On Mon, 11 Dec 2000 technews@egsx.com wrote: > For example I found this site http://www.psc.edu/networking/perf_tune.html > that discusses in detail how to do that. My question is this. Does > changing the size of the buffer in > > /proc/sys/net/core/rmem_default - default receive window > /proc/sys/net/core/rmem_max - maximum receive window > /proc/sys/net/core/wmem_default - default send window > /proc/sys/net/core/wmem_max - maximum send window > > improve the performance in Linux? have anyone done this on their It does over high-latency links (relatively). Let's say you have 100mbit clean pipe to something that is 15 ms away. You will not be able to use those 100mbit in one TCP connection with the default settings. However, I did this on a freshly installed redhat 7.0 machine with the redhat kernel and it just rebooted twice after taking some traffic. Removing the changes I did to the above files made it perfectly stable again). I did these changes earlier on a standard 2.2.16 kernel and it worked just perfectly. I managed to reach 50mbit over the public internet to a machine 400 miles away. This was on a 15 megabyte file. Quite impressing considering that the lowest denomiator was a 100mbit fddi public exchange point. -- Mikael Abrahamsson email: swmike@swm.pp.se - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org
