| 00:33.25 | seanbright | coreyfarrell: still need to look into it more, but it's not about the number of threadpools |
| 00:33.50 | seanbright | 'threads' are constantly moving between 2/3 of those containers as tasks are executed |
| 00:34.18 | seanbright | and if the containers are only being manipulated by a single thread, then there is no reason for locking |
| 00:34.54 | seanbright | i could make the argument that they shouldn't be ao2_containers at all, but i know people have a low tolerance when it comes to my suggestions |
| 00:56.56 | snuff-work | seanbright: when u decide to make a gerrit for the master/trunk of that Ignore RTP we havent negotiated.. i can test |
| 01:22.54 | seanbright | k |
| 01:24.31 | coreyfarrell | seanbright: the number of threadpools matters for the memory overhead of the ao2 locks. |
| 02:18.39 | seanbright | yes |
| 02:20.04 | seanbright | memory isn't my prinary driver for considering removing the locks |
| 04:35.01 | *** join/#asterisk-dev snuff-work (~snuff-wor@202-161-112-134.tpgi.com.au) |
| 04:35.01 | *** mode/#asterisk-dev [+o snuff-work] by ChanServ |
| 04:45.45 | *** join/#asterisk-dev snuff-work (~snuff-wor@202-161-112-134.tpgi.com.au) |
| 04:45.45 | *** mode/#asterisk-dev [+o snuff-work] by ChanServ |