Merge pull request #418 from dralley/benches

Parameterize benchmarks, add throughput measurements

Author: Mingun

Changelog.md

@@ -121,6 +121,7 @@
 - [#393]: Added tests for reserved names (started with "xml"i) -- see <https://www.w3.org/TR/xml-names11/#xmlReserved>
 - [#363]: Add tests for `Reader::read_event_impl` to ensure that proper events generated for corresponding inputs
 - [#407]: Improved benchmark suite to cover whole-document parsing, escaping and unescaping text
+- [#418]: Parameterized macrobenchmarks and comparative benchmarks, added throughput measurements via criterion
 
 [#8]: https://github.com/Mingun/fast-xml/pull/8
 [#9]: https://github.com/Mingun/fast-xml/pull/9
@@ -137,6 +138,7 @@
 [#407]: https://github.com/tafia/quick-xml/pull/407
 [#412]: https://github.com/tafia/quick-xml/pull/412
 [#416]: https://github.com/tafia/quick-xml/pull/416
+[#418]: https://github.com/tafia/quick-xml/pull/418
 
 ## 0.23.0 -- 2022-05-08
 

benches/macrobenches.rs

@@ -1,4 +1,4 @@
-use criterion::{self, criterion_group, Criterion};
+use criterion::{self, criterion_group, Criterion, Throughput};
 use quick_xml::events::Event;
 use quick_xml::Reader;
 use quick_xml::Result as XmlResult;
@@ -44,89 +44,39 @@ fn parse_document(doc: &[u8]) -> XmlResult<()> {
 pub fn bench_fully_parse_document(c: &mut Criterion) {
     let mut group = c.benchmark_group("fully_parse_document");
 
-    // long, mix of attributes and text, not much escaping, mix of attribute lengths, some namespaces
-    group.bench_function("rpm_primary.xml", |b| {
-        b.iter(|| {
-            parse_document(RPM_PRIMARY).unwrap();
-        })
-    });
-
-    // long, mix of attributes and text, not much escaping, mix of attribute lengths, some namespaces
-    group.bench_function("rpm_primary2.xml", |b| {
-        b.iter(|| {
-            parse_document(RPM_PRIMARY2).unwrap();
-        })
-    });
-
-    // long, mostly medium-length text elements, not much escaping
-    group.bench_function("rpm_filelists.xml", |b| {
-        b.iter(|| {
-            parse_document(RPM_FILELISTS).unwrap();
-        })
-    });
-
-    // long, mix of attributes and text, lots of escaping (both entity and char literal), long attributes
-    group.bench_function("rpm_other.xml", |b| {
-        b.iter(|| {
-            parse_document(RPM_OTHER).unwrap();
-        })
-    });
-
-    // long, mix of attributes and text, not much escaping, lots of non-ascii characters, lots of namespaces
-    group.bench_function("libreoffice_document.fodt", |b| {
-        b.iter(|| {
-            parse_document(LIBREOFFICE_DOCUMENT).unwrap();
-        })
-    });
-
-    // medium length, mostly empty tags, a few short attributes per element, no escaping
-    group.bench_function("document.xml", |b| {
-        b.iter(|| {
-            parse_document(DOCUMENT).unwrap();
-        })
-    });
-
-    // medium length, lots of namespaces, no escaping
-    group.bench_function("test_writer_ident.xml", |b| {
-        b.iter(|| {
-            parse_document(TEST_WRITER_INDENT).unwrap();
-        })
-    });
-
-    // short, mix of attributes and text, lots of escapes
-    group.bench_function("sample_1.xml", |b| {
-        b.iter(|| {
-            parse_document(SAMPLE_1).unwrap();
-        })
-    });
-
-    // medium length, lots of attributes, short attributes, few escapes
-    group.bench_function("linescore.xml", |b| {
-        b.iter(|| {
-            parse_document(LINESCORE).unwrap();
-        })
-    });
-
-    // short, lots of namespaces, no escapes
-    group.bench_function("sample_ns.xml", |b| {
-        b.iter(|| {
-            parse_document(SAMPLE_NS).unwrap();
-        })
-    });
-
-    // long, few attributes, mix of attribute lengths, escapes in text content
-    group.bench_function("sample_rss.xml", |b| {
-        b.iter(|| {
-            parse_document(SAMPLE_RSS).unwrap();
-        })
-    });
-
-    // long, lots of attributes, short attributes, no text, no escapes
-    group.bench_function("players.xml", |b| {
-        b.iter(|| {
-            parse_document(PLAYERS).unwrap();
-        })
-    });
+    let inputs = [
+        // long, mix of attributes and text, not much escaping, mix of attribute lengths, some namespaces
+        ("rpm_primary.xml", RPM_PRIMARY),
+        // long, mix of attributes and text, not much escaping, mix of attribute lengths, some namespaces
+        ("rpm_primary2.xml", RPM_PRIMARY2),
+        // long, mostly medium-length text elements, not much escaping
+        ("rpm_filelists.xml", RPM_FILELISTS),
+        // long, mix of attributes and text, lots of escaping (both entity and char literal), long attributes
+        ("rpm_other.xml", RPM_OTHER),
+        // long, mix of attributes and text, not much escaping, lots of non-ascii characters, lots of namespaces
+        ("libreoffice_document.fodt", LIBREOFFICE_DOCUMENT),
+        // medium length, mostly empty tags, a few short attributes per element, no escaping
+        ("document.xml", DOCUMENT),
+        // medium length, lots of namespaces, no escaping
+        ("test_writer_ident.xml", TEST_WRITER_INDENT),
+        // short, mix of attributes and text, lots of escapes
+        ("sample_1.xml", SAMPLE_1),
+        // medium length, lots of attributes, short attributes, few escapes
+        ("linescore.xml", LINESCORE),
+        // short, lots of namespaces, no escapes
+        ("sample_ns.xml", SAMPLE_NS),
+        // long, few attributes, mix of attribute lengths, escapes in text content
+        ("sample_rss.xml", SAMPLE_RSS),
+        // long, lots of attributes, short attributes, no text, no escapes
+        ("players.xml", PLAYERS),
+    ];
+
+    for (id, data) in inputs.iter() {
+        group.throughput(Throughput::Bytes(data.len() as u64));
+        group.bench_with_input(*id, *data, |b, input| {
+            b.iter(|| parse_document(input).unwrap())
+        });
+    }
 
     group.finish();
 }