Messages in this thread | | | Date | Wed, 18 Sep 2002 20:14:02 +0200 (CEST) | From | Ingo Molnar <> | Subject | Re: [patch] lockless, scalable get_pid(), for_each_process() elimination, 2.5.35-BK |
| |
On Wed, 18 Sep 2002, Cort Dougan wrote:
> You're talking about a different problem there. Creating a thread > within a realistic time-limit for a sensible number of threads is not a > bad idea. Doing it for a huge number of threads may not be something > that has to be done right away. Don't screw up the low to middle-end - > that trend is getting frightening in Linux.
sorry, but you must have missed the patch i posted yesterday. It can be done, it's fast, it does not hurt *any* benchmark and it actually has a worst-case cache-cold latency of 10 microseconds, even if 1 million threads are started up already.
and besides it significantly speeds up some other areas as well, like session management in shells, group-signal delivery performance, tty handling and more.
Ingo
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |