perf: Replace hash usage with gxhash (#1028)

XAMPPRocky · web-flow · commit ad5cf879a121 · 2024-10-10T11:32:50.000+02:00
* perf: Replace hash usage with gxhash

* Update ci.yml

* Update cloudbuild.yaml

* Update Dockerfile
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -29,6 +29,10 @@ concurrency:
   group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
   cancel-in-progress: true
 
+env:
+  CARGO_TERM_COLOR: always
+  RUSTFLAGS: "-C target-feature=+aes,+vaes,+avx2"
+
 jobs:
   lint:
     name: Lint
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
@@ -149,6 +149,7 @@ strum_macros = "0.26"
 cfg-if = "1.0.0"
 libflate = "2.0.0"
 form_urlencoded = "1.2.1"
+gxhash = "3.4.1"
 
 [dependencies.hyper-util]
 version = "0.1"
@@ -236,4 +237,4 @@ schemars = { version = "0.8.15", features = ["bytes", "url"] }
 url = { version = "2.4.1", features = ["serde"] }
 
 [workspace.lints.clippy]
-undocumented_unsafe_blocks = "deny"
+undocumented_unsafe_blocks = "deny"
diff --git a/build/build-image/Dockerfile b/build/build-image/Dockerfile
@@ -19,6 +19,8 @@ ARG RUST_TOOLCHAIN
 ENV RUSTUP_HOME=/usr/local/rustup \
     CARGO_HOME=/usr/local/cargo \
     PATH=/usr/local/cargo/bin:$PATH \
+    CARGO_TERM_COLOR=always \
+    RUSTFLAGS="-C target-feature=+aes,+vaes,+avx2" \
     LC_ALL=C.UTF-8 \
     LANG=C.UTF-8
 
diff --git a/cloudbuild.yaml b/cloudbuild.yaml
@@ -103,6 +103,8 @@ options:
     - "CARGO_HOME=/workspace/.cargo"
     - "REPOSITORY=${_REPOSITORY}"
     - "BUILD_IMAGE_TAG=${_BUILD_IMAGE_TAG}"
+    - "CARGO_TERM_COLOR=always"
+    - 'RUSTFLAGS="-C target-feature=+aes,+vaes,+avx2"'
   machineType: E2_HIGHCPU_32
   dynamic_substitutions: true
 timeout: 7200s
diff --git a/src/components/proxy/error.rs b/src/components/proxy/error.rs
@@ -105,17 +105,7 @@ impl Hash for PipelineError {
     }
 }
 
-pub struct SeahashBuilder;
-
-impl std::hash::BuildHasher for SeahashBuilder {
-    type Hasher = seahash::SeaHasher;
-
-    fn build_hasher(&self) -> Self::Hasher {
-        seahash::SeaHasher::new()
-    }
-}
-
-pub type ErrorMap = std::collections::HashMap<PipelineError, u64, SeahashBuilder>;
+pub type ErrorMap = gxhash::HashMap<PipelineError, u64>;
 
 pub type ErrorSender = tokio::sync::mpsc::Sender<ErrorMap>;
 //pub type ErrorReceiver = tokio::sync::mpsc::Receiver<ErrorMap>;
@@ -140,7 +130,7 @@ pub struct ErrorAccumulator {
 impl ErrorAccumulator {
     pub fn new(tx: ErrorSender) -> Self {
         Self {
-            map: ErrorMap::with_hasher(SeahashBuilder),
+            map: <_>::default(),
             tx,
             oldest: Instant::now(),
         }
@@ -160,7 +150,7 @@ impl ErrorAccumulator {
         };
 
         #[allow(clippy::mutable_key_type)]
-        let map = std::mem::replace(&mut self.map, ErrorMap::with_hasher(SeahashBuilder));
+        let map = std::mem::take(&mut self.map);
         permit.send(map);
         true
     }
diff --git a/src/components/proxy/packet_router.rs b/src/components/proxy/packet_router.rs
@@ -169,7 +169,7 @@ pub async fn spawn_receivers(
         let mut log_task = tokio::time::interval(std::time::Duration::from_secs(5));
 
         #[allow(clippy::mutable_key_type)]
-        let mut pipeline_errors = super::error::ErrorMap::with_hasher(super::error::SeahashBuilder);
+        let mut pipeline_errors = super::error::ErrorMap::default();
 
         #[allow(clippy::mutable_key_type)]
         fn report(errors: &mut super::error::ErrorMap) {
diff --git a/src/config.rs b/src/config.rs
@@ -234,7 +234,7 @@ impl Config {
                         crate::net::cluster::proto::FilterChain::try_from(&*self.filters.load())?,
                     );
                     let any = resource.try_encode()?;
-                    let version = seahash::hash(&any.value);
+                    let version = gxhash::gxhash64(&any.value, 0xdeadbeef);
 
                     let vstr = version.to_string();
 
diff --git a/src/filters/read.rs b/src/filters/read.rs
@@ -50,7 +50,7 @@ impl ReadContext {
             destinations: Vec::new(),
             source,
             contents,
-            metadata: DynamicMetadata::new(),
+            metadata: <_>::default(),
         }
     }
 }
diff --git a/src/filters/write.rs b/src/filters/write.rs
@@ -14,8 +14,6 @@
  * limitations under the License.
  */
 
-use std::collections::HashMap;
-
 use crate::{
     net::endpoint::{DynamicMetadata, EndpointAddress},
     pool::PoolBuffer,
@@ -45,7 +43,7 @@ impl WriteContext {
             source,
             dest,
             contents,
-            metadata: HashMap::new(),
+            metadata: <_>::default(),
         }
     }
 }
diff --git a/src/net/cluster.rs b/src/net/cluster.rs
@@ -15,7 +15,7 @@
  */
 
 use std::{
-    collections::{hash_map::RandomState, BTreeSet},
+    collections::BTreeSet,
     fmt,
     sync::atomic::{AtomicU64, AtomicUsize, Ordering::Relaxed},
 };
@@ -27,6 +27,7 @@ use serde::{Deserialize, Serialize};
 use crate::net::endpoint::{Endpoint, EndpointAddress, Locality};
 
 const SUBSYSTEM: &str = "cluster";
+const HASH_SEED: i64 = 0xdeadbeef;
 
 pub use crate::generated::quilkin::config::v1alpha1 as proto;
 
@@ -60,15 +61,15 @@ pub(crate) fn active_endpoints() -> &'static prometheus::IntGauge {
     &ACTIVE_ENDPOINTS
 }
 
-pub type TokenAddressMap = std::collections::BTreeMap<u64, BTreeSet<EndpointAddress>>;
+pub type TokenAddressMap = gxhash::HashMap<u64, gxhash::HashSet<EndpointAddress>>;
 
 #[derive(Copy, Clone)]
 pub struct Token(u64);
 
 impl Token {
     #[inline]
     pub fn new(token: &[u8]) -> Self {
-        Self(seahash::hash(token))
+        Self(gxhash::gxhash64(token, HASH_SEED))
     }
 }
 
@@ -123,7 +124,7 @@ impl EndpointSet {
     pub fn new(endpoints: BTreeSet<Endpoint>) -> Self {
         let mut this = Self {
             endpoints,
-            token_map: TokenAddressMap::new(),
+            token_map: <_>::default(),
             hash: 0,
             version: 0,
         };
@@ -141,7 +142,7 @@ impl EndpointSet {
     pub fn with_version(endpoints: BTreeSet<Endpoint>, hash: EndpointSetVersion) -> Self {
         let mut this = Self {
             endpoints,
-            token_map: TokenAddressMap::new(),
+            token_map: <_>::default(),
             hash: hash.number(),
             version: 1,
         };
@@ -177,14 +178,14 @@ impl EndpointSet {
     #[inline]
     pub fn update(&mut self) -> TokenAddressMap {
         use std::hash::{Hash, Hasher};
-        let mut hasher = seahash::SeaHasher::with_seeds(0, 1, 2, 3);
-        let mut token_map = TokenAddressMap::new();
+        let mut hasher = gxhash::GxHasher::with_seed(HASH_SEED);
+        let mut token_map = TokenAddressMap::default();
 
         for ep in &self.endpoints {
             ep.hash(&mut hasher);
 
             for tok in &ep.metadata.known.tokens {
-                let hash = seahash::hash(tok);
+                let hash = gxhash::gxhash64(tok, HASH_SEED);
                 token_map
                     .entry(hash)
                     .or_default()
@@ -200,12 +201,12 @@ impl EndpointSet {
     /// Creates a map of tokens -> address for the current set
     #[inline]
     pub fn build_token_map(&mut self) -> TokenAddressMap {
-        let mut token_map = TokenAddressMap::new();
+        let mut token_map = TokenAddressMap::default();
 
         // This is only called on proxies, so calculate a token map
         for ep in &self.endpoints {
             for tok in &ep.metadata.known.tokens {
-                let hash = seahash::hash(tok);
+                let hash = gxhash::gxhash64(tok, HASH_SEED);
                 token_map
                     .entry(hash)
                     .or_default()
@@ -257,7 +258,7 @@ impl EndpointSet {
 }
 
 /// Represents a full snapshot of all clusters.
-pub struct ClusterMap<S = RandomState> {
+pub struct ClusterMap<S = gxhash::GxBuildHasher> {
     map: DashMap<Option<Locality>, EndpointSet, S>,
     token_map: DashMap<u64, Vec<EndpointAddress>>,
     num_endpoints: AtomicUsize,
@@ -268,7 +269,7 @@ type DashMapRef<'inner, S> = dashmap::mapref::one::Ref<'inner, Option<Locality>,
 type DashMapRefMut<'inner, S> =
     dashmap::mapref::one::RefMut<'inner, Option<Locality>, EndpointSet, S>;
 
-impl ClusterMap<RandomState> {
+impl ClusterMap {
     pub fn new() -> Self {
         Self::default()
     }
diff --git a/src/net/endpoint/metadata.rs b/src/net/endpoint/metadata.rs
@@ -21,12 +21,10 @@ pub mod build {
     pub const GIT_COMMIT_HASH: Option<&str> = option_env!("GIT_COMMIT_HASH");
 }
 
-use std::{collections::HashMap, convert::TryFrom};
-
 pub use symbol::{Key, Reference, Symbol};
 
 /// Shared state between [`Filter`][crate::filters::Filter]s during processing for a single packet.
-pub type DynamicMetadata = HashMap<Key, Value>;
+pub type DynamicMetadata = gxhash::HashMap<Key, Value>;
 
 pub const KEY: &str = "quilkin.dev";
 

Original file line number	Diff line number	Diff line change
`@@ -50,7 +50,7 @@ impl ReadContext {`
`50`	`50`	`destinations: Vec::new(),`
`51`	`51`	`source,`
`52`	`52`	`contents,`
`53`		`- metadata: DynamicMetadata::new(),`
	`53`	`+ metadata: <_>::default(),`
`54`	`54`	`}`
`55`	`55`	`}`
`56`	`56`	`}`
Original file line number	Diff line number	Diff line change
`@@ -14,8 +14,6 @@`
`14`	`14`	`* limitations under the License.`
`15`	`15`	`*/`
`16`	`16`
`17`		`-use std::collections::HashMap;`
`18`		`-`
`19`	`17`	`use crate::{`
`20`	`18`	`net::endpoint::{DynamicMetadata, EndpointAddress},`
`21`	`19`	`pool::PoolBuffer,`
`@@ -45,7 +43,7 @@ impl WriteContext {`
`45`	`43`	`source,`
`46`	`44`	`dest,`
`47`	`45`	`contents,`
`48`		`- metadata: HashMap::new(),`
	`46`	`+ metadata: <_>::default(),`
`49`	`47`	`}`
`50`	`48`	`}`
`51`	`49`	`}`