In the case of Sqlite the shared resource is the file holding the
database. One way or another threads and processes have to serialize
their access. A single process with exclusive access is clearly the way
to achieve maximum throughput.
Databases spread across multiple spindle disk farms can have parallel
access, but they are heavy, not lite.
[EMAIL PROTECTED] wrote:
John Stanton <[EMAIL PROTECTED]> wrote:
When multi-processor machines are used thread usage may indeed implement
parallel processing when used appropriately and have tangible benefits.
True enough, in general. But with SQLite 3.5, access to each
database connection is serialized. So even though the interface
allows you to have 20 different threads all doing sqlite3_exec()
on the same connection at the same time, the SQL statements are
still being processed one by one, regardless of how many CPUs
you devote to the task.
--
D. Richard Hipp <[EMAIL PROTECTED]>
-----------------------------------------------------------------------------
To unsubscribe, send email to [EMAIL PROTECTED]
-----------------------------------------------------------------------------
-----------------------------------------------------------------------------
To unsubscribe, send email to [EMAIL PROTECTED]
-----------------------------------------------------------------------------
