Reduction of hash conflicts on GpuPreAgg

It might invalidate hash-distribution if same values are given.
charliettxx · Mar 17, 2024 · db68e60 · db68e60
1 parent 01e685a
commit db68e60
Show file tree

Hide file tree

Showing 4 changed files with 15 additions and 3 deletions.
diff --git a/src/executor.c b/src/executor.c
@@ -650,8 +650,15 @@ pgstromBuildSessionInfo(pgstromTaskState *pts,
 			pp_info->kexp_groupby_keyload &&
 			pp_info->kexp_groupby_keycomp)
 		{
+			double	n_groups = pts->css.ss.ps.plan->plan_rows;
+
 			format = KDS_FORMAT_HASH;
-			hash_nslots = 20000; //to be estimated using num_groups
+			if (n_groups <= 5000.0)
+				hash_nslots = 20000;
+			else if (n_groups <= 4000000.0)
+				hash_nslots = 20000 + (int)(2.0 * n_groups);
+			else
+				hash_nslots = 8020000 + n_groups;
 			kds_length = (1UL << 30);			/* 1GB */
 		}
 		setup_kern_data_store(kds_temp, groupby_tdesc_final, kds_length, format);

diff --git a/src/gpu_join.c b/src/gpu_join.c
@@ -1747,7 +1747,7 @@ get_tuple_hashvalue(pgstromTaskInnerState *istate,
 		bool			isnull;
 
 		datum = ExecEvalExpr(es, econtext, &isnull);
-		hash ^= h_func(isnull, datum);
+		hash = pg_hash_merge(hash, h_func(isnull, datum));
 	}
 	hash ^= 0xffffffffU;
 

diff --git a/src/xpu_common.cu b/src/xpu_common.cu
@@ -1113,7 +1113,7 @@ pgfn_HashValue(XPU_PGFUNCTION_ARGS)
 		{
 			if (!expr_ops->xpu_datum_hash(kcxt, &__hash, datum))
 				return false;
-			hash ^= __hash;
+			hash = pg_hash_merge(hash, __hash);
 		}
 	}
 	hash ^= 0xffffffffU;

diff --git a/src/xpu_common.h b/src/xpu_common.h
@@ -2941,6 +2941,11 @@ EXTERN_FUNCTION(void)
 pg_kern_ereport(kern_context *kcxt);	/* only host code */
 EXTERN_FUNCTION(uint32_t)
 pg_hash_any(const void *ptr, int sz);
+INLINE_FUNCTION(uint32_t)
+pg_hash_merge(uint32_t hash_prev, uint32_t hash_next)
+{
+	return ((hash_prev >> 3) | (hash_prev << 29)) ^ hash_next;
+}
 
 /* ----------------------------------------------------------------
  *