_kernel_direct_without_event(Range, KDDesc, CodeLoc,
+                                           IsTopCodeLoc);
+}
+
 void queue::submit_without_event_impl(
     const detail::type_erased_cgfo_ty &CGH,
     const detail::v1::SubmissionInfo &SubmitInfo,
diff --git a/sycl/test/abi/sycl_symbols_linux.dump b/sycl/test/abi/sycl_symbols_linux.dump
index 0481c5b393302..98515b118ef88 100644
--- a/sycl/test/abi/sycl_symbols_linux.dump
+++ b/sycl/test/abi/sycl_symbols_linux.dump
@@ -3755,6 +3755,12 @@ _ZNK4sycl3_V15queue22submit_with_event_implERKNS0_6detail19type_erased_cgfo_tyER
 _ZNK4sycl3_V15queue25submit_without_event_implERKNS0_6detail19type_erased_cgfo_tyERKNS2_2v114SubmissionInfoERKNS2_13code_locationEb
 _ZNK4sycl3_V15queue28ext_codeplay_supports_fusionEv
 _ZNK4sycl3_V15queue30ext_oneapi_get_last_event_implEv
+_ZNK4sycl3_V15queue36submit_kernel_direct_with_event_implENS0_8nd_rangeILi1EEERKNS0_6detail2v117KernelRuntimeInfoERKNS4_13code_locationEb
+_ZNK4sycl3_V15queue36submit_kernel_direct_with_event_implENS0_8nd_rangeILi2EEERKNS0_6detail2v117KernelRuntimeInfoERKNS4_13code_locationEb
+_ZNK4sycl3_V15queue36submit_kernel_direct_with_event_implENS0_8nd_rangeILi3EEERKNS0_6detail2v117KernelRuntimeInfoERKNS4_13code_locationEb
+_ZNK4sycl3_V15queue39submit_kernel_direct_without_event_implENS0_8nd_rangeILi1EEERKNS0_6detail2v117KernelRuntimeInfoERKNS4_13code_locationEb
+_ZNK4sycl3_V15queue39submit_kernel_direct_without_event_implENS0_8nd_rangeILi2EEERKNS0_6detail2v117KernelRuntimeInfoERKNS4_13code_locationEb
+_ZNK4sycl3_V15queue39submit_kernel_direct_without_event_implENS0_8nd_rangeILi3EEERKNS0_6detail2v117KernelRuntimeInfoERKNS4_13code_locationEb
 _ZNK4sycl3_V15queue3getEv
 _ZNK4sycl3_V15queue8get_infoINS0_4info5queue15reference_countEEENS0_6detail18is_queue_info_descIT_E11return_typeEv
 _ZNK4sycl3_V15queue8get_infoINS0_4info5queue6deviceEEENS0_6detail18is_queue_info_descIT_E11return_typeEv
diff --git a/sycl/test/abi/sycl_symbols_windows.dump b/sycl/test/abi/sycl_symbols_windows.dump
index 2c6be6c1684bb..bce53d9f078b7 100644
--- a/sycl/test/abi/sycl_symbols_windows.dump
+++ b/sycl/test/abi/sycl_symbols_windows.dump
@@ -298,6 +298,7 @@
 ??0AccessorBaseHost@detail@_V1@sycl@@QEAA@V?$id@$02@23@V?$range@$02@23@1W4mode@access@23@PEAXHH_K_NAEBVproperty_list@23@@Z
 ??0AccessorBaseHost@detail@_V1@sycl@@QEAA@V?$id@$02@23@V?$range@$02@23@1W4mode@access@23@PEAXHH_N_K4AEBVproperty_list@23@@Z
 ??0HostProfilingInfo@detail@_V1@sycl@@QEAA@XZ
+??0KernelRuntimeInfo@v1@detail@_V1@sycl@@QEAA@XZ
 ??0LocalAccessorBaseHost@detail@_V1@sycl@@IEAA@AEBV?$shared_ptr@VLocalAccessorImplHost@detail@_V1@sycl@@@std@@@Z
 ??0LocalAccessorBaseHost@detail@_V1@sycl@@QEAA@$$QEAV0123@@Z
 ??0LocalAccessorBaseHost@detail@_V1@sycl@@QEAA@AEBV0123@@Z
@@ -508,6 +509,7 @@
 ??0tls_code_loc_t@detail@_V1@sycl@@QEAA@AEBUcode_location@123@@Z
 ??0tls_code_loc_t@detail@_V1@sycl@@QEAA@XZ
 ??1AccessorBaseHost@detail@_V1@sycl@@QEAA@XZ
+??1KernelRuntimeInfo@v1@detail@_V1@sycl@@QEAA@XZ
 ??1LocalAccessorBaseHost@detail@_V1@sycl@@QEAA@XZ
 ??1SYCLCategory@detail@_V1@sycl@@UEAA@XZ
 ??1SampledImageAccessorBaseHost@detail@_V1@sycl@@QEAA@XZ
@@ -711,6 +713,8 @@
 ??_Fqueue@_V1@sycl@@QEAAXXZ
 ?AccessTargetMask@handler@_V1@sycl@@0HB
 ?Clear@exception_list@_V1@sycl@@AEAAXXZ
+?DeviceKernelInfoPtr@KernelRuntimeInfo@v1@detail@_V1@sycl@@QEAAAEAPEAVDeviceKernelInfo@345@XZ
+?DeviceKernelInfoPtr@KernelRuntimeInfo@v1@detail@_V1@sycl@@QEBAPEAVDeviceKernelInfo@345@XZ
 ?DirSep@OSUtil@detail@_V1@sycl@@2QEBDEB
 ?DisableRangeRounding@handler@_V1@sycl@@AEAA_NXZ
 ?EventMode@SubmissionInfo@detail@_V1@sycl@@QEAAAEAW4event_mode_enum@experimental@oneapi@ext@34@XZ
@@ -719,8 +723,14 @@
 ?EventMode@SubmissionInfo@v1@detail@_V1@sycl@@QEBAAEBW4event_mode_enum@experimental@oneapi@ext@45@XZ
 ?GDBMethodsAnchor@SampledImageAccessorBaseHost@detail@_V1@sycl@@IEAAXXZ
 ?GDBMethodsAnchor@UnsampledImageAccessorBaseHost@detail@_V1@sycl@@IEAAXXZ
+?GetKernelFuncPtr@KernelRuntimeInfo@v1@detail@_V1@sycl@@QEAAPEADXZ
+?GetKernelFuncPtr@KernelRuntimeInfo@v1@detail@_V1@sycl@@QEBAPEADXZ
 ?GetRangeRoundingSettings@handler@_V1@sycl@@AEAAXAEA_K00@Z
 ?HasAssociatedAccessor@handler@_V1@sycl@@AEBA_NPEAVAccessorImplHost@detail@23@W4target@access@23@@Z
+?HostKernel@KernelRuntimeInfo@v1@detail@_V1@sycl@@QEAAAEAV?$shared_ptr@VHostKernelBase@detail@_V1@sycl@@@std@@XZ
+?HostKernel@KernelRuntimeInfo@v1@detail@_V1@sycl@@QEBAAEBV?$shared_ptr@VHostKernelBase@detail@_V1@sycl@@@std@@XZ
+?KernelName@KernelRuntimeInfo@v1@detail@_V1@sycl@@QEAAAEAVstring@345@XZ
+?KernelName@KernelRuntimeInfo@v1@detail@_V1@sycl@@QEBAAEBVstring@345@XZ
 ?PostProcessorFunc@SubmissionInfo@detail@_V1@sycl@@QEAAAEAV?$optional@V?$function@$$A6AX_N0AEAVevent@_V1@sycl@@@Z@std@@@234@XZ
 ?PostProcessorFunc@SubmissionInfo@detail@_V1@sycl@@QEBAAEBV?$optional@V?$function@$$A6AX_N0AEAVevent@_V1@sycl@@@Z@std@@@234@XZ
 ?PostProcessorFunc@SubmissionInfo@v1@detail@_V1@sycl@@QEAAAEAV?$optional@V?$function@$$A6AX_N0AEAVevent@_V1@sycl@@@Z@std@@@345@XZ
@@ -4467,6 +4477,12 @@
 ?storeRawArg@handler@_V1@sycl@@AEAAPEAXAEBVraw_kernel_arg@experimental@oneapi@ext@23@@Z
 ?storeRawArg@handler@_V1@sycl@@AEAAPEAXPEBX_K@Z
 ?stringifyErrorCode@detail@_V1@sycl@@YAPEBDH@Z
+?submit_direct_with_event_impl@queue@_V1@sycl@@AEBA?AVevent@23@V?$nd_range@$00@23@AEBVSubmissionInfo@v1@detail@23@AEBVKernelRuntimeInfo@7823@AEBUcode_location@823@_N@Z
+?submit_direct_with_event_impl@queue@_V1@sycl@@AEBA?AVevent@23@V?$nd_range@$01@23@AEBVSubmissionInfo@v1@detail@23@AEBVKernelRuntimeInfo@7823@AEBUcode_location@823@_N@Z
+?submit_direct_with_event_impl@queue@_V1@sycl@@AEBA?AVevent@23@V?$nd_range@$02@23@AEBVSubmissionInfo@v1@detail@23@AEBVKernelRuntimeInfo@7823@AEBUcode_location@823@_N@Z
+?submit_direct_without_event_impl@queue@_V1@sycl@@AEBAXV?$nd_range@$00@23@AEBVSubmissionInfo@v1@detail@23@AEBVKernelRuntimeInfo@6723@AEBUcode_location@723@_N@Z
+?submit_direct_without_event_impl@queue@_V1@sycl@@AEBAXV?$nd_range@$01@23@AEBVSubmissionInfo@v1@detail@23@AEBVKernelRuntimeInfo@6723@AEBUcode_location@723@_N@Z
+?submit_direct_without_event_impl@queue@_V1@sycl@@AEBAXV?$nd_range@$02@23@AEBVSubmissionInfo@v1@detail@23@AEBVKernelRuntimeInfo@6723@AEBUcode_location@723@_N@Z
 ?submit_impl@queue@_V1@sycl@@AEAA?AVevent@23@V?$function@$$A6AXAEAVhandler@_V1@sycl@@@Z@std@@AEBUcode_location@detail@23@@Z
 ?submit_impl@queue@_V1@sycl@@AEAA?AVevent@23@V?$function@$$A6AXAEAVhandler@_V1@sycl@@@Z@std@@AEBUcode_location@detail@23@_N@Z
 ?submit_impl@queue@_V1@sycl@@AEAA?AVevent@23@V?$function@$$A6AXAEAVhandler@_V1@sycl@@@Z@std@@V123@AEBUcode_location@detail@23@@Z
diff --git a/sycl/unittests/Extensions/CommandGraph/CommandGraph.cpp b/sycl/unittests/Extensions/CommandGraph/CommandGraph.cpp
index dedd4ebbcb407..f8cb84900d4f8 100644
--- a/sycl/unittests/Extensions/CommandGraph/CommandGraph.cpp
+++ b/sycl/unittests/Extensions/CommandGraph/CommandGraph.cpp
@@ -626,6 +626,8 @@ TEST_F(CommandGraphTest, AccessorModeEdges) {
 
 // Tests the transitive queue recording behaviour with queue shortcuts.
 TEST_F(CommandGraphTest, TransitiveRecordingShortcuts) {
+// Graphs not supported yet for the no-handler submit path
+#ifndef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
   device Dev;
   context Ctx{{Dev}};
   queue Q1{Ctx, Dev};
@@ -669,6 +671,7 @@ TEST_F(CommandGraphTest, TransitiveRecordingShortcuts) {
             ext::oneapi::experimental::queue_state::executing);
   ASSERT_EQ(Q3.ext_oneapi_get_state(),
             ext::oneapi::experimental::queue_state::executing);
+#endif
 }
 
 // Tests that dynamic_work_group_memory.get() will throw on the host side.
diff --git a/sycl/unittests/Extensions/FreeFunctionCommands/FreeFunctionEventsHelpers.hpp b/sycl/unittests/Extensions/FreeFunctionCommands/FreeFunctionEventsHelpers.hpp
index c45d72ea4c343..6272b3d7d7a8a 100644
--- a/sycl/unittests/Extensions/FreeFunctionCommands/FreeFunctionEventsHelpers.hpp
+++ b/sycl/unittests/Extensions/FreeFunctionCommands/FreeFunctionEventsHelpers.hpp
@@ -26,40 +26,52 @@ inline ur_result_t after_urKernelGetInfo(void *pParams) {
 static thread_local size_t counter_urEnqueueKernelLaunch = 0;
 inline ur_result_t redefined_urEnqueueKernelLaunch(void *pParams) {
   ++counter_urEnqueueKernelLaunch;
+// The no-handler scheduler submission includes a fix for the event return,
+// which is not yet applied to the handler-based path
+#ifndef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
   auto params = *static_cast<ur_enqueue_kernel_launch_params_t *>(pParams);
   EXPECT_EQ(*params.pphEvent, nullptr);
+#endif
   return UR_RESULT_SUCCESS;
 }
 
 static thread_local size_t counter_urUSMEnqueueMemcpy = 0;
 inline ur_result_t redefined_urUSMEnqueueMemcpy(void *pParams) {
   ++counter_urUSMEnqueueMemcpy;
+#ifndef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
   auto params = *static_cast<ur_enqueue_usm_memcpy_params_t *>(pParams);
   EXPECT_EQ(*params.pphEvent, nullptr);
+#endif
   return UR_RESULT_SUCCESS;
 }
 
 static thread_local size_t counter_urUSMEnqueueFill = 0;
 inline ur_result_t redefined_urUSMEnqueueFill(void *pParams) {
   ++counter_urUSMEnqueueFill;
+#ifndef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
   auto params = *static_cast<ur_enqueue_usm_fill_params_t *>(pParams);
   EXPECT_EQ(*params.pphEvent, nullptr);
+#endif
   return UR_RESULT_SUCCESS;
 }
 
 static thread_local size_t counter_urUSMEnqueuePrefetch = 0;
 inline ur_result_t redefined_urUSMEnqueuePrefetch(void *pParams) {
   ++counter_urUSMEnqueuePrefetch;
+#ifndef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
   auto params = *static_cast<ur_enqueue_usm_prefetch_params_t *>(pParams);
   EXPECT_EQ(*params.pphEvent, nullptr);
+#endif
   return UR_RESULT_SUCCESS;
 }
 
 static thread_local size_t counter_urUSMEnqueueMemAdvise = 0;
 inline ur_result_t redefined_urUSMEnqueueMemAdvise(void *pParams) {
   ++counter_urUSMEnqueueMemAdvise;
+#ifndef __DPCPP_ENABLE_UNFINISHED_NO_CGH_SUBMIT
   auto params = *static_cast<ur_enqueue_usm_advise_params_t *>(pParams);
   EXPECT_EQ(*params.pphEvent, nullptr);
+#endif
   return UR_RESULT_SUCCESS;
 }
 
diff --git a/sycl/unittests/scheduler/InOrderQueueHostTaskDeps.cpp b/sycl/unittests/scheduler/InOrderQueueHostTaskDeps.cpp
index 31a2914e2c803..7dae968013948 100644
--- a/sycl/unittests/scheduler/InOrderQueueHostTaskDeps.cpp
+++ b/sycl/unittests/scheduler/InOrderQueueHostTaskDeps.cpp
@@ -48,7 +48,7 @@ TEST_F(SchedulerTest, InOrderQueueHostTaskDeps) {
   EXPECT_EQ(GEventsWaitCounter, expectedCount);
 }
 
-enum class CommandType { KERNEL = 1, MEMSET = 2 };
+enum class CommandType { KERNEL = 1, MEMSET = 2, HOST_TASK = 3 };
 std::vector<std::pair<CommandType, size_t>> ExecutedCommands;
 
 inline ur_result_t customEnqueueKernelLaunch(void *pParams) {
@@ -162,3 +162,44 @@ TEST_F(SchedulerTest, InOrderQueueCrossDepsShortcutFuncs) {
   EXPECT_EQ(ExecutedCommands[1].first /*CommandType*/, CommandType::KERNEL);
   EXPECT_EQ(ExecutedCommands[1].second /*EventsCount*/, 0u);
 }
+
+TEST_F(SchedulerTest, InOrderQueueCrossDepsShortcutFuncsParallelFor) {
+  ExecutedCommands.clear();
+  sycl::unittest::UrMock<> Mock;
+  mock::getCallbacks().set_before_callback("urEnqueueKernelLaunch",
+                                           &customEnqueueKernelLaunch);
+
+  sycl::platform Plt = sycl::platform();
+
+  context Ctx{Plt};
+  queue InOrderQueue{Ctx, default_selector_v, property::queue::in_order()};
+
+  std::mutex CvMutex;
+  std::condition_variable Cv;
+  bool ready = false;
+
+  InOrderQueue.submit([&](sycl::handler &CGH) {
+    CGH.host_task([&] {
+      std::unique_lock<std::mutex> lk(CvMutex);
+      Cv.wait(lk, [&ready] { return ready; });
+      ExecutedCommands.push_back({CommandType::HOST_TASK, 0});
+    });
+  });
+
+  event Ev2 = InOrderQueue.parallel_for<TestKernel>(
+      nd_range<1>{range{32}, range{32}}, [](nd_item<1>) {});
+
+  {
+    std::unique_lock<std::mutex> lk(CvMutex);
+    ready = true;
+  }
+  Cv.notify_one();
+
+  InOrderQueue.wait();
+
+  ASSERT_EQ(ExecutedCommands.size(), 2u);
+  EXPECT_EQ(ExecutedCommands[0].first /*CommandType*/, CommandType::HOST_TASK);
+  EXPECT_EQ(ExecutedCommands[0].second /*EventsCount*/, 0u);
+  EXPECT_EQ(ExecutedCommands[1].first /*CommandType*/, CommandType::KERNEL);
+  EXPECT_EQ(ExecutedCommands[1].second /*EventsCount*/, 0u);
+}
diff --git a/sycl/unittests/xpti_trace/CMakeLists.txt b/sycl/unittests/xpti_trace/CMakeLists.txt
index 44574d36a9aa4..0fbb3be4046fc 100644
--- a/sycl/unittests/xpti_trace/CMakeLists.txt
+++ b/sycl/unittests/xpti_trace/CMakeLists.txt
@@ -8,3 +8,4 @@ add_sycl_unittest(XptiTraceTests OBJECT
 )
 target_link_libraries(XptiTraceTests_non_preview PRIVATE xpti xptitest_subscriber)
 target_link_libraries(XptiTraceTests_preview PRIVATE xpti xptitest_subscriber)
+target_link_libraries(XptiTraceTests_no_cgh PRIVATE xpti xptitest_subscriber)