builds crds filters in parallel (#12360)

Based on run-time profiles, the majority time of new_pull_requests is
spent building bloom filters, in hashing and bit-vec ops.

This commit builds crds filters in parallel using rayon constructs. The
added benchmark shows ~5x speedup (4-core machine, 8 threads).

core/benches/crds_gossip_pull.rs

@@ -3,19 +3,19 @@
 extern crate test;
 
 use rand::{thread_rng, Rng};
-use solana_core::crds_gossip_pull::CrdsFilter;
-use solana_sdk::hash::{Hash, HASH_BYTES};
+use rayon::ThreadPoolBuilder;
+use solana_core::cluster_info::MAX_BLOOM_SIZE;
+use solana_core::crds::Crds;
+use solana_core::crds_gossip_pull::{CrdsFilter, CrdsGossipPull};
+use solana_core::crds_value::CrdsValue;
+use solana_sdk::hash::Hash;
 use test::Bencher;
 
 #[bench]
 fn bench_hash_as_u64(bencher: &mut Bencher) {
     let mut rng = thread_rng();
-    let hashes: Vec<_> = (0..1000)
-        .map(|_| {
-            let mut buf = [0u8; HASH_BYTES];
-            rng.fill(&mut buf);
-            Hash::new(&buf)
-        })
+    let hashes: Vec<_> = std::iter::repeat_with(|| Hash::new_rand(&mut rng))
+        .take(1000)
         .collect();
     bencher.iter(|| {
         hashes
@@ -24,3 +24,30 @@ fn bench_hash_as_u64(bencher: &mut Bencher) {
             .collect::<Vec<_>>()
     });
 }
+
+#[bench]
+fn bench_build_crds_filters(bencher: &mut Bencher) {
+    let thread_pool = ThreadPoolBuilder::new().build().unwrap();
+    let mut rng = thread_rng();
+    let mut crds_gossip_pull = CrdsGossipPull::default();
+    let mut crds = Crds::default();
+    for _ in 0..50_000 {
+        crds_gossip_pull
+            .purged_values
+            .push_back((Hash::new_rand(&mut rng), rng.gen()));
+    }
+    let mut num_inserts = 0;
+    for _ in 0..90_000 {
+        if crds
+            .insert(CrdsValue::new_rand(&mut rng), rng.gen())
+            .is_ok()
+        {
+            num_inserts += 1;
+        }
+    }
+    assert_eq!(num_inserts, 90_000);
+    bencher.iter(|| {
+        let filters = crds_gossip_pull.build_crds_filters(&thread_pool, &crds, MAX_BLOOM_SIZE);
+        assert_eq!(filters.len(), 128);
+    });
+}