Clean up nvtx macros (#15038)

This PR includes several cleanups for the cudf nvtx wrappers:

- Removed the unused `NVTX3_FUNC_RANGE` macro
- Fixed a typo in the doc
- Added an example in the `cudf::thread_range` doc
- Updated the `NVTX` section in the developer guide doc

Authors:
  - Yunsong Wang (https://github.com/PointKernel)

Approvers:
  - Karthikeyan (https://github.com/karthikeyann)
  - Nghia Truong (https://github.com/ttnghia)

URL: #15038

cpp/doxygen/developer_guide/DEVELOPER_GUIDE.md

@@ -659,10 +659,14 @@ defaults.
 ## NVTX Ranges
 
 In order to aid in performance optimization and debugging, all compute intensive libcudf functions
-should have a corresponding NVTX range. libcudf has a convenience macro `CUDF_FUNC_RANGE()` that
-automatically annotates the lifetime of the enclosing function and uses the function's name as
-the name of the NVTX range. For more information about NVTX, see
-[here](https://github.com/NVIDIA/NVTX/tree/dev/c).
+should have a corresponding NVTX range. Choose between `CUDF_FUNC_RANGE` or `cudf::thread_range`
+for declaring NVTX ranges in the current scope:
+- Use the `CUDF_FUNC_RANGE()` macro if you want to use the name of the function as the name of the
+NVTX range
+- Use `cudf::thread_range rng{"custom_name"};` to provide a custom name for the current scope's
+NVTX range
+
+For more information about NVTX, see [here](https://github.com/NVIDIA/NVTX/tree/dev/c).
 
 ## Input/Output Style
 

cpp/include/cudf/detail/nvtx/nvtx3.hpp

@@ -1,5 +1,5 @@
 /*
- *  Copyright (c) 2020-2023, NVIDIA CORPORATION.
+ *  Copyright (c) 2020-2024, NVIDIA CORPORATION.
  *
  *  Licensed under the Apache License, Version 2.0 (the "License");
  *  you may not use this file except in compliance with the License.
@@ -1901,33 +1901,9 @@ inline void mark(event_attributes const& attr) noexcept
  *
  * @param[in] D Type containing `name` member used to identify the
  * `domain` to which the `registered_message` belongs. Else,
- * `domain::global` to  indicate that the global NVTX domain should be used.
+ * `domain::global` to indicate that the global NVTX domain should be used.
  */
 #define NVTX3_FUNC_RANGE_IN(D)                                                 \
   static ::nvtx3::registered_message<D> const nvtx3_func_name__{__func__};     \
   static ::nvtx3::event_attributes const nvtx3_func_attr__{nvtx3_func_name__}; \
   [[maybe_unused]] ::nvtx3::domain_thread_range<D> const nvtx3_range__{nvtx3_func_attr__};
-
-/**
- * @brief Convenience macro for generating a range in the global domain from the
- * lifetime of a function.
- *
- * This macro is useful for generating an NVTX range in the global domain from
- * the entry point of a function to its exit. It is intended to be the first
- * line of the function.
- *
- * Constructs a static `registered_message` using the name of the immediately
- * enclosing function returned by `__func__` and constructs a
- * `nvtx3::thread_range` using the registered function name as the range's
- * message.
- *
- * Example:
- * ```
- * void foo(...){
- *    NVTX3_FUNC_RANGE(); // Range begins on entry to foo()
- *    // do stuff
- *    ...
- * } // Range ends on return from foo()
- * ```
- */
-#define NVTX3_FUNC_RANGE() NVTX3_FUNC_RANGE_IN(::nvtx3::domain::global)

cpp/include/cudf/detail/nvtx/ranges.hpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2020, NVIDIA CORPORATION.
+ * Copyright (c) 2020-2024, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -28,6 +28,16 @@ struct libcudf_domain {
 
 /**
  * @brief Alias for an NVTX range in the libcudf domain.
+ *
+ * Customizes an NVTX range with the given input.
+ *
+ * Example:
+ * ```
+ * void some_function(){
+ *    cudf::thread_range rng{"custom_name"}; // Customizes range name
+ *    ...
+ * }
+ * ```
  */
 using thread_range = ::nvtx3::domain_thread_range<libcudf_domain>;