From b7c1c9e72643aff82cec4cc37443d728c892e3bc Mon Sep 17 00:00:00 2001
From: David Wendt <45795991+davidwendt@users.noreply.github.com>
Date: Thu, 8 Apr 2021 23:24:29 -0400
Subject: [PATCH] Update strings extract gbenchmark to measure multiple groups
 (#7789)

Reference https://github.com/rapidsai/custreamz/issues/2
Updating the gbenchmark for `cudf::strings::extract` to measure varying number of groups. Previously the regex pattern only had 1 group. This is to more closely align to the referenced issue which includes more than 20 groups.

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

Approvers:
  - Nghia Truong (https://github.com/ttnghia)
  - Devavret Makkar (https://github.com/devavret)
  - Conor Hoekstra (https://github.com/codereport)

URL: https://github.com/rapidsai/cudf/pull/7789
---
 cpp/benchmarks/string/extract_benchmark.cpp | 68 +++++++++++++--------
 1 file changed, 41 insertions(+), 27 deletions(-)

diff --git a/cpp/benchmarks/string/extract_benchmark.cpp b/cpp/benchmarks/string/extract_benchmark.cpp
index f53ea60cdf9..aa1e59a22bf 100644
--- a/cpp/benchmarks/string/extract_benchmark.cpp
+++ b/cpp/benchmarks/string/extract_benchmark.cpp
@@ -14,6 +14,8 @@
  * limitations under the License.
  */
 
+#include "string_bench_args.hpp"
+
 #include <benchmark/benchmark.h>
 #include <benchmarks/common/generate_benchmark_input.hpp>
 #include <benchmarks/fixture/benchmark_fixture.hpp>
@@ -23,43 +25,55 @@
 #include <cudf/strings/strings_column_view.hpp>
 #include <cudf_test/column_wrapper.hpp>
 
-#include "string_bench_args.hpp"
+#include <random>
 
 class StringExtract : public cudf::benchmark {
 };
 
-static void BM_extract(benchmark::State& state, int re_instructions)
+static void BM_extract(benchmark::State& state, int groups)
 {
-  cudf::size_type const n_rows{static_cast<cudf::size_type>(state.range(0))};
-  cudf::size_type const max_str_length{static_cast<cudf::size_type>(state.range(1))};
-  data_profile table_profile;
-  table_profile.set_distribution_params(
-    cudf::type_id::STRING, distribution_id::NORMAL, 0, max_str_length);
-  auto const table =
-    create_random_table({cudf::type_id::STRING}, 1, row_count{n_rows}, table_profile);
-  cudf::strings_column_view input(table->view().column(0));
-  std::string const raw_pattern =
-    "1234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234"
-    "5678901234567890123456789012345678901234567890";
-  std::string const pattern = "(" + raw_pattern.substr(0, re_instructions) + ")";
+  auto const n_rows   = static_cast<cudf::size_type>(state.range(0));
+  auto const n_length = static_cast<cudf::size_type>(state.range(1));
+
+  std::default_random_engine generator;
+  std::uniform_int_distribution<int> words_dist(0, 999);
+
+  std::vector<std::string> samples(100);  // 100 unique rows of data to reuse
+  std::generate(samples.begin(), samples.end(), [&]() {
+    std::string row;  // build a row of random tokens
+    while (static_cast<int>(row.size()) < n_length) {
+      row += std::to_string(words_dist(generator)) + " ";
+    }
+    return row;
+  });
+
+  std::string pattern;
+  while (static_cast<int>(pattern.size()) < groups) { pattern += "(\\d+) "; }
+
+  std::uniform_int_distribution<int> distribution(0, samples.size() - 1);
+  auto elements = cudf::detail::make_counting_transform_iterator(
+    0, [&](auto idx) { return samples.at(distribution(generator)); });
+  cudf::test::strings_column_wrapper input(elements, elements + n_rows);
+  cudf::strings_column_view view(input);
 
   for (auto _ : state) {
-    cuda_event_timer raii(state, true, rmm::cuda_stream_default);
-    auto results = cudf::strings::extract(input, pattern);
+    cuda_event_timer raii(state, true);
+    auto results = cudf::strings::extract(view, pattern);
   }
 
-  state.SetBytesProcessed(state.iterations() * input.chars_size());
+  state.SetBytesProcessed(state.iterations() * view.chars_size());
 }
 
 static void generate_bench_args(benchmark::internal::Benchmark* b)
 {
-  int const min_rows   = 1 << 12;
-  int const max_rows   = 1 << 24;
-  int const row_mult   = 8;
-  int const min_rowlen = 1 << 5;
-  int const max_rowlen = 1 << 13;
-  int const len_mult   = 4;
-  generate_string_bench_args(b, min_rows, max_rows, row_mult, min_rowlen, max_rowlen, len_mult);
+  int const min_rows          = 1 << 12;
+  int const max_rows          = 1 << 24;
+  int const row_multiplier    = 8;
+  int const min_row_length    = 1 << 5;
+  int const max_row_length    = 1 << 13;
+  int const length_multiplier = 4;
+  generate_string_bench_args(
+    b, min_rows, max_rows, row_multiplier, min_row_length, max_row_length, length_multiplier);
 }
 
 #define STRINGS_BENCHMARK_DEFINE(name, instructions)          \
@@ -70,6 +84,6 @@ static void generate_bench_args(benchmark::internal::Benchmark* b)
     ->UseManualTime()                                         \
     ->Unit(benchmark::kMillisecond);
 
-STRINGS_BENCHMARK_DEFINE(small, 4)
-STRINGS_BENCHMARK_DEFINE(medium, 48)
-STRINGS_BENCHMARK_DEFINE(large, 128)
+STRINGS_BENCHMARK_DEFINE(small, 2)
+STRINGS_BENCHMARK_DEFINE(medium, 10)
+STRINGS_BENCHMARK_DEFINE(large, 30)