Date: Tue, 30 Jan 1996 19:54:27 +1100 From: Bruce Evans <bde@zeta.org.au> To: bde@zeta.org.au, gerg@stallion.oz.au Cc: hackers@freebsd.org Subject: Re: Multi-Port Async Cards Message-ID: <199601300854.TAA32122@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
>> For comparison, a Cyclades 8Yo ISA uses about 0.4% host CPU per kbyte >> output in opost mode (test done on FreeBSD-current, 486DX/33, 8MB RAM). >> Why is the Stallion so slow? :-). The Cyclades board "intelligent" but >Hmmm, comparing apples and oranges to some extent. As I pointed out this >bench mark result was done on a comercial UNIX flavor, I expect the above >number was done on a FreeBSD system? Yes, I said so above :-). >> [Cyclades, 486DX/33], -opost 0.3%/KB >> >> 16550, 486DX2/66, -opost 0.25% (one port) >>... >> Linux Comtrol Rocketport driver as reported in Linux Journal: >> >> 486DX/33, -opost 0.08% >> 486DX/33, opost 0.77% >Again, different OS and different number of ports. It does make a >difference... The only fair comparison is all boards similarly configured, >in the same system, running the same OS... It's important to compare with other OS's to make sure that FreeBSD is the best. My numbers indicate that the OS doesn't actually make much difference unless the driver or OS does something wrong. The 0.08% for the Comtrol can be predicted from the 0.25% for the 16550 by dividing by a bit less than 2 for the improvement from 8-bit i/o's to 16-bit i/o's and then reducing a bit more for the lower interrupt overhead. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199601300854.TAA32122>