cache available_parallelism

Author: conradludgate

src/raw/utils/counter.rs

@@ -1,4 +1,7 @@
-use std::sync::atomic::{AtomicIsize, Ordering};
+use std::sync::{
+    atomic::{AtomicIsize, Ordering},
+    OnceLock,
+};
 
 use super::CachePadded;
 
@@ -14,9 +17,13 @@ pub struct Counter(Box<[CachePadded<AtomicIsize>]>);
 impl Default for Counter {
     /// Create a new `Counter`.
     fn default() -> Counter {
-        let num_cpus = std::thread::available_parallelism()
-            .map(usize::from)
-            .unwrap_or(1);
+        // available_parallelism is quite slow (microseconds).
+        static CPUS: OnceLock<usize> = OnceLock::new();
+        let num_cpus = *CPUS.get_or_init(|| {
+            std::thread::available_parallelism()
+                .map(Into::into)
+                .unwrap_or(1)
+        });
 
         // Round up to the next power-of-two for fast modulo.
         let shards = (0..num_cpus.next_power_of_two())