[Date Prev][Date Next]
Re: better malloc strategies
Howard Chu wrote:
Kinda interesting - with hoard this shows us processing 23000 entries
per second in the single-threaded case, vs only 3521 per second with
It is possible that you are seeing the syndrome that I wrote about here:
AFAIK the poorly behaving code is still present in today's glibc malloc.
(the problem affects UMich-derived LDAP servers because entries in
the cache comprise a significant proportion of the heap traffic, and they
tend to get allocated by a different thread than frees them when the
cache fills up). malloc exhibits lock contention when blocks are freed
by a different thread than allocated them (more exactly when the threads
hash to different arenas).