Add Cast and CoinFlip GPU benchmarks (NVIDIA#3541)

Signed-off-by: Szymon Karpiński <hugo@staszic.waw.pl>

Author: szkarpinski

dali/benchmark/CMakeLists.txt

@@ -38,6 +38,8 @@ if (BUILD_BENCHMARK)
     "${CMAKE_CURRENT_SOURCE_DIR}/copy_bench.cc"
     "${CMAKE_CURRENT_SOURCE_DIR}/gaussian_blur_bench.cc"
     "${CMAKE_CURRENT_SOURCE_DIR}/flip_bench.cc"
+    "${CMAKE_CURRENT_SOURCE_DIR}/cast_bench.cc"
+    "${CMAKE_CURRENT_SOURCE_DIR}/coin_flip_bench.cc"
   )
 
   if (BUILD_LMDB)

dali/benchmark/cast_bench.cc

@@ -0,0 +1,43 @@
+// Copyright (c) 2021, Szymon Karpiński. All rights reserved.
+//
+// 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 "dali/benchmark/operator_bench.h"
+#include "dali/benchmark/dali_bench.h"
+
+namespace dali {
+
+BENCHMARK_DEFINE_F(OperatorBench, CastGPU)(benchmark::State& st) {
+  int batch_size = st.range(0);
+  int H = st.range(1);
+  int W = st.range(1);
+
+  this->RunGPU<uint8_t>(st,
+                        OpSpec("Cast")
+                          .AddArg("max_batch_size", batch_size)
+                          .AddArg("num_threads", 1)
+                          .AddArg("device", "gpu")
+                          .AddArg("dtype", DALI_FLOAT),
+                        batch_size, H, W);
+}
+
+BENCHMARK_REGISTER_F(OperatorBench, CastGPU)->Iterations(100)
+->Unit(benchmark::kMicrosecond)
+->UseRealTime()
+->ArgsProduct({
+  benchmark::CreateRange(1, 256, 2),
+  {500, 1000},
+});
+
+}  // namespace dali

dali/benchmark/coin_flip_bench.cc

@@ -0,0 +1,43 @@
+// Copyright (c) 2021, Szymon Karpiński. All rights reserved.
+//
+// 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 "dali/benchmark/operator_bench.h"
+#include "dali/benchmark/dali_bench.h"
+
+namespace dali {
+
+BENCHMARK_DEFINE_F(OperatorBench, CoinFlipGPU)(benchmark::State& st) {
+  int batch_size = st.range(0);
+  int H = st.range(1);
+  int W = st.range(1);
+
+  this->RunGPU<uint8_t>(st,
+                        OpSpec("CoinFlip")
+                          .AddArg("max_batch_size", batch_size)
+                          .AddArg("num_threads", 1)
+                          .AddArg("device", "gpu")
+                          .AddInput("data", "gpu"),
+                        batch_size, H, W, 1);
+}
+
+BENCHMARK_REGISTER_F(OperatorBench, CoinFlipGPU)->Iterations(100)
+->Unit(benchmark::kMicrosecond)
+->UseRealTime()
+->ArgsProduct({
+  benchmark::CreateRange(1, 256, 2),
+  {500, 1000, 2000},
+});
+
+}  // namespace dali