Improve granularity of benchmarks (#122)

AaronO · web-flow · commit 45b60feb4927 · 2023-04-13T21:22:48.000Z
To provide better baselines for perf work, etc...

It unfortunately requires exposing a few functions as public, though they are hidden from the docs and re-exported under a `_benchable` mod. They have warnings.

Don't use them. We will change them. You will be sad. You've been warned.
diff --git a/benches/parse.rs b/benches/parse.rs
@@ -83,9 +83,112 @@ fn resp_short(c: &mut Criterion) {
         }));
 }
 
+fn uri(c: &mut Criterion) {
+    fn _uri(c: &mut Criterion, name: &str, input: &'static [u8]) {
+        c.benchmark_group("uri")
+        .throughput(Throughput::Bytes(input.len() as u64))
+        .bench_function(name, |b| b.iter(|| {
+            black_box({
+                let mut b = httparse::_benchable::Bytes::new(input);
+                httparse::_benchable::parse_uri(&mut b).unwrap()
+            });
+        }));
+    }
+
+    const S: &[u8] = b" ";
+    const CHUNK64: &[u8] = b"/wp-content/uploads/2022/08/31/hello-kitty-darth-vader-pink.webp";
+    let chunk_4k = CHUNK64.repeat(64);
+    
+    // 1b to 4096b
+    for p in 0..=12 {
+        let n = 1 << p;
+        _uri(c, &format!("uri_{}b", n), [chunk_4k[..n].to_vec(), S.into()].concat().leak());
+    }
+}
+
+fn header(c: &mut Criterion) {
+    fn _header(c: &mut Criterion, name: &str, input: &'static [u8]) {
+        let mut headers = [httparse::EMPTY_HEADER; 128];
+        c.benchmark_group("header")
+        .throughput(Throughput::Bytes(input.len() as u64))
+        .bench_function(name, |b| b.iter(|| {
+            black_box({
+                let _ = httparse::parse_headers(input, &mut headers).unwrap();
+            });
+        }));
+    }
+        
+    const RN: &[u8] = b"\r\n";
+    const RNRN: &[u8] = b"\r\n\r\n";
+    const TINY_RN: &[u8] = b"a: b\r\n"; // minimal header line
+    const XFOOBAR: &[u8] = b"X-Foobar";
+    let xfoobar_4k = XFOOBAR.repeat(4096/XFOOBAR.len());
+    
+    // header names 1b to 4096b
+    for p in 0..=12 {
+        let n = 1 << p;
+        let payload = [&xfoobar_4k[..n], b": b", RNRN].concat().leak();
+        _header(c, &format!("name_{}b", n), payload);
+    }
+    
+    // header values 1b to 4096b
+    for p in 0..=12 {
+        let n = 1 << p;
+        let payload = [b"a: ", &xfoobar_4k[..n], RNRN].concat().leak();
+        _header(c, &format!("value_{}b", n), payload);
+    }
+    
+    // 1 to 128
+    for p in 0..=7 {
+        let n = 1 << p;
+        _header(c, &format!("count_{}", n), [TINY_RN.repeat(n), RN.into()].concat().leak());
+    }
+}
+
+fn version(c: &mut Criterion) {
+    fn _version(c: &mut Criterion, name: &str, input: &'static [u8]) {
+        c.benchmark_group("version")
+        .throughput(Throughput::Bytes(input.len() as u64))
+        .bench_function(name, |b| b.iter(|| {
+            black_box({
+                let mut b = httparse::_benchable::Bytes::new(input);
+                httparse::_benchable::parse_version(&mut b).unwrap()
+            });
+        }));
+    }
+    
+    _version(c, "http10", b"HTTP/1.0\r\n");
+    _version(c, "http11", b"HTTP/1.1\r\n");
+    _version(c, "partial", b"HTTP/1.");
+}
+
+fn method(c: &mut Criterion) {
+    fn _method(c: &mut Criterion, name: &str, input: &'static [u8]) {
+        c.benchmark_group("method")
+        .throughput(Throughput::Bytes(input.len() as u64))
+        .bench_function(name, |b| b.iter(|| {
+            black_box({
+                let mut b = httparse::_benchable::Bytes::new(input);
+                httparse::_benchable::parse_method(&mut b).unwrap()
+            });
+        }));
+    }
+    
+    // Common methods should be fast-pathed
+    const COMMON_METHODS: &[&str] = &["GET", "HEAD", "POST", "PUT", "DELETE", "CONNECT", "OPTIONS", "TRACE", "PATCH"];
+    for method in COMMON_METHODS {
+        _method(c, &method.to_lowercase(), format!("{} / HTTP/1.1\r\n", method).into_bytes().leak());
+    }
+    // Custom methods should be infrequent and thus not worth optimizing
+    _method(c, "custom", b"CUSTOM / HTTP/1.1\r\n");
+}
+
+const WARMUP: Duration = Duration::from_millis(100);
+const MTIME: Duration = Duration::from_millis(100);
+const SAMPLES: usize = 200;
 criterion_group!{
     name = benches;
-    config = Criterion::default().sample_size(100).measurement_time(Duration::from_secs(10));
-    targets = req, req_short, resp, resp_short
+    config = Criterion::default().sample_size(SAMPLES).warm_up_time(WARMUP).measurement_time(MTIME);
+    targets = req, req_short, resp, resp_short, uri, header, version, method
 }
 criterion_main!(benches);
diff --git a/src/iter.rs b/src/iter.rs
@@ -2,11 +2,13 @@ use core::slice;
 use core::convert::TryInto;
 use core::convert::TryFrom;
 
+#[allow(missing_docs)]
 pub struct Bytes<'a> {
     slice: &'a [u8],
     pos: usize
 }
 
+#[allow(missing_docs)]
 impl<'a> Bytes<'a> {
     #[inline]
     pub fn new(slice: &'a [u8]) -> Bytes<'a> {
diff --git a/src/lib.rs b/src/lib.rs
@@ -30,6 +30,16 @@ mod iter;
 #[macro_use] mod macros;
 mod simd;
 
+#[doc(hidden)]
+// Expose some internal functions so we can bench them individually
+// WARNING: Exported for internal benchmarks, not fit for public consumption
+pub mod _benchable {
+    pub use super::parse_uri;
+    pub use super::parse_version;
+    pub use super::parse_method;
+    pub use super::iter::Bytes;
+}
+
 /// Determines if byte is a token char.
 ///
 /// > ```notrust
@@ -476,23 +486,7 @@ impl<'h, 'b> Request<'h, 'b> {
         let orig_len = buf.len();
         let mut bytes = Bytes::new(buf);
         complete!(skip_empty_lines(&mut bytes));
-        const GET: [u8; 4] = *b"GET ";
-        const POST: [u8; 4] = *b"POST";
-        let method = match bytes.peek_n::<[u8; 4]>(4) {
-            Some(GET) => {
-                unsafe {
-                    bytes.advance_and_commit(4);
-                }
-                "GET"
-            }
-            Some(POST) if bytes.peek_ahead(4) == Some(b' ') => {
-                unsafe {
-                    bytes.advance_and_commit(5);
-                }
-                "POST"
-            }
-            _ => complete!(parse_token(&mut bytes)),
-        };
+        let method = complete!(parse_method(&mut bytes));
         self.method = Some(method);
         if config.allow_multiple_spaces_in_request_line_delimiters {
             complete!(skip_spaces(&mut bytes));
@@ -743,7 +737,10 @@ impl<'a> fmt::Debug for Header<'a> {
 pub const EMPTY_HEADER: Header<'static> = Header { name: "", value: b"" };
 
 #[inline]
-fn parse_version(bytes: &mut Bytes<'_>) -> Result<u8> {
+#[doc(hidden)]
+#[allow(missing_docs)]
+// WARNING: Exported for internal benchmarks, not fit for public consumption
+pub fn parse_version(bytes: &mut Bytes) -> Result<u8> {
     if let Some(eight) = bytes.peek_n::<[u8; 8]>(8) {
         unsafe { bytes.advance(8); }
         return match &eight {
@@ -767,6 +764,30 @@ fn parse_version(bytes: &mut Bytes<'_>) -> Result<u8> {
     Ok(Status::Partial)
 }
 
+#[inline]
+#[doc(hidden)]
+#[allow(missing_docs)]
+// WARNING: Exported for internal benchmarks, not fit for public consumption
+pub fn parse_method<'a>(bytes: &mut Bytes<'a>) -> Result<&'a str> {
+    const GET: [u8; 4] = *b"GET ";
+    const POST: [u8; 4] = *b"POST";
+    match bytes.peek_n::<[u8; 4]>(4) {
+        Some(GET) => {
+            unsafe {
+                bytes.advance_and_commit(4);
+            }
+            Ok(Status::Complete("GET"))
+        }
+        Some(POST) if bytes.peek_ahead(4) == Some(b' ') => {
+            unsafe {
+                bytes.advance_and_commit(5);
+            }
+            Ok(Status::Complete("POST"))
+        }
+        _ => parse_token(bytes),
+    }
+}
+
 /// From [RFC 7230](https://tools.ietf.org/html/rfc7230):
 ///
 /// > ```notrust
@@ -838,7 +859,10 @@ fn parse_token<'a>(bytes: &mut Bytes<'a>) -> Result<&'a str> {
 }
 
 #[inline]
-fn parse_uri<'a>(bytes: &mut Bytes<'a>) -> Result<&'a str> {
+#[doc(hidden)]
+#[allow(missing_docs)]
+// WARNING: Exported for internal benchmarks, not fit for public consumption
+pub fn parse_uri<'a>(bytes: &mut Bytes<'a>) -> Result<&'a str> {
     let b = next!(bytes);
     if !is_uri_token(b) {
         // First char must be a URI char, it can't be a space which would indicate an empty path.
