I think it should be considered for Feisty, because it is really serious performance degradation. It's not hypotetical, this bug showed up to me while I evaluating Ubuntu 7.04 server on x86_64 for our production environment, and it happens in real program.
To demonstrate how bad it is look at results of small test program on Xeon 5160 with 32 GB ram: stock libc: 72 mallocs per second patched libc: 2251354 mallocs per second Program attached, it took number of seconds to run test as argument. ** Attachment added: "test_malloc.c" http://librarian.launchpad.net/7829424/test_malloc.c ** Description changed: Current malloc implementation in glibc-2.5 have serious problem with large number of free chunks, this bug reported to upstream http://sources.redhat.com/bugzilla/show_bug.cgi?id=4349 and fixed in current cvs and glibc-2.6 release. - This bug does not happens in edgy glibc-2.3.2. + This bug does not happens in Debian Sarge glibc-2.3.2. -- [feisty] serious performance regression in malloc https://bugs.launchpad.net/bugs/116621 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
