Code for gathering stats about extremely slow HashMap operations

Author: pczarn

Cargo.toml

@@ -36,4 +36,4 @@ repository  = "https://github.com/carllerche/hashmap2"
 homepage    = "https://github.com/carllerche/hashmap2"
 
 [dependencies]
-rand = "0.3.12"
+rand = "0.3.15"

examples/stats.rs

@@ -0,0 +1,26 @@
+extern crate hashmap2;
+extern crate rand;
+
+use hashmap2::HashMap;
+use rand::Rng;
+
+fn main() {
+    let mut map: HashMap<i32, ()> = HashMap::new();
+    assert_eq!(map.len(), 0);
+    let mut rng = rand::weak_rng();
+    let mut iter = rng.gen_iter();
+    let len = 2 << 20;
+    let usable_cap = (len as f32 * 0.833) as usize;
+    let mut stats = vec![];
+    for _ in 0..10000 {
+        while map.len() < usable_cap {
+            map.insert(iter.next().unwrap(), ());
+        }
+        map.stats(&mut stats);
+        map.clear();
+    }
+    for (i, (displacement, forward_shift)) in stats.into_iter().enumerate() {
+        println!("{}: {}\t{}", i, displacement, forward_shift);
+    }
+    println!("map len={:?} capacity={:?}", map.len(), map.capacity());
+}

src/adaptive_map.rs

@@ -22,7 +22,7 @@ use HashMap;
 
 // Beyond this displacement, we switch to safe hashing or grow the table.
 const DISPLACEMENT_THRESHOLD: usize = 128;
-const FORWARD_SHIFT_THRESHOLD: usize = 512;
+const FORWARD_SHIFT_THRESHOLD: usize = 1024;
 // When the map's load factor is below this threshold, we switch to safe hashing.
 // Otherwise, we grow the table.
 // const LOAD_FACTOR_THRESHOLD: f32 = 0.625;

src/lib.rs

@@ -1239,6 +1239,10 @@ impl<K, V, S> HashMap<K, V, S>
 
         self.search_mut(k).into_occupied_bucket().map(|bucket| pop_internal(bucket))
     }
+
+    pub fn stats(&self, stats: &mut Vec<(u64, u64)>) {
+        self.table.stats(stats);
+    }
 }
 
 // Not copying this requires specialization

src/table.rs

@@ -590,6 +590,39 @@ fn test_offset_calculation() {
 }
 
 impl<K, V> RawTable<K, V> {
+    pub fn stats(&self, stats: &mut Vec<(u64, u64)>) {
+        if stats.is_empty() {
+            *stats = vec![(0, 0); 2000];
+        }
+        // stats.clear();
+        // stats.extend(iter::repeat((0, 0)).take(1000));
+        let mut iter = self.raw_buckets();
+        let first_hash = iter.raw.hash;
+        let mut latest_hash = iter.raw.hash;
+        let mut chunk_info = vec![];
+        while let Some(raw) = iter.next() {
+            let num_empty = (raw.hash as usize - latest_hash as usize) / size_of::<u64>();
+            let idx = (raw.hash as usize - first_hash as usize) / size_of::<u64>();
+            stats[0].0 += num_empty as u64;
+            stats[0].1 += num_empty as u64;
+            if num_empty > 0 {
+                for n in chunk_info.drain(..) {
+                    stats[n as usize].1 += 1;
+                }
+            } else {
+                for n in chunk_info.iter_mut() {
+                    *n += 1;
+                }
+            }
+            chunk_info.push(0);
+            let ib = unsafe {
+                (*raw.hash) as usize & (self.capacity - 1)
+            };
+            let displacement = (idx as isize - ib as isize) as usize & (self.capacity - 1);
+            stats[displacement].0 += 1;
+            latest_hash = iter.raw.hash;
+        }
+    }
     /// Does not initialize the buckets. The caller should ensure they,
     /// at the very least, set every hash to EMPTY_BUCKET.
     unsafe fn new_uninitialized(capacity: usize) -> RawTable<K, V> {