update

apache · Jun 5, 2024 · 2605d84 · 2605d84
1 parent ecebb29
commit 2605d84
Show file tree

Hide file tree

Showing 5 changed files with 18 additions and 11 deletions.
diff --git a/cpp/velox/benchmarks/GenericBenchmark.cc b/cpp/velox/benchmarks/GenericBenchmark.cc
@@ -169,11 +169,12 @@ void populateWriterMetrics(
 
 void setCpu(::benchmark::State& state) {
   // Pin each threads to different CPU# starting from 0 or --cpu.
+  auto cpu = state.thread_index();
   if (FLAGS_cpu != -1) {
-    gluten::setCpu(FLAGS_cpu + state.thread_index());
-  } else {
-    gluten::setCpu(state.thread_index());
+    cpu += FLAGS_cpu;
   }
+  LOG(INFO) << "Setting CPU for thread " << state.thread_index() << " to " << cpu;
+  gluten::setCpu(cpu);
 }
 
 void runShuffle(
@@ -361,8 +362,8 @@ int main(int argc, char** argv) {
   gflags::ParseCommandLineFlags(&argc, &argv, true);
 
   // Init Velox backend.
-  std::unordered_map<std::string, std::string> backendConf;
-  std::unordered_map<std::string, std::string> sessionConf;
+  auto backendConf = gluten::defaultConf();
+  auto sessionConf = gluten::defaultConf();
   backendConf.insert({gluten::kSparkBatchSize, std::to_string(FLAGS_batch_size)});
   if (!FLAGS_conf.empty()) {
     abortIfFileNotExists(FLAGS_conf);

diff --git a/cpp/velox/benchmarks/common/BenchmarkUtils.cc b/cpp/velox/benchmarks/common/BenchmarkUtils.cc
@@ -33,10 +33,15 @@ DEFINE_int32(iterations, 1, "The number of iterations to run this benchmark");
 
 namespace gluten {
 namespace {
+std::unordered_map<std::string, std::string> bmConfMap = defaultConf();
+}
 
-std::unordered_map<std::string, std::string> bmConfMap = {{gluten::kSparkBatchSize, std::to_string(FLAGS_batch_size)}};
-
-} // namespace
+std::unordered_map<std::string, std::string> defaultConf() {
+  return {
+      {gluten::kSparkBatchSize, std::to_string(FLAGS_batch_size)},
+      {gluten::kDebugModeEnabled, "true"},
+  };
+}
 
 void initVeloxBackend(std::unordered_map<std::string, std::string>& conf) {
   gluten::VeloxBackend::create(conf);

diff --git a/cpp/velox/benchmarks/common/BenchmarkUtils.h b/cpp/velox/benchmarks/common/BenchmarkUtils.h
@@ -43,6 +43,9 @@ DECLARE_int32(threads);
 DECLARE_int32(iterations);
 
 namespace gluten {
+
+std::unordered_map<std::string, std::string> defaultConf();
+
 /// Initialize the Velox backend with default value.
 void initVeloxBackend();
 

diff --git a/cpp/velox/benchmarks/common/FileReaderIterator.h b/cpp/velox/benchmarks/common/FileReaderIterator.h
@@ -38,8 +38,6 @@ class FileReaderIterator : public ColumnarBatchIterator {
 
   virtual ~FileReaderIterator() = default;
 
-  virtual void createReader() = 0;
-
   virtual std::shared_ptr<arrow::Schema> getSchema() = 0;
 
   int64_t getCollectBatchTime() const {

diff --git a/cpp/velox/benchmarks/common/ParquetReaderIterator.h b/cpp/velox/benchmarks/common/ParquetReaderIterator.h
@@ -27,7 +27,7 @@ class ParquetReaderIterator : public FileReaderIterator {
  public:
   explicit ParquetReaderIterator(const std::string& path) : FileReaderIterator(path) {}
 
-  void createReader() override {
+  void createReader() {
     parquet::ArrowReaderProperties properties = parquet::default_arrow_reader_properties();
     properties.set_batch_size(FLAGS_batch_size);
     GLUTEN_THROW_NOT_OK(parquet::arrow::FileReader::Make(