improvements from gts repo

GraphStreamingProject · Sep 24, 2024 · 1408005 · 1408005
1 parent f68d883
commit 1408005
Show file tree

Hide file tree

Showing 3 changed files with 11 additions and 6 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -48,7 +48,7 @@ FetchContent_Declare(
   GutterTree
 
   GIT_REPOSITORY  https://github.com/GraphStreamingProject/GutterTree.git
-  GIT_TAG         main
+  GIT_TAG         better_pht
 )
 
 # Get StreamingUtilities

diff --git a/include/graph_sketch_driver.h b/include/graph_sketch_driver.h
@@ -147,6 +147,9 @@ class GraphSketchDriver {
 
     auto task = [&](int thr_id) {
       GraphStreamUpdate update_array[update_array_size];
+
+      // buffer of updates for gts
+      update_t gts_buffer[2 * update_array_size];
 #ifdef VERIFY_SAMPLES_F
       GraphVerifier local_verifier(sketching_alg->get_num_vertices());
 #endif
@@ -159,6 +162,7 @@ class GraphSketchDriver {
           upd.type = static_cast<UpdateType>(update_array[i].type);
           if (upd.type == BREAKPOINT) {
             // reached the breakpoint. Update verifier if applicable and return
+            gts->batch_insert(gts_buffer, 2 * i, thr_id);
 #ifdef VERIFY_SAMPLES_F
             std::lock_guard<std::mutex> lk(verifier_mtx);
             verifier->combine(local_verifier);
@@ -167,14 +171,14 @@ class GraphSketchDriver {
           }
           else {
             sketching_alg->pre_insert(upd, thr_id);
-            Edge edge = upd.edge;
-            gts->insert({edge.src, edge.dst}, thr_id);
-            gts->insert({edge.dst, edge.src}, thr_id);
+            gts_buffer[2 * i] = {upd.edge.src, upd.edge.dst};
+            gts_buffer[2 * i + 1] = {upd.edge.dst, upd.edge.src};
 #ifdef VERIFY_SAMPLES_F
-            local_verifier.edge_update(edge);
+            local_verifier.edge_update(upd.edge);
 #endif
           }
         }
+        gts->batch_insert(gts_buffer, 2 * updates, thr_id);
       }
     };
 

diff --git a/tools/process_stream.cpp b/tools/process_stream.cpp
@@ -85,7 +85,8 @@ int main(int argc, char **argv) {
   std::cout << std::endl;
 
   auto driver_config = DriverConfiguration().gutter_sys(CACHETREE).worker_threads(num_threads);
-  auto cc_config = CCAlgConfiguration().batch_factor(1);
+  driver_config.gutter_conf().wq_batch_per_elm(8);
+  auto cc_config = CCAlgConfiguration().batch_factor(2);
   CCSketchAlg cc_alg{num_nodes, get_seed(), cc_config};
   GraphSketchDriver<CCSketchAlg> driver{&cc_alg, &stream, driver_config, reader_threads};