[SYCL][NFC] Fix some spelling errors in SYCL project (#20930)

bader · web-flow · commit 31d8329f8f90 · 2025-12-18T08:41:23.000-08:00
The errors are fixed in the documentation and the code comments.
diff --git a/sycl/doc/FAQ.md b/sycl/doc/FAQ.md
@@ -60,7 +60,7 @@ design/CompilerAndRuntimeDesign.md) document.
 ### Q: What happens if I run my SYCL application on a machine without a supported backend?
 **A:** If a SYCL application is unable to find a suitable device, the SYCL
 objects (`sycl::context`, `sycl::queue`, etc.) or the SYCL device selector (See
-[SYCL 2020 specificaion 4.6.1.1](https://registry.khronos.org/SYCL/specs/sycl-2020/html/sycl-2020.html#sec:device-selection))
+[SYCL 2020 specification 4.6.1.1](https://registry.khronos.org/SYCL/specs/sycl-2020/html/sycl-2020.html#sec:device-selection))
 will throw a `sycl::exception` with `errc::runtime` error code.
 
 
diff --git a/sycl/doc/design/CompileTimeProperties.md b/sycl/doc/design/CompileTimeProperties.md
@@ -611,7 +611,7 @@ names and values of the properties in the following sequence:
 * Etc.
 
 Every field in the global variable pointed to by this parameter are string
-literals in seperate global variables. Property values are converted to strings
+literals in separate global variables. Property values are converted to strings
 in the same way as described above, except that the `nullptr` value and the
 empty string (`""`) is represented as `null` in the global variable field.
 
diff --git a/sycl/doc/design/GlobalObjectsInRuntime.md b/sycl/doc/design/GlobalObjectsInRuntime.md
@@ -5,7 +5,7 @@
 C++ standard does not specify the order in which global objects are constructed
 or destroyed. If global objects somehow interact with each other, there's a
 chance, that one of the objects has not been initialized or has been destroyed
-by the time of interaction. This problem is also refered to as 
+by the time of interaction. This problem is also referred to as
 [static initialization order fiasco].
 
 The only two things C++ guarantees is that global objects are constructed before
diff --git a/sycl/doc/design/KernelParameterPassing.md b/sycl/doc/design/KernelParameterPassing.md
@@ -136,7 +136,7 @@ other host fields replaced with device-specific fields and
 the host data pointer field must be translated to an OpenCL
 or L0 memory object before it can be passed as a kernel parameter.
 To enable all of this, the parameters of the `sycl::accessor`
-and `sycl::sampler` init functions are transfered from 
+and `sycl::sampler` init functions are transferred from
 host to device separately. The values received on the device
 are passed to the `init` functions executed on the device,
 which results in the reassembly of the SYCL object in a form usable on the device.
diff --git a/sycl/doc/design/OffloadDesign.md b/sycl/doc/design/OffloadDesign.md
@@ -78,7 +78,7 @@ Example usage of the external `llvm-offload-binary` call:
 The `clang-linker-wrapper` provides the interface to perform the needed link
 steps when consuming fat binaries.  The linker wrapper performs a majority of
 the work involved during the link step during an offload compilation,
-significantly reducing the amount of work that is occuring in the compiler
+significantly reducing the amount of work that is occurring in the compiler
 driver.  From the compilation perspective, the linker wrapper replaces the
 typical call to the host link.  This allows for the responsibility of the
 compiler driver to be nearly identical when performing a regular compilation
diff --git a/sycl/doc/syclgraph/SYCLGraphUsageGuide.md b/sycl/doc/syclgraph/SYCLGraphUsageGuide.md
@@ -576,7 +576,7 @@ sycl::buffer bufferB{...};
 sycl_ext::dynamic_parameter dynParamAccessor(myGraph, bufferA.get_access());
 
 sycl_ext::node kernelNode = myGraph.add([&](handler& cgh) {
-    // Require the accessor contained in the dynamic paramter
+    // Require the accessor contained in the dynamic parameter
     cgh.require(dynParamAccessor);
     // Set the arg on the kernel using the dynamic parameter directly
     cgh.set_args(dynParamAccessor);
diff --git a/sycl/include/sycl/detail/accessor_iterator.hpp b/sycl/include/sycl/detail/accessor_iterator.hpp
@@ -165,7 +165,7 @@ template <typename DataT, int Dimensions> class accessor_iterator {
   // Fields below are used (and changed to be non-zero) only if we deal with
   // a ranged accessor.
   //
-  // TODO: consider making their existance dependable on Dimensions template
+  // TODO: consider making their existence dependable on Dimensions template
   // parameter, because not all of them are needed for all possible dimensions.
 
   // Three field below allow us to calculate an absolute offset to an accessor's
diff --git a/sycl/include/sycl/ext/intel/esimd/detail/simd_obj_impl.hpp b/sycl/include/sycl/ext/intel/esimd/detail/simd_obj_impl.hpp
@@ -1022,7 +1022,7 @@ class [[__sycl_detail__::__uses_aspects__(
   /// Unary logical negation operator, available in all subclasses, but only for
   /// integral element types (\c simd_mask included).
   /// Similarly to C++, where !x returns bool, !simd returns a simd_mask, where
-  /// each element is a result of comparision with zero.
+  /// each element is a result of comparison with zero.
   /// @return A \c simd_mask instance where each element is a result of
   ///   comparison of the original element with zero.
   template <class T1 = Ty, class = std::enable_if_t<std::is_integral_v<T1>>>
diff --git a/sycl/include/sycl/group.hpp b/sycl/include/sycl/group.hpp
@@ -39,7 +39,7 @@ inline namespace _V1 {
 template <int Dimensions> class h_item;
 
 namespace detail {
-// Implements a barrier accross work items within a work group.
+// Implements a barrier across work items within a work group.
 inline void workGroupBarrier() {
 #ifdef __SYCL_DEVICE_ONLY__
   constexpr uint32_t flags =
diff --git a/sycl/include/sycl/interop_handle.hpp b/sycl/include/sycl/interop_handle.hpp
@@ -109,7 +109,7 @@ class interop_handle {
 #endif
   }
 
-  /// Returns an underlying native backend object associated with teh queue
+  /// Returns an underlying native backend object associated with the queue
   /// that the host task was submitted to. If the command group was submitted
   /// with a secondary queue and the fall-back was triggered, the queue that
   /// is associated with the interop_handle must be the fall-back queue.
diff --git a/sycl/include/sycl/multi_ptr.hpp b/sycl/include/sycl/multi_ptr.hpp
@@ -71,7 +71,7 @@ struct LegacyReferenceTypes<ElementType,
 /// qualified pointers to allow interoperability between plain C++ and OpenCL C.
 ///
 /// \ingroup sycl_api
-// TODO: Default value for DecorateAddress is for backwards compatiblity. It
+// TODO: Default value for DecorateAddress is for backwards compatibility. It
 //       should be removed.
 template <typename ElementType, access::address_space Space,
           access::decorated DecorateAddress = access::decorated::legacy>
diff --git a/sycl/source/detail/device_filter.cpp b/sycl/source/detail/device_filter.cpp
@@ -229,7 +229,7 @@ Parse_ONEAPI_DEVICE_SELECTOR(const std::string &envString) {
 
       backend be = Parse_ODS_Backend(input_be, Entry);
 
-      // For each backend, we can have multiple targets, seperated by ','.
+      // For each backend, we can have multiple targets, separated by ','.
       std::vector<std::string_view> Targets = tokenize(Pair[1], ",");
       for (auto TargetStr : Targets) {
         ods_target DeviceTarget(be);
diff --git a/sycl/source/detail/global_handler.hpp b/sycl/source/detail/global_handler.hpp
@@ -43,7 +43,7 @@ class GlobalHandler {
 public:
   static bool isInstanceAlive() { return RTGlobalObjHandler != nullptr; }
   /// \return a reference to a GlobalHandler singleton instance. The reference
-  /// is valid as long as runtime library is loaded (i.e. untill `DllMain` or
+  /// is valid as long as runtime library is loaded (i.e. until `DllMain` or
   /// `__attribute__((destructor))` is called).
   static GlobalHandler &instance() { return *RTGlobalObjHandler; }
 
diff --git a/sycl/source/detail/kernel_data.cpp b/sycl/source/detail/kernel_data.cpp
@@ -144,7 +144,7 @@ void KernelData::processArg(void *Ptr, const detail::kernel_param_kind_t &Kind,
         static_cast<detail::AccessorBaseHost *>(&S->GlobalFlushBuf);
     detail::Requirement *GFlushReq = &*detail::getSyclObjImpl(*GFlushBase);
 
-    // If work group size wasn't set explicitly then it must be recieved
+    // If work group size wasn't set explicitly then it must be received
     // from kernel attribute or set to default values.
     // For now we can't get this attribute here.
     // So we just suppose that WG size is always default for stream.
diff --git a/sycl/source/detail/kernel_info.hpp b/sycl/source/detail/kernel_info.hpp
@@ -142,7 +142,7 @@ get_kernel_device_specific_info(ur_kernel_handle_t Kernel,
 }
 
 // TODO: This is used by a deprecated version of
-// info::kernel_device_specific::max_sub_group_size taking an input paramter.
+// info::kernel_device_specific::max_sub_group_size taking an input parameter.
 // This should be removed when the deprecated info query is removed.
 template <typename Param>
 uint32_t get_kernel_device_specific_info_with_input(ur_kernel_handle_t Kernel,
diff --git a/sycl/source/detail/scheduler/scheduler.hpp b/sycl/source/detail/scheduler/scheduler.hpp
@@ -628,7 +628,7 @@ class Scheduler {
                       std::vector<Command *> &ToCleanUp);
 
     /// Perform connection of events in multiple contexts
-    /// \param Cmd dependant command
+    /// \param Cmd dependent command
     /// \param DepEvent event to depend on
     /// \param Dep optional DepDesc to perform connection properly
     /// \param ToCleanUp container for commands that can be cleaned up due to
diff --git a/sycl/test-e2e/Adapters/level_zero/interop-device.cpp b/sycl/test-e2e/Adapters/level_zero/interop-device.cpp
@@ -30,7 +30,7 @@ void for_each_descendent_dev(sycl::device dev, FuncTy F) {
 
   // I'm not sure if remaining partitioning schems are actually supported by any
   // of the existing Level-Zero devices. Make sure we still cover that
-  // possibility in this test to accomodate any future situation.
+  // possibility in this test to accommodate any future situation.
 
   if (supports(sycl::info::partition_property::partition_equally)) {
     std::cout << "Equally" << std::endl;
diff --git a/sycl/test-e2e/ESIMD/Prefix_Local_sum3.cpp b/sycl/test-e2e/ESIMD/Prefix_Local_sum3.cpp
@@ -343,7 +343,7 @@ int main(int argc, char *argv[]) {
   unsigned int *pDeviceOutputs =
       malloc_shared<unsigned int>(size * TUPLE_SZ, q);
 
-  // allocate & intialize expected result
+  // allocate & initialize expected result
   unsigned int *pExpectOutputs = static_cast<unsigned int *>(
       malloc(size * TUPLE_SZ * sizeof(unsigned int)));
   memcpy(pExpectOutputs, pInputs, size * TUPLE_SZ * sizeof(unsigned));
diff --git a/sycl/test-e2e/ESIMD/api/functional/README.md b/sycl/test-e2e/ESIMD/api/functional/README.md
@@ -3,5 +3,5 @@ Tests within this directory has additional preprocessor definitions available.
 
 `ESIMD_TESTS_VERBOSE_LOG` (default: not defined)
 Enables verbose log messages, including the trace messages for the current test
-case running. It could be usefull to investigate test failures, crashes or
+case running. It could be useful to investigate test failures, crashes or
 freezes.
diff --git a/sycl/test-e2e/ESIMD/api/functional/ctors/ctor_vector_core.cpp b/sycl/test-e2e/ESIMD/api/functional/ctors/ctor_vector_core.cpp
@@ -12,7 +12,7 @@
 // This test uses different data types, sizes and different simd constructor
 // invocation contexts.
 // The test do the following actions:
-//  - call init_simd.data() to retreive vector_type and then provide it to the
+//  - call init_simd.data() to retrieve vector_type and then provide it to the
 //    simd constructor
 //  - bitwise comparing expected and retrieved values
 
diff --git a/sycl/test-e2e/ESIMD/api/functional/ctors/ctor_vector_fp_extra.cpp b/sycl/test-e2e/ESIMD/api/functional/ctors/ctor_vector_fp_extra.cpp
@@ -12,7 +12,7 @@
 // This test uses different data types, sizes and different simd constructor
 // invocation contexts.
 // The test do the following actions:
-//  - call init_simd.data() to retreive vector_type and then provide it to the
+//  - call init_simd.data() to retrieve vector_type and then provide it to the
 //    simd constructor
 //  - bitwise comparing expected and retrieved values
 
diff --git a/sycl/test-e2e/ESIMD/api/functional/logger.hpp b/sycl/test-e2e/ESIMD/api/functional/logger.hpp
@@ -166,7 +166,7 @@ template <typename... argsT> inline void debug(const argsT &...args) {
 #endif
 }
 
-// Syntax sugar for tracing test cases; could be usefull in case of test crash
+// Syntax sugar for tracing test cases; could be useful in case of test crash
 // or freeze
 template <typename TestCaseDescriptionT, typename... argsT>
 inline void trace(const argsT &...args) {
diff --git a/sycl/test-e2e/ESIMD/api/functional/sycl_accessor.hpp b/sycl/test-e2e/ESIMD/api/functional/sycl_accessor.hpp
@@ -7,7 +7,7 @@
 //===----------------------------------------------------------------------===//
 ///
 /// \file
-/// Provides type packs and all neccessary helpers and wrappers to use with
+/// Provides type packs and all necessary helpers and wrappers to use with
 /// tests with sycl::accessor
 ///
 //===----------------------------------------------------------------------===//
diff --git a/sycl/test-e2e/ESIMD/api/functional/sycl_range.hpp b/sycl/test-e2e/ESIMD/api/functional/sycl_range.hpp
@@ -7,7 +7,7 @@
 //===----------------------------------------------------------------------===//
 ///
 /// \file
-/// Provides all neccessary helpers and wrappers to use with tests with
+/// Provides all necessary helpers and wrappers to use with tests with
 /// sycl::range
 ///
 //===----------------------------------------------------------------------===//
diff --git a/sycl/test-e2e/ESIMD/api/functional/type_coverage.hpp b/sycl/test-e2e/ESIMD/api/functional/type_coverage.hpp
@@ -287,7 +287,7 @@ template <typename T, T... values> struct value_pack {
 
   // Factory function to generate the type pack using generator function
   //
-  // It could be especially usefull for value packs with enums. For example:
+  // It could be especially useful for value packs with enums. For example:
   //   enum class access {read, write};
   //   template <access ... values>
   //   using access_pack = value_pack<access, values...>;
diff --git a/sycl/test-e2e/Graph/Inputs/release_while_executing.cpp b/sycl/test-e2e/Graph/Inputs/release_while_executing.cpp
@@ -16,7 +16,7 @@ int main() {
 
   T *PtrA = malloc_device<T>(Size, Queue);
 
-  // Create the command_graph in a seperate scope so that it's destroyed before
+  // Create the command_graph in a separate scope so that it's destroyed before
   // Queue.wait()
   {
     exp_ext::command_graph Graph{Queue.get_context(), Queue.get_device()};
diff --git a/sycl/test-e2e/Reduction/reduction_usm.cpp b/sycl/test-e2e/Reduction/reduction_usm.cpp
@@ -45,7 +45,7 @@ int test(queue &Q, OptionalIdentity<T, HasIdentity> Identity, T Init,
 
   // Compute.
   Q.submit([&](handler &CGH) {
-     // Helper for creating the reductions depending on the existance of an
+     // Helper for creating the reductions depending on the existence of an
      // identity.
      auto CreateReduction = [&]() {
        if constexpr (HasIdentity) {
diff --git a/sycl/test-e2e/Reduction/reduction_utils.hpp b/sycl/test-e2e/Reduction/reduction_utils.hpp
@@ -365,7 +365,7 @@ int testInner(queue &Q, OptionalIdentity<T, HasIdentity> Identity, T Init,
 
   // Compute.
   Q.submit([&](handler &CGH) {
-    // Helper for creating the reductions depending on the existance of an
+    // Helper for creating the reductions depending on the existence of an
     // identity.
     auto CreateReduction = [&]() {
       if constexpr (HasIdentity) {
@@ -476,7 +476,7 @@ int testUSMInner(queue &Q, OptionalIdentity<T, HasIdentity> Identity, T Init,
 
   // Compute.
   Q.submit([&](handler &CGH) {
-     // Helper for creating the reductions depending on the existance of an
+     // Helper for creating the reductions depending on the existence of an
      // identity.
      auto CreateReduction = [&]() {
        if constexpr (HasIdentity) {
diff --git a/sycl/test-e2e/VirtualFunctions/helpers.hpp b/sycl/test-e2e/VirtualFunctions/helpers.hpp
@@ -16,7 +16,7 @@ template <typename... T> struct aligned_storage {
 };
 
 // Helper data structure that automatically creates a right (in terms of size
-// and alignment) storage to accomodate a value of any of types T...
+// and alignment) storage to accommodate a value of any of types T...
 template <typename... T> struct obj_storage_t {
   static_assert(std::max({alignof(T)...}) == std::min({alignof(T)...}),
                 "Unsupported alignment of input types");
diff --git a/sycl/test-e2e/bindless_images/array/read_sampled_array.cpp b/sycl/test-e2e/bindless_images/array/read_sampled_array.cpp
@@ -66,7 +66,7 @@ runNDimTestHost(sycl::range<NDims> globalSize, float offset,
           coords[1] = j;
         }
 
-        // Sampling is done only within array layers, not accross layers.
+        // Sampling is done only within array layers, not across layers.
         // To mimic this on host, extract the layer and sample that.
         std::vector<sycl::vec<DType, NChannels>> layer(
             inputImage.begin() + arr_idx * globalSizeTwoComp.size(),
diff --git a/sycl/test/basic_tests/reduction/known_identities.cpp b/sycl/test/basic_tests/reduction/known_identities.cpp
@@ -1,6 +1,6 @@
 // RUN: %clangxx -fsycl -fsyntax-only %s
 
-// Tests the existance and specializations of known identities.
+// Tests the existence and specializations of known identities.
 
 #include <sycl/sycl.hpp>
 
diff --git a/sycl/test/esimd/simd_view.cpp b/sycl/test/esimd/simd_view.cpp
@@ -149,7 +149,7 @@ SYCL_ESIMD_FUNCTION void test_simd_view_move_assign() {
   v0_view = std::move(v1_view);
 }
 
-// Check that the same holds for specialization with lenght==1.
+// Check that the same holds for specialization with length==1.
 SYCL_ESIMD_FUNCTION void test_simd_view_ctors_length1() {
   simd<int, 16> v = 1;
   auto v0_view = v[0];
diff --git a/sycl/test/native_cpu/example-sycl-application.cpp b/sycl/test/native_cpu/example-sycl-application.cpp
@@ -83,7 +83,7 @@ int main() {
     /* This kernel will actually perform the computation C = A * B. Since
      * A and B are only read from, we specify read accessors for those two
      * buffers, which the SYCL runtime recognises as a dependency on the
-     * previous kernels. If the data were initialised on a different device,
+     * previous kernels. If the data were initialized on a different device,
      * or on the host, the SYCL runtime would ensure that the data were
      * copied between contexts etc. properly. */
     myQueue.submit([&](handler &cgh) {
diff --git a/sycl/unittests/buffer/BufferReleaseBase.cpp b/sycl/unittests/buffer/BufferReleaseBase.cpp
@@ -256,7 +256,7 @@ TEST_F(BufferDestructionCheck, ReadyToReleaseLogic) {
   ExpectedEventStatus[WriteCmd->getEvent()->getHandle()] =
       UR_EVENT_STATUS_COMPLETE;
   EXPECT_TRUE(MockSchedulerPtr->checkLeavesCompletion(Rec));
-  // previous expect_call is still valid and will generate failure if we recieve
+  // previous expect_call is still valid and will generate failure if we receive
   // call here, no need for extra limitation
   EXPECT_CALL(*ReadCmd, Release).Times(1);
   EXPECT_CALL(*WriteCmd, Release).Times(1);
diff --git a/sycl/unittests/helpers/UrMock.hpp b/sycl/unittests/helpers/UrMock.hpp
@@ -617,7 +617,7 @@ template <sycl::backend Backend = backend::opencl> class UrMock {
     // these between tests
     detail::GlobalHandler::instance().prepareSchedulerToRelease(true);
     detail::GlobalHandler::instance().releaseDefaultContexts();
-    // the platform cache is cleared at the BEGINING of the mock.
+    // the platform cache is cleared at the BEGINNING of the mock.
 
     mock::getCallbacks().resetCallbacks();
   }
diff --git a/sycl/unittests/queue/GetProfilingInfo.cpp b/sycl/unittests/queue/GetProfilingInfo.cpp
@@ -32,7 +32,7 @@ static ur_result_t redefinedUrEventGetProfilingInfo(void *) {
 
 static ur_result_t redefinedUrDeviceGet(void *pParams) {
   auto params = *static_cast<ur_device_get_params_t *>(pParams);
-  // Host/Device timer syncronization isn't done all the time (cached), so we
+  // Host/Device timer synchronization isn't done all the time (cached), so we
   // need brand new device for some of the testcases.
   static std::intptr_t device_id = 10;
   if (*params.ppNumDevices)

Original file line number	Diff line number	Diff line change
`@@ -109,7 +109,7 @@ class interop_handle {`
`109`	`109`	`#endif`
`110`	`110`	`}`
`111`	`111`
`112`		`- /// Returns an underlying native backend object associated with teh queue`
	`112`	`+ /// Returns an underlying native backend object associated with the queue`
`113`	`113`	`/// that the host task was submitted to. If the command group was submitted`
`114`	`114`	`/// with a secondary queue and the fall-back was triggered, the queue that`
`115`	`115`	`/// is associated with the interop_handle must be the fall-back queue.`