Date: Sat, 11 Nov 1995 20:02:27 -0800 From: "Russell L. Carter" <rcarter@geli.com> To: hackers@freebsd.org Subject: New lmbench results, FreeBSD anomaly for TCP large transfers Message-ID: <199511120402.UAA03076@geli.clusternet>
next in thread | raw e-mail | index | archive | help
Finally got the (trivial) performance anomaly figured out for the TCP bandwith tests. As Larry suggested, setting the "SOCKOPT_*" to SOCKOPT_NONE fixed the atrocious results from bw_tcp.c. My systems' performance using this change will show up as *.clusternet in post 11/11/95 lmbench papers. FreeBSD's performance is ahem not bad. Here's the q, why is FreeBSD sensitive here? Thanks, Russell
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199511120402.UAA03076>