"Confilict" -> "conflict" This increases conflicts by restricting the pool of random numbers to a limited number of bits. Actually, this approach replaces the "pool size" parameters I already have there, with slightly less flexibility though. Please look into the "random number pool size" to try to achieve the same effect.
Thanks, Mathieu * Lai Jiangshan ([email protected]) wrote: > Signed-off-by: Lai Jiangshan <[email protected]> > --- > tests/test_urcu_hash.c | 20 +++++++++++++++++++- > 1 files changed, 19 insertions(+), 1 deletions(-) > > diff --git a/tests/test_urcu_hash.c b/tests/test_urcu_hash.c > index 02d2c63..eb42522 100644 > --- a/tests/test_urcu_hash.c > +++ b/tests/test_urcu_hash.c > @@ -141,6 +141,7 @@ static unsigned long min_hash_alloc_size = > DEFAULT_MIN_ALLOC_SIZE; > static unsigned long init_populate; > static int opt_auto_resize; > static int add_only, add_unique, add_replace; > +static int opt_test_bucket_conflict; > > static int count_pipe[2]; > > @@ -390,8 +391,22 @@ int lookup_first(struct cds_lfht_node *node, void *key) > static inline > void lfht_test_node_init(struct lfht_test_node *node) > { > + unsigned long hash = test_rand_get(); > + > cds_lfht_node_init(&node->node); > - node->hash = test_rand_get(); > + > + if (opt_test_bucket_conflict) { > +#if (CAA_BITS_PER_LONG == 32) > + int hash_bits= test_rand_get_bits(5); > +#else > + int hash_bits= test_rand_get_bits(6); > +#endif > + > + if (hash_bits > 4) > + hash &= (1UL << hash_bits) - 1; > + } > + > + node->hash = hash; > } > > static > @@ -674,6 +689,7 @@ void show_usage(int argc, char **argv) > printf(" [-i] Add only (no removal).\n"); > printf(" [-k nr_nodes] Number of nodes to insert initially.\n"); > printf(" [-A] Automatically resize hash table.\n"); > + printf(" [-C] Use some hash value Confilict with buckets.\n"); > printf("\n\n"); > } > > @@ -792,6 +808,8 @@ int main(int argc, char **argv) > case 'A': > opt_auto_resize = 1; > break; > + case 'C': > + opt_test_bucket_conflict = 1; > break; > > } > -- > 1.7.4.4 > -- Mathieu Desnoyers Operating System Efficiency R&D Consultant EfficiOS Inc. http://www.efficios.com _______________________________________________ ltt-dev mailing list [email protected] http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
