From cf4cfc1908bb04c4f5680aa011fcf0d93075bd7f Mon Sep 17 00:00:00 2001
From: Emmanuel Keller <emmanuel.keller@surrealdb.com>
Date: Fri, 7 Jul 2023 19:43:16 +0100
Subject: [PATCH] processor micro-benchmark (#2216)

---
 Cargo.lock               |   1 +
 lib/Cargo.toml           |   6 ++-
 lib/benches/processor.rs | 112 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 118 insertions(+), 1 deletion(-)
 create mode 100644 lib/benches/processor.rs

diff --git a/Cargo.lock b/Cargo.lock
index 74d264a7..f4f4b5be 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -1191,6 +1191,7 @@ dependencies = [
  "ciborium",
  "clap 3.2.25",
  "criterion-plot",
+ "futures 0.3.28",
  "itertools 0.10.5",
  "lazy_static",
  "num-traits",
diff --git a/lib/Cargo.toml b/lib/Cargo.toml
index 9b3b8b56..bf7641b0 100644
--- a/lib/Cargo.toml
+++ b/lib/Cargo.toml
@@ -114,7 +114,7 @@ url = "2.4.0"
 bytes = "1.4.0"
 
 [dev-dependencies]
-criterion = "0.4"
+criterion = { version="0.4", features= [  "async_futures" ] }
 env_logger = "0.10.0"
 test-log = "0.2.12"
 pprof = { version = "0.11.1", features = [ "flamegraph", "criterion" ] }
@@ -147,3 +147,7 @@ harness = false
 [[bench]]
 name = "parser"
 harness = false
+
+[[bench]]
+name = "processor"
+harness = false
\ No newline at end of file
diff --git a/lib/benches/processor.rs b/lib/benches/processor.rs
new file mode 100644
index 00000000..e8a1eb30
--- /dev/null
+++ b/lib/benches/processor.rs
@@ -0,0 +1,112 @@
+use criterion::async_executor::FuturesExecutor;
+use criterion::{black_box, criterion_group, criterion_main, Criterion, Throughput};
+use std::time::Duration;
+use surrealdb::dbs::Session;
+use surrealdb::kvs::Datastore;
+use surrealdb::sql::Value;
+
+fn bench_processor(c: &mut Criterion) {
+	let rt = tokio::runtime::Runtime::new().unwrap();
+	let i = rt.block_on(prepare_data());
+
+	let mut group = c.benchmark_group("processor");
+	group.throughput(Throughput::Elements(1));
+	group.sample_size(10);
+	group.measurement_time(Duration::from_secs(15));
+
+	group.bench_function("table-iterator", |b| {
+		b.to_async(FuturesExecutor).iter(|| run(&i, "SELECT * FROM item", i.count * 5))
+	});
+
+	group.bench_function("table-iterator-parallel", |b| {
+		b.to_async(FuturesExecutor).iter(|| run(&i, "SELECT * FROM item PARALLEL", i.count * 5))
+	});
+
+	group.bench_function("non-uniq-index-iterator", |b| {
+		b.to_async(FuturesExecutor).iter(|| run(&i, "SELECT * FROM item WHERE number=4", i.count))
+	});
+
+	group.bench_function("non-uniq-index-iterator-parallel", |b| {
+		b.to_async(FuturesExecutor)
+			.iter(|| run(&i, "SELECT * FROM item WHERE number=4 PARALLEL", i.count))
+	});
+
+	group.bench_function("full-text-index-iterator", |b| {
+		b.to_async(FuturesExecutor)
+			.iter(|| run(&i, "SELECT * FROM item WHERE label @@ 'charlie'", i.count))
+	});
+
+	group.bench_function("full-text-index-iterator-parallel", |b| {
+		b.to_async(FuturesExecutor)
+			.iter(|| run(&i, "SELECT * FROM item WHERE label @@ 'charlie' PARALLEL", i.count))
+	});
+
+	group.finish();
+}
+
+struct Input {
+	dbs: Datastore,
+	ses: Session,
+	count: usize,
+}
+
+async fn prepare_data() -> Input {
+	let dbs = Datastore::new("memory").await.unwrap();
+	let ses = Session::for_kv().with_ns("bench").with_db("bench");
+	let sql = format!(
+		r"DEFINE INDEX number ON item FIELDS number;
+		DEFINE ANALYZER simple TOKENIZERS blank,class;
+		DEFINE INDEX search ON item FIELDS label SEARCH ANALYZER simple BM25"
+	);
+	let res = &mut dbs.execute(&sql, &ses, None).await.unwrap();
+	for _ in 0..3 {
+		assert!(res.remove(0).result.is_ok());
+	}
+
+	let count = if cfg!(debug_assertions) {
+		100 // debug is much slower!
+	} else {
+		10_000
+	};
+
+	for i in 0..count {
+		let j = i * 5;
+		let a = j;
+		let b = j + 1;
+		let c = j + 2;
+		let d = j + 3;
+		let e = j + 4;
+		let sql = format!(
+			r"CREATE item SET id = {a}, name = '{a}', number = 0, label='alpha';
+		CREATE item SET id = {b}, name = '{b}', number = 1, label='bravo';
+		CREATE item SET id = {c}, name = '{c}', number = 2, label='charlie';
+		CREATE item SET id = {d}, name = '{d}', number = 3, label='delta';
+		CREATE item SET id = {e}, name = '{e}', number = 4, label='echo';",
+		);
+		let res = &mut dbs.execute(&sql, &ses, None).await.unwrap();
+		for _ in 0..5 {
+			assert!(res.remove(0).result.is_ok());
+		}
+	}
+	Input {
+		dbs,
+		ses,
+		count,
+	}
+}
+
+async fn run(i: &Input, q: &str, expected: usize) {
+	let mut r = i.dbs.execute(black_box(q), &i.ses, None).await.unwrap();
+	if cfg!(debug_assertions) {
+		assert_eq!(r.len(), 1);
+		if let Value::Array(a) = r.remove(0).result.unwrap() {
+			assert_eq!(a.len(), expected);
+		} else {
+			panic!("Fail");
+		}
+	}
+	black_box(r);
+}
+
+criterion_group!(benches, bench_processor);
+criterion_main!(benches);