<br><br><div><span class="gmail_quote">On 12/11/06, <b class="gmail_sendername">Toby Douglass</b> &lt;<a href="mailto:toby.douglass@summerblue.net">toby.douglass@summerblue.net</a>&gt; wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br>Ah.&nbsp;&nbsp;Go look at socket_iocp_add_socket().&nbsp;&nbsp;You'll find the very first<br>call to socket_internal_iocp_read() occurs at the end of *that*<br>function, which gets the read ball rolling.</blockquote><div><br>Yeah, you're right! I&nbsp; didn't&nbsp; pay attention to the prior&nbsp; socket_internal_iocp_read. 
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Ye-e-e-s-s...and no.&nbsp;&nbsp;I think I understand you to mean that we could,<br>for example, arrange it so we have 2xCPU threads running GQCS and when
<br>one of those threads returns with a complete, we trigger or signal the<br>main event loop thread?&nbsp;&nbsp;the problem with that surely is going to be the<br>large amount of thread switching.</blockquote><div><br>You may check out this paper:
<a class="l" href="www.scs.cs.nyu.edu/%7Edm/papers/zeldovich:async-mp.pdf" target="_blank" onmousedown="return rwt(this,'','','res','1','__EwS7tSi_KvwWWV4fv8g_FNvxJDE=','&amp;sig2=KW4P_vZnd6E9m7gsLWWsQg')">Multiprocessor Support for Event-Driven Programs.
</a><br><a class="l" href="www.scs.cs.nyu.edu/%7Edm/papers/zeldovich:async-mp.pdf" target="_blank" onmousedown="return rwt(this,'','','res','1','__EwS7tSi_KvwWWV4fv8g_FNvxJDE=','&amp;sig2=KW4P_vZnd6E9m7gsLWWsQg')"> </a><span class="m">
</span></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">_______________________________________________<br>Libevent-users mailing list<br>
<a href="mailto:Libevent-users@monkey.org">Libevent-users@monkey.org</a><br><a href="http://monkey.org/mailman/listinfo/libevent-users">http://monkey.org/mailman/listinfo/libevent-users</a><br></blockquote></div><br><br clear="all">
<br>-- <br>best regards,<br>hanzhu