Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add gbenchmark for nvtext replace-tokens function #7708

Merged
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions cpp/benchmarks/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -177,8 +177,9 @@ ConfigureBench(BINARYOP_BENCH binaryop/binaryop_benchmark.cu)
ConfigureBench(TEXT_BENCH
text/normalize_benchmark.cpp
text/normalize_spaces_benchmark.cpp
text/tokenize_benchmark.cpp
text/subword_benchmark.cpp)
text/replace_benchmark.cpp
text/subword_benchmark.cpp
text/tokenize_benchmark.cpp)

###################################################################################################
# - strings benchmark -------------------------------------------------------------------
Expand Down
85 changes: 85 additions & 0 deletions cpp/benchmarks/text/replace_benchmark.cpp
Original file line number Diff line number Diff line change
@@ -0,0 +1,85 @@
/*
* Copyright (c) 2021, 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 <benchmark/benchmark.h>
#include <benchmarks/fixture/benchmark_fixture.hpp>
#include <benchmarks/string/string_bench_args.hpp>
#include <benchmarks/synchronization/synchronization.hpp>

#include <cudf/strings/strings_column_view.hpp>
#include <cudf_test/base_fixture.hpp>
#include <cudf_test/column_wrapper.hpp>

#include <nvtext/replace.hpp>

class TextReplace : public cudf::benchmark {
};

static void BM_replace(benchmark::State& state)
{
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::vector<std::string> words{" ", "one ", "two ", "three ", "four ",
"five ", "six ", "sevén ", "eight ", "nine ",
"ten ", "eleven ", "twelve ", "thirteen ", "fourteen ",
"fifteen ", "sixteen ", "seventeen ", "eighteen ", "nineteen "};

std::default_random_engine generator;
std::uniform_int_distribution<int> tokens_dist(0, words.size() - 1);
std::string row; // build a row of random tokens
while (static_cast<int>(row.size()) < n_length) row += words[tokens_dist(generator)];

std::uniform_int_distribution<int> position_dist(0, 16);
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm curious, why 16? Is this a good size test? Do we need to benchmark if each string is of size around 1000?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just an arbitrary power of two number less than 32 to force some amount of warp divergence.
This benchmark tests with string lengths ranging from 32 to 8K and also some row sizes between 4K to 16M (within limits of column size boundaries).


auto elements = cudf::detail::make_counting_transform_iterator(
0, [&](auto idx) { return row.c_str() + position_dist(generator); });
cudf::test::strings_column_wrapper input(elements, elements + n_rows);
cudf::strings_column_view view(input);

cudf::test::strings_column_wrapper targets({"one", "two", "sevén", "zero"});
cudf::test::strings_column_wrapper replacements({"1", "2", "7", "0"});

for (auto _ : state) {
cuda_event_timer raii(state, true, 0);
nvtext::replace_tokens(
view, cudf::strings_column_view(targets), cudf::strings_column_view(replacements));
}

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_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 NVTEXT_BENCHMARK_DEFINE(name) \
BENCHMARK_DEFINE_F(TextReplace, name) \
(::benchmark::State & st) { BM_replace(st); } \
BENCHMARK_REGISTER_F(TextReplace, name) \
->Apply(generate_bench_args) \
->UseManualTime() \
->Unit(benchmark::kMillisecond);

NVTEXT_BENCHMARK_DEFINE(replace)