Partially fix issue #43 (only for count)

Author: bwyogatama

src/cuda/cudf/cudf_aggregate.cu

@@ -25,11 +25,15 @@ static std::unique_ptr<cudf::reduce_aggregation> make_reduce_aggregation()
 
 void cudf_aggregate(vector<shared_ptr<GPUColumn>>& column, uint64_t num_aggregates, AggregationType* agg_mode) 
 {
+    GPUBufferManager *gpuBufferManager = &(GPUBufferManager::GetInstance());
+    cudf::set_current_device_resource(gpuBufferManager->mr);
     if (column[0]->column_length == 0) {
         SIRIUS_LOG_DEBUG("Input size is 0");
         for (int agg_idx = 0; agg_idx < num_aggregates; agg_idx++) {
             if (agg_mode[agg_idx] == AggregationType::COUNT_STAR || agg_mode[agg_idx] == AggregationType::COUNT) {
-                column[agg_idx] = make_shared_ptr<GPUColumn>(0, ColumnType::INT64, column[agg_idx]->data_wrapper.data);
+                uint64_t* temp = gpuBufferManager->customCudaMalloc<uint64_t>(1, 0, 0);
+                cudaMemset(temp, 0, sizeof(uint64_t));
+                column[agg_idx] = make_shared_ptr<GPUColumn>(1, ColumnType::INT64, reinterpret_cast<uint8_t*>(temp));
             } else {
                 column[agg_idx] = make_shared_ptr<GPUColumn>(0, column[agg_idx]->data_wrapper.type, column[agg_idx]->data_wrapper.data);
             }
@@ -42,9 +46,6 @@ void cudf_aggregate(vector<shared_ptr<GPUColumn>>& column, uint64_t num_aggregat
     SETUP_TIMING();
     START_TIMER();
 
-    GPUBufferManager *gpuBufferManager = &(GPUBufferManager::GetInstance());
-    cudf::set_current_device_resource(gpuBufferManager->mr);
-
     uint64_t size = 0;
     for (int agg = 0; agg < num_aggregates; agg++) {
         if (column[agg]->data_wrapper.data != nullptr || (column[agg]->data_wrapper.data == nullptr && agg_mode[agg] == AggregationType::COUNT_STAR && column[agg]->column_length > 0)) {