Network Fix #2The problem:http://elitistjerks.com/f15/t40955-wow_ ... ost1252436Each time there has been a patch (and some times very sporadically after a patch) my comp lags out when i try to log my main into dalaran. Its always the same, I log in, graphics load but everyone around me is "unknown the noble" etc. After logging in and getting disconnected (this takes from 10-100 secs) what seems like a set number of times (around 10-20), I can log in no problem.
The Fix:Next, go into your network settings in the system properties if you are playing under windows. Select the network card and select the "configure" button. You should get dialog box with some tabs on it. Select advanced. You will find several options. Ensure that specific features that might be called like "TCP offloading" is switched off. Also you might want to shut off powersafe functions (unless you are playing on a laptop they might not make much sense anyway). I can't give you specific names for the options to turn off, because the naming of those are driver dependant and thus change depending which hardware and driver you are using. In general, every option that sounds like the hardware is "accelerating" or "offloading" some functionality from the operating system into the hardware or driver should be switched off. These features have in the past proven to be more counter-productive than useful. Again a system reboot is necessary after changing something here.
Here's a 10 step screenshot to show you how to navigate to the correct options if it sounds too confusing (windows XP users skip #3, I think):
http://khornn.pp.fi/images/network.jpgWhy it workshttp://en.wikipedia.org/wiki/Tcp_checksum_offloadhttp://publib.boulder.ibm.com/infocente ... ffload.htmTCP checksum offload
The TCP checksum offload option enables the network adapter to compute the TCP checksum on transmit and receive, which saves the AIX® host CPU from having to compute the checksum.
The savings vary by packet size. Small packets have little or no savings with this option, while large packets have larger savings. On the PCI-X GigE adapters, the savings for MTU 1500 are typically about 5% reduction in CPU utilization, and for MTU 9000 (Jumbo Frames) the savings is approximately a 15% reduction in CPU utilization.
TCP streaming throughput with MTU 1500 is slower on machines that have processors faster than 400 MHz if the TCP checksum offload option is enabled because the host system can run the checksum faster than the Gigabit Ethernet PCI adapters, FC2969 and FC 2975. Therefore, by default, this option is off on these adapters. When these adapters use jumbo frames, it can run at wire speed even when it has to compute the checksum.
The PCI-X Gigabit Ethernet adapters can run at wire speeds with the TCP checksum offload option enabled and it reduces host CPU processing so it is enabled by default.
This should probably fix some of the random disconnect issues by taking a lot of workload off of your shitty network card and making your CPU work more. Again, if you experience any problems just reverse these steps to restore the original settings.