Michael McCormack mccormac@aals27.alcatel.com.au writes:
Reducing the number of server calls but making them more expensive is not necessarily a gain. Show me the numbers...
I'm not sure it's a good idea, particularly if we want to support signals (which I think would be a good thing since it allows getting rid of the service thread).
You'll probably still need a server call to find out about file locks at some point. And again you'll get in trouble if you want to support signals.
Hi Alexandre,
i have rewritten my patch for the latest wine release. The main changes from the previous are that it only checks overlapped operations in WaitForMultipleObjectswEx, and it now uses SetEvent to end the operation. (courtesy of your new servercall architecture). I have done a simple benchmark. The results of the benchmark on my Pii350 and an internal modem are as follows:
Wine 2001/03/05 vanilla + async-fix.diff: average write time for "AT" command: 675 usec average read time for response: 634 usec
Wine 2001/03/05 + client-async3.diff: average write time for "AT" command: 362 usec average read time for response: 322 usec
The times in themselves are pretty meaningless, but it is the same program, under the same conditions.
The patches and program i used are included.
Mike
On Tue, 6 Mar 2001, Alexandre Julliard wrote: