HashTable: Fix bug HashTable htable set to empty_slot

htable needs to set to empty_slot or in `Insert()` it will set probing to continue which results in additional probe at slot=0
sethtroisi · Nov 16, 2022 · e533780 · e533780
1 parent a7d08e1
commit e533780
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 40 deletions.
diff --git a/core/HashTable.cpp b/core/HashTable.cpp
@@ -25,39 +25,6 @@ HashTable::HashTable(uint32_t elements)
       elements = HASH_MINIMUM_ELTS;
 
    for (hsize = 1<<HASH_MINIMUM_SHIFT; hsize < elements/HASH_MAX_DENSITY; )
-      hsize <<= 1;
-
-   hsize_minus1 = hsize - 1;
-
-   htable = (uint16_t *) xmalloc(hsize*sizeof(uint16_t));
-   olist = (uint16_t *) xmalloc(elements*sizeof(uint16_t));
-
-   // The j values are all in the range 0 <= j < M, so we can use M as an
-   // empty slot marker as long as we fill BJ[M] with a value that will never
-   // match a real b^j value. Since b^j is always in the range 0 <= b^j < p
-   // for some prime p, any value larger than all 32/64 bit primes will do.
-   empty_slot = elements;
-   BJ64 = (uint64_t *) xmalloc((elements+1)*sizeof(uint64_t));
-   BJ64[empty_slot] = UINT64_MAX;
-}
-
-HashTable::HashTable(uint32_t elements, uint32_t bestQ, uint32_t powerResidueLcm)
-{
-   uint32_t bjLen;
-
-   assert(elements <= HASH_MAX_ELTS);
-
-   if (elements < HASH_MINIMUM_ELTS)
-      elements = HASH_MINIMUM_ELTS;
-
-   for (hsize = 1<<HASH_MINIMUM_SHIFT; hsize*HASH_MAX_DENSITY < elements; )
-      hsize *= 2;
-
-   // Double size of hashtable if it will fit into L1 cache
-   if (hsize*sizeof(uint16_t) < L1_CACHE_SIZE*1024 && hsize <= elements/HASH_MIN_DENSITY/2)
-      hsize *= 2;
-
-   if (hsize*sizeof(uint16_t)*2 < L1_CACHE_SIZE*1024 && hsize <= elements/HASH_MIN_DENSITY/2)
       hsize *= 2;
 
    // Don't exceed maximum hash
@@ -69,15 +36,14 @@ HashTable::HashTable(uint32_t elements, uint32_t bestQ, uint32_t powerResidueLcm
    htable = (uint16_t *) xmalloc(hsize*sizeof(uint16_t));
    olist = (uint16_t *) xmalloc(elements*sizeof(uint16_t));
 
-   bjLen = bestQ+powerResidueLcm+9;
-
    // The j values are all in the range 0 <= j < M, so we can use M as an
    // empty slot marker as long as we fill BJ[M] with a value that will never
    // match a real b^j value. Since b^j is always in the range 0 <= b^j < p
    // for some prime p, any value larger than all 32/64 bit primes will do.
-   empty_slot = elements+1;
-   BJ64 = (uint64_t *) xmalloc(bjLen*sizeof(uint64_t));
-   BJ64[empty_slot] = UINT64_MAX;
+   empty_slot = elements;
+   BJ64 = (uint64_t *) xmalloc((elements+1)*sizeof(uint64_t));
+   // Point all slots at empty_slot
+   Clear();
 }
 
 HashTable::~HashTable(void)

diff --git a/core/HashTable.h b/core/HashTable.h
@@ -19,15 +19,18 @@ class HashTable
 {
 public:
    HashTable(uint32_t elements);
-   HashTable(uint32_t babySteps, uint32_t bestQ, uint32_t powerResidueLcm);
 
    ~HashTable(void);
 
    void  Cleanup(void);
 
    inline void Clear(void)
    {
-      memset(htable, 0, hsize*sizeof(uint16_t));
+      for(uint32_t i=0; i < hsize; i++) {
+          htable[i] = empty_slot;
+      }
+      // will never match a Lookup value
+      BJ64[empty_slot] = UINT64_MAX;
    }
 
    inline uint64_t get(uint32_t x) {return BJ64[x]; };
@@ -55,6 +58,8 @@ class HashTable
       slot = bj & hsize_minus1;
       elt = htable[slot];
 
+      // Could check elt == empty_slot to avoid checking BJ64
+
       if (BJ64[elt & HASH_MASK2] == bj)
          return elt & HASH_MASK2;