From 909ba1e59ce654db3ea9294201dec2bc826b0b72 Mon Sep 17 00:00:00 2001 From: Douglas Rumbaugh Date: Tue, 23 Apr 2024 13:25:17 -0400 Subject: Added vptree parmsweep benchmark and fixed some CMake issues --- CMakeLists.txt | 12 ++++ benchmarks/vldb/vptree_parmsweep.cpp | 129 +++++++++++++++++++++++++++++++++++ 2 files changed, 141 insertions(+) create mode 100644 benchmarks/vldb/vptree_parmsweep.cpp diff --git a/CMakeLists.txt b/CMakeLists.txt index 3d812a0..37926ca 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -155,31 +155,43 @@ if (vldb_bench) target_link_libraries(irs_bench PUBLIC gsl pthread atomic) target_include_directories(irs_bench PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include) target_link_options(irs_bench PUBLIC -mcx16) + target_compile_options(irs_bench PUBLIC -fopenmp) add_executable(vptree_bench ${CMAKE_CURRENT_SOURCE_DIR}/benchmarks/vldb/vptree_bench.cpp) target_link_libraries(vptree_bench PUBLIC gsl pthread atomic) target_include_directories(vptree_bench PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include) target_link_options(vptree_bench PUBLIC -mcx16) + target_compile_options(vptree_bench PUBLIC -fopenmp) + + add_executable(vptree_parmsweep ${CMAKE_CURRENT_SOURCE_DIR}/benchmarks/vldb/vptree_parmsweep.cpp) + target_link_libraries(vptree_parmsweep PUBLIC gsl pthread atomic) + target_include_directories(vptree_parmsweep PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include) + target_link_options(vptree_parmsweep PUBLIC -mcx16) + target_compile_options(vptree_parmsweep PUBLIC -fopenmp) add_executable(ts_bench ${CMAKE_CURRENT_SOURCE_DIR}/benchmarks/vldb/ts_bench.cpp) target_link_libraries(ts_bench PUBLIC gsl pthread atomic) target_include_directories(ts_bench PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include) target_link_options(ts_bench PUBLIC -mcx16) + target_compile_options(ts_bench PUBLIC -fopenmp) add_executable(ts_parmsweep ${CMAKE_CURRENT_SOURCE_DIR}/benchmarks/vldb/ts_parmsweep.cpp) target_link_libraries(ts_parmsweep PUBLIC gsl pthread atomic) target_include_directories(ts_parmsweep PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include) target_link_options(ts_parmsweep PUBLIC -mcx16) + target_compile_options(ts_parmsweep PUBLIC -fopenmp) add_executable(ts_bsm_bench ${CMAKE_CURRENT_SOURCE_DIR}/benchmarks/vldb/ts_bsm_bench.cpp) target_link_libraries(ts_bsm_bench PUBLIC gsl pthread atomic) target_include_directories(ts_bsm_bench PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include) target_link_options(ts_bsm_bench PUBLIC -mcx16) + target_compile_options(ts_bsm_bench PUBLIC -fopenmp) add_executable(ts_mdsp_bench ${CMAKE_CURRENT_SOURCE_DIR}/benchmarks/vldb/ts_mdsp_bench.cpp) target_link_libraries(ts_mdsp_bench PUBLIC gsl pthread atomic) target_include_directories(ts_mdsp_bench PRIVATE include external external/m-tree/cpp external/PGM-index/include external/PLEX/include benchmarks/include external/psudb-common/cpp/include) target_link_options(ts_mdsp_bench PUBLIC -mcx16) + target_compile_options(ts_mdsp_bench PUBLIC -fopenmp) add_executable(pgm_bench ${CMAKE_CURRENT_SOURCE_DIR}/benchmarks/vldb/pgm_bench.cpp) target_link_libraries(pgm_bench PUBLIC gsl pthread atomic gomp) diff --git a/benchmarks/vldb/vptree_parmsweep.cpp b/benchmarks/vldb/vptree_parmsweep.cpp new file mode 100644 index 0000000..2cbd521 --- /dev/null +++ b/benchmarks/vldb/vptree_parmsweep.cpp @@ -0,0 +1,129 @@ +/* + * + */ + +#define ENABLE_TIMER + +#include "framework/DynamicExtension.h" +#include "shard/VPTree.h" +#include "query/knn.h" +#include "framework/interface/Record.h" +#include "file_util.h" +#include "standard_benchmarks.h" + +#include + +#include "psu-util/timer.h" + + +typedef Word2VecRec Rec; + +typedef de::VPTree Shard; +typedef de::knn::Query Q; +typedef de::DynamicExtension Ext; +typedef de::DynamicExtension Ext2; +typedef de::knn::Parms QP; + +void usage(char *progname) { + fprintf(stderr, "%s reccnt datafile queryfile\n", progname); +} + +int main(int argc, char **argv) { + + if (argc < 4) { + usage(argv[0]); + exit(EXIT_FAILURE); + } + + size_t n = atol(argv[1]); + std::string d_fname = std::string(argv[2]); + std::string q_fname = std::string(argv[3]); + + gsl_rng * rng = gsl_rng_alloc(gsl_rng_mt19937); + + auto data = read_vector_file(d_fname, n); + + std::vector to_delete(n * delete_proportion); + size_t j=0; + for (size_t i=0; i(q_fname, 10); + + + const std::vector policies = {de::LayoutPolicy::LEVELING, de::LayoutPolicy::TEIRING}; + const std::vector buffer_sizes = {100, 400, 800, 1200, 1500, 2000}; + const std::vector scale_factors = {2, 4, 6, 8, 10, 12}; + + for (const auto &bs : buffer_sizes) { + for (const auto &sf : scale_factors) { + auto extension = new Ext(bs, bs, sf, 0, 64); + + /* warmup structure w/ 10% of records */ + size_t warmup = .1 * n; + size_t delete_idx = 0; + insert_records(extension, 0, warmup, data, to_delete, delete_idx, false, rng); + + extension->await_next_epoch(); + + TIMER_INIT(); + + TIMER_START(); + insert_records(extension, warmup, data.size(), data, to_delete, delete_idx, true, rng); + TIMER_STOP(); + + auto insert_latency = TIMER_RESULT(); + size_t insert_throughput = (size_t) ((double) (n - warmup) / (double) insert_latency * 1e9); + + TIMER_START(); + run_queries(extension, queries); + TIMER_STOP(); + + auto query_latency = TIMER_RESULT() / queries.size(); + + auto ext_size = extension->get_memory_usage() + extension->get_aux_memory_usage(); + + fprintf(stdout, "TIERING\t%ld\t%ld\t%ld\t%ld\t%ld\n", bs, sf, insert_throughput, query_latency, ext_size); + delete extension; + } + } + + for (const auto &bs : buffer_sizes) { + for (const auto &sf : scale_factors) { + auto extension = new Ext2(bs, bs, sf, 0, 64); + + /* warmup structure w/ 10% of records */ + size_t warmup = .1 * n; + size_t delete_idx = 0; + insert_records(extension, 0, warmup, data, to_delete, delete_idx, false, rng); + + extension->await_next_epoch(); + + TIMER_INIT(); + + TIMER_START(); + insert_records(extension, warmup, data.size(), data, to_delete, delete_idx, true, rng); + TIMER_STOP(); + + auto insert_latency = TIMER_RESULT(); + size_t insert_throughput = (size_t) ((double) (n - warmup) / (double) insert_latency * 1e9); + + TIMER_START(); + run_queries(extension, queries); + TIMER_STOP(); + + auto query_latency = TIMER_RESULT() / queries.size(); + + auto ext_size = extension->get_memory_usage() + extension->get_aux_memory_usage(); + + fprintf(stdout, "LEVELING\t%ld\t%ld\t%ld\t%ld\t%ld\n", bs, sf, insert_throughput, query_latency, ext_size); + delete extension; + } + } + + gsl_rng_free(rng); + fflush(stderr); +} + -- cgit v1.2.3