use IntMap

Cargo.lock

@@ -1087,6 +1087,12 @@ dependencies = [
  "libc",
 ]
 
+[[package]]
+name = "nohash-hasher"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2bf50223579dc7cdcfb3bfcacf7069ff68243f8c363f62ffa99cf000a6b9c451"
+
 [[package]]
 name = "nom"
 version = "7.1.1"
@@ -1826,6 +1832,7 @@ dependencies = [
  "axum",
  "clap 4.0.22",
  "futures",
+ "nohash-hasher",
  "parking_lot",
  "serde",
  "serde_json",

router/Cargo.toml

@@ -17,6 +17,7 @@ axum = { version = "0.5.16", features = ["json", "serde_json"] }
 text-generation-client = { path = "client" }
 clap = { version = "4.0.15", features = ["derive", "env"] }
 futures = "0.3.24"
+nohash-hasher = "0.2.0"
 parking_lot = "0.12.1"
 serde = "1.0.145"
 serde_json = "1.0.85"

router/src/batcher.rs

@@ -1,15 +1,14 @@
-use std::collections::HashMap;
 /// Batching and inference logic
 use crate::{Db, Entry};
 use crate::{ErrorResponse, GenerateRequest};
 use axum::http::StatusCode;
 use axum::Json;
+use nohash_hasher::IntMap;
 use std::future::Future;
 use std::sync::Arc;
 use text_generation_client::{Batch, ClientError, GeneratedText, ShardedClient};
 use thiserror::Error;
 use tokio::sync::{oneshot, Notify};
-use tokio::sync::{Semaphore, TryAcquireError};
 use tokio::time::Instant;
 use tracing::instrument;
 
@@ -26,8 +25,6 @@ pub struct Batcher {
 struct Shared {
     /// Batching background Tokio task notifier
     batching_task: Notify,
-    /// Inference request limit
-    limit_concurrent_requests: Semaphore,
 }
 
 impl Batcher {
@@ -35,13 +32,11 @@ impl Batcher {
         client: ShardedClient,
         max_batch_size: usize,
         max_waiting_tokens: usize,
-        max_concurrent_requests: usize,
     ) -> Self {
         // Batcher shared state
         let db = Db::new();
         let shared = Arc::new(Shared {
             batching_task: Notify::new(),
-            limit_concurrent_requests: Semaphore::new(max_concurrent_requests),
         });
 
         // Spawn batching background task that contains all the inference logic
@@ -62,9 +57,6 @@ impl Batcher {
         input_length: usize,
         request: GenerateRequest,
     ) -> Result<InferResponse, InferError> {
-        // Limit concurrent requests by acquiring a permit from the semaphore
-        let _permit = self.shared.limit_concurrent_requests.try_acquire()?;
-
         // One shot channel to communicate with the background batching task
         let (response_tx, response_rx) = oneshot::channel();
 
@@ -151,8 +143,7 @@ async fn batching_task(
                     }
                 }
 
-                cached_batch =
-                    wrap_future(client.generate_with_cache(batches), &mut entries).await;
+                cached_batch = wrap_future(client.generate_with_cache(batches), &mut entries).await;
                 waiting_tokens += 1;
             }
         }
@@ -162,7 +153,7 @@ async fn batching_task(
 /// Wrap a future inside a match statement to handle errors and send the response to the Batcher
 async fn wrap_future(
     future: impl Future<Output = Result<(Vec<GeneratedText>, Option<Batch>), ClientError>>,
-    entries: &mut HashMap<u64, Entry>,
+    entries: &mut IntMap<u64, Entry>,
 ) -> Option<Batch> {
     match future.await {
         Ok((generated_texts, next_batch)) => {
@@ -178,7 +169,7 @@ async fn wrap_future(
 }
 
 /// Send errors to the Batcher for all `entries`
-fn send_error(error: ClientError, entries: &mut HashMap<u64, Entry>) {
+fn send_error(error: ClientError, entries: &mut IntMap<u64, Entry>) {
     entries.drain().for_each(|(_, entry)| {
         // unwrap_or is valid here as we don't care if the receiver is gone.
         entry.response_tx.send(Err(error.clone())).unwrap_or(());
@@ -186,7 +177,7 @@ fn send_error(error: ClientError, entries: &mut HashMap<u64, Entry>) {
 }
 
 /// Send `generated_text` to the Batcher for all `finished`
-fn send_generated(finished: Vec<GeneratedText>, entries: &mut HashMap<u64, Entry>) {
+fn send_generated(finished: Vec<GeneratedText>, entries: &mut IntMap<u64, Entry>) {
     finished.into_iter().for_each(|output| {
         // We can `expect` here as the request id should always be in the entries
         let entry = entries
@@ -226,30 +217,18 @@ pub(crate) struct InferResponse {
 pub enum InferError {
     #[error("Request failed during generation: {0}")]
     GenerationError(String),
-    #[error("Model is overloaded")]
-    Overloaded,
-}
-
-/// Convert semaphore error
-impl From<TryAcquireError> for InferError {
-    fn from(_: TryAcquireError) -> Self {
-        InferError::Overloaded
-    }
 }
 
 /// Convert to Axum supported format
 impl From<InferError> for (StatusCode, Json<ErrorResponse>) {
     fn from(err: InferError) -> Self {
-        let status_code = match err {
-            InferError::GenerationError(_) => StatusCode::FAILED_DEPENDENCY,
-            InferError::Overloaded => StatusCode::TOO_MANY_REQUESTS,
-        };
-
-        (
-            status_code,
-            Json(ErrorResponse {
-                error: err.to_string(),
-            }),
-        )
+        match err {
+            InferError::GenerationError(_) => (
+                StatusCode::FAILED_DEPENDENCY,
+                Json(ErrorResponse {
+                    error: err.to_string(),
+                }),
+            ),
+        }
     }
 }

router/src/db.rs

@@ -1,8 +1,9 @@
-use crate::InferResponse;
 /// This code is massively inspired by Tokio mini-redis
+use crate::InferResponse;
 use crate::{GenerateParameters, GenerateRequest};
+use nohash_hasher::{BuildNoHashHasher, IntMap};
 use parking_lot::Mutex;
-use std::collections::{BTreeMap, HashMap};
+use std::collections::BTreeMap;
 use std::sync::Arc;
 use text_generation_client::{
     Batch, ClientError, NextTokenChooserParameters, Request, StoppingCriteriaParameters,
@@ -117,7 +118,7 @@ impl Db {
         &self,
         min_size: Option<usize>,
         max_size: usize,
-    ) -> Option<(HashMap<u64, Entry>, Batch)> {
+    ) -> Option<(IntMap<u64, Entry>, Batch)> {
         // Acquire lock
         let mut state = self.shared.state.lock();
 
@@ -132,13 +133,13 @@ impl Db {
             // Batch size
             let size = requests.len();
 
-            let mut entries = HashMap::with_capacity(size);
+            let mut entries = IntMap::with_capacity_and_hasher(size, BuildNoHashHasher::default());
             ids.iter().for_each(|id| {
                 // Remove entry from db
                 let mut entry = state.entries.remove(id).unwrap();
                 // Set batch_time
                 entry.batch_time = Some(Instant::now());
-                // Insert in entries hashmap
+                // Insert in entries IntMap
                 entries.insert(*id, entry);
             });
 

router/src/server.rs

@@ -7,9 +7,11 @@ use axum::response::IntoResponse;
 use axum::routing::{get, post};
 use axum::{Json, Router};
 use std::net::SocketAddr;
+use std::sync::Arc;
 use text_generation_client::ShardedClient;
 use tokenizers::Tokenizer;
 use tokio::signal;
+use tokio::sync::Semaphore;
 use tokio::time::Instant;
 use tracing::instrument;
 
@@ -18,6 +20,7 @@ use tracing::instrument;
 struct ServerState {
     validation: Validation,
     batcher: Batcher,
+    limit_concurrent_requests: Arc<Semaphore>,
 }
 
 /// Health check method
@@ -27,6 +30,16 @@ async fn health(state: Extension<ServerState>) -> Result<(), (StatusCode, Json<E
     //       be a bit too slow for a health check.
     //       What we should do instead if check if the gRPC channels are still healthy.
 
+    // Limit concurrent requests by acquiring a permit from the semaphore
+    let _permit = state.limit_concurrent_requests.try_acquire().map_err(|_| {
+        (
+            StatusCode::TOO_MANY_REQUESTS,
+            Json(ErrorResponse {
+                error: "Model is overloaded".to_string(),
+            }),
+        )
+    })?;
+
     // Send a small inference request
     state
         .batcher
@@ -65,6 +78,16 @@ async fn generate(
     req: Json<GenerateRequest>,
 ) -> Result<impl IntoResponse, (StatusCode, Json<ErrorResponse>)> {
     let start_time = Instant::now();
+    // Limit concurrent requests by acquiring a permit from the semaphore
+    let _permit = state.limit_concurrent_requests.try_acquire().map_err(|_| {
+        tracing::error!("Model is overloaded");
+        (
+            StatusCode::TOO_MANY_REQUESTS,
+            Json(ErrorResponse {
+                error: "Model is overloaded".to_string(),
+            }),
+        )
+    })?;
 
     // Validate request
     let details = req.0.parameters.details;
@@ -162,16 +185,12 @@ pub async fn run(
     addr: SocketAddr,
 ) {
     // Create state
-    let batcher = Batcher::new(
-        client,
-        max_batch_size,
-        max_waiting_tokens,
-        max_concurrent_requests,
-    );
+    let batcher = Batcher::new(client, max_batch_size, max_waiting_tokens);
     let validation = Validation::new(validation_workers, tokenizer, max_input_length);
     let shared_state = ServerState {
         validation,
         batcher,
+        limit_concurrent_requests: Arc::new(Semaphore::new(max_concurrent_requests)),
     };
 
     // Create router