From 265610435e1164a9acc39ca02ea1139acd37c46c Mon Sep 17 00:00:00 2001 From: Douglas Rumbaugh Date: Thu, 9 May 2024 14:10:29 -0400 Subject: Added benchmarks for BigANN --- benchmarks/vldb/btree_bench.cpp | 2 +- benchmarks/vldb/irs_bench.cpp | 2 +- benchmarks/vldb/mtree_bench_alt.cpp | 80 ++++++++++++++++++++++++ benchmarks/vldb/vptree_bench_alt.cpp | 102 +++++++++++++++++++++++++++++++ benchmarks/vldb/vptree_bsm_bench_alt.cpp | 92 ++++++++++++++++++++++++++++ 5 files changed, 276 insertions(+), 2 deletions(-) create mode 100644 benchmarks/vldb/mtree_bench_alt.cpp create mode 100644 benchmarks/vldb/vptree_bench_alt.cpp create mode 100644 benchmarks/vldb/vptree_bsm_bench_alt.cpp (limited to 'benchmarks/vldb') diff --git a/benchmarks/vldb/btree_bench.cpp b/benchmarks/vldb/btree_bench.cpp index 673da33..fa72831 100644 --- a/benchmarks/vldb/btree_bench.cpp +++ b/benchmarks/vldb/btree_bench.cpp @@ -57,7 +57,7 @@ int main(int argc, char **argv) { } /* warmup structure w/ 10% of records */ - size_t warmup = .3 * n; + size_t warmup = .1 * n; size_t delete_idx = 0; insert_records(&btree, 0, warmup, data, to_delete, delete_idx, false, rng); diff --git a/benchmarks/vldb/irs_bench.cpp b/benchmarks/vldb/irs_bench.cpp index f68de63..e062e80 100644 --- a/benchmarks/vldb/irs_bench.cpp +++ b/benchmarks/vldb/irs_bench.cpp @@ -56,7 +56,7 @@ int main(int argc, char **argv) { } /* warmup structure w/ 10% of records */ - size_t warmup = .3 * n; + size_t warmup = .1 * n; size_t delete_idx = 0; insert_records(extension, 0, warmup, data, to_delete, delete_idx, false, rng); diff --git a/benchmarks/vldb/mtree_bench_alt.cpp b/benchmarks/vldb/mtree_bench_alt.cpp new file mode 100644 index 0000000..6b08df7 --- /dev/null +++ b/benchmarks/vldb/mtree_bench_alt.cpp @@ -0,0 +1,80 @@ +/* + * + */ + +#define ENABLE_TIMER + +#include "query/knn.h" +#include "file_util.h" +#include "standard_benchmarks.h" + +#include + +#include "psu-util/timer.h" + + +typedef ANNRec Rec; +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]); + + auto mtree = new MTree_alt(); + gsl_rng * rng = gsl_rng_alloc(gsl_rng_mt19937); + + fprintf(stderr, "[I] Reading data file...\n"); + auto data = read_binary_vector_file(d_fname, n); + + fprintf(stderr, "[I] Generating delete vector\n"); + std::vector to_delete(n * delete_proportion); + size_t j=0; + for (size_t i=0; i(q_fname, 1000, 100); + + fprintf(stderr, "[I] Warming up structure...\n"); + /* warmup structure w/ 10% of records */ + size_t warmup = .1 * n; + size_t delete_idx = 0; + insert_records(mtree, 0, warmup, data, to_delete, delete_idx, false, rng); + + TIMER_INIT(); + + fprintf(stderr, "[I] Running Insertion Benchmark\n"); + TIMER_START(); + insert_records(mtree, 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); + + fprintf(stderr, "[I] Running Query Benchmark\n"); + TIMER_START(); + run_queries(mtree, queries); + TIMER_STOP(); + + auto query_latency = TIMER_RESULT() / queries.size(); + + fprintf(stdout, "%ld\t%ld\n", insert_throughput, query_latency); + + gsl_rng_free(rng); + delete mtree; + fflush(stderr); +} + diff --git a/benchmarks/vldb/vptree_bench_alt.cpp b/benchmarks/vldb/vptree_bench_alt.cpp new file mode 100644 index 0000000..b09ee7d --- /dev/null +++ b/benchmarks/vldb/vptree_bench_alt.cpp @@ -0,0 +1,102 @@ +/* + * + */ + +#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 ANNRec Rec; + +typedef de::VPTree Shard; +typedef de::knn::Query Q; +typedef de::DynamicExtension Ext; +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]); + + auto extension = new Ext(1400, 1400, 8, 0, 64); + gsl_rng * rng = gsl_rng_alloc(gsl_rng_mt19937); + + fprintf(stderr, "[I] Reading data file...\n"); + auto data = read_binary_vector_file(d_fname, n); + + fprintf(stderr, "[I] Generating delete vector\n"); + std::vector to_delete(n * delete_proportion); + size_t j=0; + for (size_t i=0; i(q_fname, 1000, 100); + + fprintf(stderr, "[I] Warming up structure...\n"); + /* 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(); + + fprintf(stderr, "[I] Running Insertion Benchmark\n"); + 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); + + fprintf(stderr, "[I] Running Query Benchmark\n"); + TIMER_START(); + run_queries(extension, queries); + TIMER_STOP(); + + auto query_latency = TIMER_RESULT() / queries.size(); + + auto shard = extension->create_static_structure(); + + fprintf(stderr, "Running Static query tests\n\n"); + TIMER_START(); + run_static_queries(shard, queries); + TIMER_STOP(); + + auto static_latency = TIMER_RESULT() / queries.size(); + + auto ext_size = extension->get_memory_usage() + extension->get_aux_memory_usage(); + auto static_size = shard->get_memory_usage(); // + shard->get_aux_memory_usage(); + + fprintf(stdout, "%ld\t%ld\t%ld\t%ld\t%ld\n", insert_throughput, query_latency, ext_size, static_latency, static_size); + + gsl_rng_free(rng); + delete extension; + fflush(stderr); + fflush(stdout); +} + diff --git a/benchmarks/vldb/vptree_bsm_bench_alt.cpp b/benchmarks/vldb/vptree_bsm_bench_alt.cpp new file mode 100644 index 0000000..63baf8b --- /dev/null +++ b/benchmarks/vldb/vptree_bsm_bench_alt.cpp @@ -0,0 +1,92 @@ +/* + * + */ + +#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 ANNRec Rec; + +typedef de::VPTree Shard; +typedef de::knn::Query Q; +typedef de::DynamicExtension Ext; +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]); + + auto extension = new Ext(1, 1400, 2, 0, 64); + gsl_rng * rng = gsl_rng_alloc(gsl_rng_mt19937); + + fprintf(stderr, "[I] Reading data file...\n"); + auto data = read_binary_vector_file(d_fname, n); + + fprintf(stderr, "[I] Generating delete vector\n"); + std::vector to_delete(n * delete_proportion); + size_t j=0; + for (size_t i=0; i(q_fname, 1000, 100); + + fprintf(stderr, "[I] Warming up structure...\n"); + /* 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(); + + fprintf(stderr, "[I] Running Insertion Benchmark\n"); + 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); + + fprintf(stderr, "[I] Running Query Benchmark\n"); + 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, "%ld\t%ld\t\t%ld\n", insert_throughput, query_latency, ext_size); + + gsl_rng_free(rng); + delete extension; + fflush(stderr); + fflush(stdout); +} + -- cgit v1.2.3