On 6/28/2010 4:00 AM, zeromq-dev-requ...@lists.zeromq.org wrote:
As a quick test, try dumping libzmq.dll into the same directory as your
Python executable and see if that works. I bet it will.
Generally you'll need to add the location of libzmq.dll to your PATH. The
reason this will work
Hi,
The current load balancing algorithm used by the XREQ socket is
defined in lb.cpp. Requests are simply round robin'ed through the
active pipes:
current = (current + 1) % active
The problem with this approach is that it has a potential severe
bottleneck. Let's say every N request takes a
On Mon, 28 Jun 2010 11:08:13 -0500, Brian Granger wrote:
The current load balancing algorithm used by the XREQ socket is defined in
lb.cpp. Requests are simply round robin'ed through the active pipes:
[snip]
I am wondering if we can do better than this. My idea is to pick a client
at
On Mon, Jun 28, 2010 at 2:20 PM, Holger Hoffstaette
holger.hoffstae...@googlemail.com wrote:
On Mon, 28 Jun 2010 11:08:13 -0500, Brian Granger wrote:
The current load balancing algorithm used by the XREQ socket is defined in
lb.cpp. Â Requests are simply round robin'ed through the active pipes:
On Mon, Jun 28, 2010 at 3:46 PM, Peter Alexander vel.ac...@gmail.com wrote:
On Mon, Jun 28, 2010 at 2:20 PM, Holger Hoffstaette
holger.hoffstae...@googlemail.com wrote:
On Mon, 28 Jun 2010 11:08:13 -0500, Brian Granger wrote:
The current load balancing algorithm used by the XREQ socket is
I'm using latest zeromq from git (zeromq2-ba91644...) compiled with
Visual C++ 2008 SP1, tested on Vista 32-bit.
Please consider adding NOMINMAX to preprocessor definitions in
libzmq.vcproj. Without NOMINMAX, msvc fails to compile msg_store.cpp
because Windef.h defines min.
Also, here's a C++
Salvor Hardin wrote:
I'm using latest zeromq from git (zeromq2-ba91644...) compiled with
Visual C++ 2008 SP1, tested on Vista 32-bit.
Please consider adding NOMINMAX to preprocessor definitions in
libzmq.vcproj. Without NOMINMAX, msvc fails to compile msg_store.cpp
because Windef.h