Rework cudf::strings::detail::copy_range for offsetalator (#15010)

This reworks the `cudf::strings::detail::copy_range()` function to use the offsetalator instead of accessing the output offsets directly. Also refactored the code to remove the unnecessary template arguments. And added a benchmark to ensure these changes did not cause a performance impact.

Most of the code in `cpp/include/cudf/strings/detail/copy_range.cuh` was rewritten and moved to `cpp/src/strings/copying/copy_range.cu`.

Authors:
  - David Wendt (https://github.com/davidwendt)

Approvers:
  - Robert Maynard (https://github.com/robertmaynard)
  - Vukasin Milovanovic (https://github.com/vuule)

URL: #15010

cpp/CMakeLists.txt

@@ -570,6 +570,7 @@ add_library(
   src/strings/convert/convert_lists.cu
   src/strings/copying/concatenate.cu
   src/strings/copying/copying.cu
+  src/strings/copying/copy_range.cu
   src/strings/copying/shift.cu
   src/strings/count_matches.cu
   src/strings/extract/extract.cu

cpp/benchmarks/CMakeLists.txt

@@ -312,6 +312,7 @@ ConfigureNVBench(
   string/char_types.cpp
   string/contains.cpp
   string/copy_if_else.cpp
+  string/copy_range.cpp
   string/count.cpp
   string/extract.cpp
   string/gather.cpp

cpp/benchmarks/string/copy_range.cpp

@@ -0,0 +1,60 @@
+/*
+ * Copyright (c) 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.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <benchmarks/common/generate_input.hpp>
+
+#include <cudf/copying.hpp>
+#include <cudf/strings/strings_column_view.hpp>
+#include <cudf/utilities/default_stream.hpp>
+
+#include <nvbench/nvbench.cuh>
+
+static void bench_copy_range(nvbench::state& state)
+{
+  auto const num_rows  = static_cast<cudf::size_type>(state.get_int64("num_rows"));
+  auto const row_width = static_cast<cudf::size_type>(state.get_int64("row_width"));
+
+  if (static_cast<std::size_t>(num_rows) * static_cast<std::size_t>(row_width) >=
+      static_cast<std::size_t>(std::numeric_limits<cudf::size_type>::max())) {
+    state.skip("Skip benchmarks greater than size_type limit");
+  }
+
+  data_profile const table_profile =
+    data_profile_builder()
+      .distribution(cudf::type_id::STRING, distribution_id::NORMAL, 0, row_width)
+      .no_validity();
+  auto const source_tables = create_random_table(
+    {cudf::type_id::STRING, cudf::type_id::STRING}, row_count{num_rows}, table_profile);
+
+  auto const start  = num_rows / 4;
+  auto const end    = (num_rows * 3) / 4;
+  auto const source = source_tables->view().column(0);
+  auto const target = source_tables->view().column(1);
+
+  state.set_cuda_stream(nvbench::make_cuda_stream_view(cudf::get_default_stream().value()));
+  auto chars_size = cudf::strings_column_view(target).chars_size(cudf::get_default_stream());
+  state.add_global_memory_reads<nvbench::int8_t>(chars_size);   // all bytes are read;
+  state.add_global_memory_writes<nvbench::int8_t>(chars_size);  // both columns are similar size
+
+  state.exec(nvbench::exec_tag::sync, [&](nvbench::launch& launch) {
+    [[maybe_unused]] auto result = cudf::copy_range(source, target, start, end, start / 2);
+  });
+}
+
+NVBENCH_BENCH(bench_copy_range)
+  .set_name("copy_range")
+  .add_int64_axis("row_width", {32, 64, 128, 256, 512, 1024, 2048, 4096})
+  .add_int64_axis("num_rows", {4096, 32768, 262144, 2097152, 16777216});

cpp/include/cudf/strings/detail/copy_range.hpp

@@ -0,0 +1,60 @@
+/*
+ * Copyright (c) 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.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+#pragma once
+
+#include <cudf/strings/strings_column_view.hpp>
+#include <cudf/types.hpp>
+
+#include <rmm/cuda_stream_view.hpp>
+
+namespace cudf {
+namespace strings {
+namespace detail {
+
+/**
+ * @brief Internal API to copy a range of string elements out-of-place from
+ * a source column to a target column
+ *
+ * Creates a new column as if an in-place copy was performed into `target`.
+ * The elements indicated by the indices `source_begin`, `source_end`)
+ * replace with the elements in the target column starting at `target_begin`.
+ * Elements outside the range are copied from `target` into the new target
+ * column to return.
+ *
+ * @throws cudf::logic_error for invalid range (if `target_begin < 0`,
+ * or `target_begin >= target.size()`,
+ * or `target_begin + (source_end-source_begin)` > target.size()`).
+ *
+ * @param source The strings column to copy from inside the `target_begin` range
+ * @param target The strings column to copy from outside the range
+ * @param source_end The index of the first element in the source range
+ * @param source_end The index of the last element in the source range (exclusive)
+ * @param target_begin The starting index of the target range (inclusive)
+ * @param stream CUDA stream used for device memory operations and kernel launches.
+ * @param mr Device memory resource used to allocate the returned column's device memory.
+ * @return The result target column
+ */
+std::unique_ptr<column> copy_range(strings_column_view const& source,
+                                   strings_column_view const& target,
+                                   size_type source_begin,
+                                   size_type source_end,
+                                   size_type target_begin,
+                                   rmm::cuda_stream_view stream,
+                                   rmm::mr::device_memory_resource* mr);
+
+}  // namespace detail
+}  // namespace strings
+}  // namespace cudf