fix trt engine memory sharing (#49584)

paddle/fluid/inference/api/analysis_config.cc

@@ -434,6 +434,7 @@ AnalysisConfig::AnalysisConfig(const AnalysisConfig &other) {
   CP_MEMBER(shape_range_info_path_);
   CP_MEMBER(trt_use_inspector_);
   CP_MEMBER(trt_engine_memory_sharing_);
+  CP_MEMBER(trt_engine_memory_sharing_identifier_);
   // Dlnne related
   CP_MEMBER(use_dlnne_);
   CP_MEMBER(dlnne_min_subgraph_size_);

paddle/fluid/inference/api/analysis_predictor.h

@@ -103,16 +103,17 @@ class AnalysisPredictor : public PaddlePredictor {
     if (config_.shape_range_info_collected()) {
       config_.SwitchIrOptim(false);
     }
-    auto trt_identifier = config_.trt_engine_memory_sharing_identifier_;
+    int trt_identifier = config_.trt_engine_memory_sharing_identifier_;
     if (trt_identifier > 0) {
       // NOTE(liuyuanle): For convenience, we set the id of the predictor to
       // negative sharing_identifier directly. In the future, this may affect
       // the meaning of negative predictor id.
       predictor_id_ = -trt_identifier;
-      LOG_FIRST_N(WARNING, 1)
+      LOG(WARNING)
           << "Since the engine context memory of multiple predictors "
-             "is enabled in Paddle-TRT, we set the id of current predictor to "
-             "negative sharing_identifier you specified.";
+             "is enabled in Paddle-TRT, we set the id of these predictors to "
+             "negative sharing_identifier you specified : "
+          << predictor_id_;
     } else {
       predictor_id_ = inference::GetUniqueId();
     }