diff options
| author | Douglas B. Rumbaugh <dbr4@psu.edu> | 2024-05-14 16:31:05 -0400 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-05-14 16:31:05 -0400 |
| commit | 47916da2ba5ed5bee2dda3cbcc58d39e1e931bfc (patch) | |
| tree | ee5613ce182b2c9caa228d3abeb65dc27fef2db3 /benchmarks/vldb/fst_bench.cpp | |
| parent | 4a834497d5f82c817d634925250158d85ca825c2 (diff) | |
| parent | 8643fe194dec05b4e3f3ea31e162ac0b2b00e162 (diff) | |
| download | dynamic-extension-47916da2ba5ed5bee2dda3cbcc58d39e1e931bfc.tar.gz | |
Merge pull request #4 from dbrumbaugh/master
Updates for VLDB revision
Diffstat (limited to 'benchmarks/vldb/fst_bench.cpp')
| -rw-r--r-- | benchmarks/vldb/fst_bench.cpp | 100 |
1 files changed, 100 insertions, 0 deletions
diff --git a/benchmarks/vldb/fst_bench.cpp b/benchmarks/vldb/fst_bench.cpp new file mode 100644 index 0000000..276a922 --- /dev/null +++ b/benchmarks/vldb/fst_bench.cpp @@ -0,0 +1,100 @@ +/* + * + */ + +#define ENABLE_TIMER +#define TS_TEST + +#include <thread> + +#include "framework/DynamicExtension.h" +#include "shard/FSTrie.h" +#include "query/pointlookup.h" +#include "framework/interface/Record.h" +#include "file_util.h" +#include "standard_benchmarks.h" + +#include <gsl/gsl_rng.h> + +#include "psu-util/timer.h" + + +typedef de::Record<const char *, uint64_t> Rec; +typedef de::FSTrie<Rec> Shard; +typedef de::pl::Query<Rec, Shard> Q; +typedef de::DynamicExtension<Rec, Shard, Q, de::LayoutPolicy::TEIRING, de::DeletePolicy::TOMBSTONE, de::SerialScheduler> Ext; +typedef de::pl::Parms<Rec> QP; + +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 extension = new Ext(12000, 12001, 8, 0, 64); + gsl_rng * rng = gsl_rng_alloc(gsl_rng_mt19937); + + auto strings = read_string_file(d_fname, n); + auto queries = generate_string_lookup_queries<QP>(strings, 1000, rng); + + std::vector<Rec> data; + for (size_t i=0; i<strings.size(); i++) { + data.push_back({strings[i].get(), i, strlen(strings[i].get())}); + } + + std::vector<size_t> to_delete(n * delete_proportion); + size_t j=0; + for (size_t i=0; i<strings.size() && j<to_delete.size(); i++) { + if (gsl_rng_uniform(rng) <= delete_proportion) { + to_delete[j++] = i; + } + } + + /* warmup structure w/ 10% of records */ + size_t warmup = .1 * n; + size_t delete_idx = 0; + insert_records<Ext, Rec>(extension, 0, warmup, data, to_delete, delete_idx, false, rng); + + extension->await_next_epoch(); + + TIMER_INIT(); + + TIMER_START(); + insert_records<Ext, Rec>(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<Ext, QP>(extension, queries); + TIMER_STOP(); + + auto query_latency = TIMER_RESULT() / queries.size(); + + auto shard = extension->create_static_structure(); + + TIMER_START(); + run_static_queries<Shard, QP, Q>(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); +} + |