Add missing stream to cudaEventRecord in bench apps

Author: tbensonatl

examples/channelize_poly_bench.cu

@@ -94,11 +94,11 @@ void ChannelizePolyBench(matx::index_t channel_start, matx::index_t channel_stop
       cudaStreamSynchronize(stream);
 
       float elapsed_ms = 0.0f;
-      cudaEventRecord(start);
+      cudaEventRecord(start, stream);
       for (int k = 0; k < NUM_ITERATIONS; k++) {
         (output = channelize_poly(input, filter, num_channels, decimation_factor)).run(stream);
       }
-      cudaEventRecord(stop);
+      cudaEventRecord(stop, stream);
       cudaStreamSynchronize(stream);
       CUDA_CHECK_LAST_ERROR();
       cudaEventElapsedTime(&elapsed_ms, start, stop);

examples/resample_poly_bench.cu

@@ -115,11 +115,11 @@ void ResamplePolyBench()
       cudaStreamSynchronize(stream);
 
       float elapsed_ms = 0.0f;
-      cudaEventRecord(start);
+      cudaEventRecord(start, stream);
       for (int k = 0; k < NUM_ITERATIONS; k++) {
         (output = matx::resample_poly(input, filter, up, down)).run(stream);
       }
-      cudaEventRecord(stop);
+      cudaEventRecord(stop, stream);
       cudaStreamSynchronize(stream);
       CUDA_CHECK_LAST_ERROR();
       cudaEventElapsedTime(&elapsed_ms, start, stop);

include/matx/transforms/resample_poly.h

@@ -137,9 +137,8 @@ inline void resample_poly_impl(OutType &out, const InType &in, const FilterType
     MATX_ASSERT_STR(out.Size(i) == in.Size(i), matxInvalidDim, "resample_poly: input/output must have matched batch sizes");
   }
 
-  const index_t up_size = in.Size(RANK-1) * up;
-  const index_t outlen = up_size / down + ((up_size % down) ? 1 : 0);
-
+  [[maybe_unused]] const index_t up_size = in.Size(RANK-1) * up;
+  [[maybe_unused]] const index_t outlen = up_size / down + ((up_size % down) ? 1 : 0);
   MATX_ASSERT_STR(out.Size(RANK-1) == outlen, matxInvalidDim, "resample_poly: output size mismatch");
 
   const index_t g = gcd(up, down);