Per record cost estimation progress

4 files changed, 102 insertions, 8 deletions

diff --git a/CMakeLists.txt b/CMakeLists.txt
index a7bcad6..31cdc50 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -348,6 +348,11 @@ if (tail_bench)
     target_include_directories(stall_benchmark_vptree PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include)
     target_link_options(stall_benchmark_vptree PUBLIC -mcx16)
 
+    add_executable(isam_construction_times ${CMAKE_CURRENT_SOURCE_DIR}/benchmarks/tail-latency/isam_construction_times.cpp)
+    target_link_libraries(isam_construction_times PUBLIC gsl pthread atomic)
+    target_include_directories(isam_construction_times PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include)
+    target_link_options(isam_construction_times PUBLIC -mcx16)
+
     add_executable(selectivity_sweep ${CMAKE_CURRENT_SOURCE_DIR}/benchmarks/tail-latency/selectivity_sweep.cpp)
     target_link_libraries(selectivity_sweep PUBLIC gsl pthread atomic)
     target_include_directories(selectivity_sweep PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include)
diff --git a/benchmarks/tail-latency/isam_construction_times.cpp b/benchmarks/tail-latency/isam_construction_times.cpp
new file mode 100644
index 0000000..ecf085c
--- /dev/null
+++ b/benchmarks/tail-latency/isam_construction_times.cpp
@@ -0,0 +1,87 @@
+/*
+ *
+ */
+
+#define ENABLE_TIMER
+#define TS_TEST
+
+#include "framework/scheduling/FIFOScheduler.h"
+#include "framework/DynamicExtension.h"
+#include "shard/ISAMTree.h"
+#include "query/rangecount.h"
+#include "framework/interface/Record.h"
+#include "file_util.h"
+#include "standard_benchmarks.h"
+#include "framework/util/Configuration.h"
+
+#include <gsl/gsl_rng.h>
+
+
+typedef de::Record<uint64_t, uint64_t> Rec;
+typedef de::ISAMTree<Rec>Shard;
+typedef de::rc::Query<Shard> Q;
+typedef de::DynamicExtension<Shard, Q, de::DeletePolicy::TOMBSTONE, de::FIFOScheduler> Ext;
+typedef Q::Parameters QP;
+typedef de::DEConfiguration<Shard, Q, de::DeletePolicy::TOMBSTONE,
+                            de::FIFOScheduler>
+    Conf;
+
+void usage(char *progname) {
+    fprintf(stderr, "%s reccnt datafile\n", progname);
+}
+
+int main(int argc, char **argv) {
+
+    if (argc < 3) {
+        usage(argv[0]);
+        exit(EXIT_FAILURE);
+    }
+
+    size_t n = atol(argv[1]);
+    std::string d_fname = std::string(argv[2]);
+
+    
+    auto data = read_sosd_file<Rec>(d_fname, n);
+
+    std::vector<int> policies = {6};
+    std::vector<size_t> buffers = {12000}; 
+    std::vector<size_t> sfs = {8};
+
+    for (size_t l=0; l<policies.size(); l++) {
+    for (size_t j=0; j<buffers.size(); j++) {
+    for (size_t k=0; k<sfs.size(); k++) {
+        auto policy = get_policy<Shard, Q>(sfs[k], buffers[j], policies[l]);
+        auto config = Conf(std::move(policy));
+        config.recon_enable_maint_on_flush = true;
+        config.recon_maint_disabled = false;
+        // config.buffer_flush_trigger = 4000;
+        config.maximum_threads = 6;
+        auto extension = new Ext(std::move(config));
+
+        /* warmup structure w/ 10% of records */
+        size_t warmup = .1 * n;
+        for (size_t i=0; i<warmup; i++) {
+            while (!extension->insert(data[i])) {
+                usleep(1);
+            }
+        }
+
+        extension->await_version();
+
+        for (size_t i=warmup; i<data.size(); i++) {
+            while (!extension->insert(data[i])) {
+                usleep(1);
+            }
+        }
+
+        extension->await_version();
+        
+        extension->print_scheduler_statistics();
+        delete extension;
+       
+    }}}
+
+
+    fflush(stderr);
+}
+
diff --git a/include/framework/scheduling/statistics.h b/include/framework/scheduling/statistics.h
index 34699f1..6d9f9f0 100644
--- a/include/framework/scheduling/statistics.h
+++ b/include/framework/scheduling/statistics.h
@@ -118,10 +118,9 @@ public:
     size_t first_query = UINT64_MAX;
 
 
-    /* hard-coded for the moment to only consider queries */
     for (auto &job : m_jobs) {
       if (job.second.type != 1) {
-        continue;
+        fprintf(stdout, "%ld %ld %ld %ld\n", job.second.id, job.second.size, job.second.runtime(), job.second.runtime() / (job.second.size));
       }
 
       if (job.first < first_query) {
@@ -152,8 +151,8 @@ public:
     }
     
 
-    int64_t average_queue_time = total_queue_time / query_cnt;
-    int64_t average_runtime = total_runtime / query_cnt;
+    int64_t average_queue_time = (query_cnt) ? total_queue_time / query_cnt : 0;
+    int64_t average_runtime = (query_cnt) ? total_runtime / query_cnt : 0;
 
     /* calculate standard deviations */
     int64_t queue_deviation_sum = 0;
@@ -168,8 +167,8 @@ public:
       
     }
 
-    int64_t queue_stddev = std::sqrt(queue_deviation_sum / query_cnt);
-    int64_t runtime_stddev = std::sqrt(runtime_deviation_sum / query_cnt);
+    int64_t queue_stddev = (query_cnt) ? std::sqrt(queue_deviation_sum / query_cnt) : 0;
+    int64_t runtime_stddev = (query_cnt) ? std::sqrt(runtime_deviation_sum / query_cnt) : 0;
     
 
     fprintf(stdout, "Query Count: %ld\tWorst Query: %ld\tFirst Query: %ld\n", query_cnt, worst_query, first_query);
diff --git a/include/util/types.h b/include/util/types.h
index c10f7ff..88774f5 100644
--- a/include/util/types.h
+++ b/include/util/types.h
@@ -115,7 +115,7 @@ public:
                           size_t reccnt, ReconstructionType type) {
 
     m_tasks.push_back({std::move(shards), target, reccnt, type});
-    
+    total_reccnt += reccnt;
   }
 
   void add_reconstruction(level_index source, level_index target,
@@ -130,7 +130,10 @@ public:
     total_reccnt += reccnt;
   }
 
-  void add_reconstruction(ReconstructionTask task) { m_tasks.push_back(task); }
+  void add_reconstruction(ReconstructionTask task) {
+    m_tasks.push_back(task);
+    total_reccnt += task.reccnt;
+  }
 
   size_t get_total_reccnt() { return total_reccnt; }