Dan Nelson wrote:
In the last episode (Mar 27), Mirar @ Pike developers forum said:
So if you are doing a lot of image operations, it will use several real threads for the operations.
Any place in C code that you see THREADS_ALLOW / THREADS_DISALLOW pairs, that's where the code has released the main pike interpreter lock, allowing other threads to run.
Yes, but AFAIK those refer to Pike-threads, which are still being run on a single core.
One of the few modules that actually makes use of multi-cores is the shuffler module in Pike. But that solves more of a latency problem than a CPU problem. The (current) best way to utilise multicores in combination with a database is by using stored procedures in SQL. The SQL-backend will probably use one core per open SQL session.