Date: Wed, 12 Apr 1995 18:16:44 +0000 From: Matt Thomas <matt@lkg.dec.com> To: freebsd-hackers@FreeBSD.org Subject: Fast Ethernet (100baseT) on FreeBSD Message-ID: <199504121816.SAA06094@whydos.lkg.dec.com>
next in thread | raw e-mail | index | archive | help
For whom ever cares, I just sent off the latest DC21040/DC21140 driver for inclusion into the FreeBSD source pool. This still supports the DC21040 cards that it always has and support for the DC21140 (Fast Ethernet Controller) has been added. Unlike the DC21040, the DC21140 design doesn't allow a truly generic driver to be written. This mean each DC21140 card will probably need to have code written to detect and configure it (this is about ~50 lines of code). Currently only the DEC DE500-XA and the DEC DC21140 10/100 Evaluation Board are known to work. The (or a) SMC 10/100 card is supposed to work the same at the evaluation board but as of now it's still only a rumor. I know everyone will want to know about performance. Unfortunately my test system is a DECpc XL433; not exactly the fastest machine in the world. Connected to it (though a cross-over cat5 cable) is an AlphaStation 400 4/233; now there's a fast machine. Not the best test environment (but then not bad considering this is a hobby and not my job) but enough to do proof of concept. So the numbers might be at the low end of what you might see. Sending to the AlphaStation: fbsd# ./ttcp -t -s 3.3.3.3 ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001 tcp -> 3.3.3.3 ttcp-t: socket ttcp-t: connect ttcp-t: 16777216 bytes in 6.13 real seconds = 2673.76 KB/sec +++ ttcp-t: 2048 I/O calls, msec/call = 3.06, calls/sec = 334.22 ttcp-t: 0.0user 4.6sys 0:06real 75% 20i+316d 206maxrss 0+2pf 4+68csw Receiving from the AlphaStation: fbsd# ./ttcp -r -s ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001 tcp ttcp-r: socket ttcp-r: accept from 3.3.3.3 ttcp-r: 16777216 bytes in 4.32 real seconds = 3791.66 KB/sec +++ ttcp-r: 2054 I/O calls, msec/call = 2.15, calls/sec = 475.35 ttcp-r: 0.0user 1.8sys 0:04real 44% 33i+526d 208maxrss 0+2pf 8+46csw In both cases, vmstat reports no idle CPU. Larger tranfers (64MB, 256MB) are within %5 of throughput. As always, your mileage may vary. Void where prohibited. Blah Blah Blah. Matt Thomas Internet: matt@lkg.dec.com U*X Networking WWW URL: http://ftp.dec.com/%7Ethomas/ Digital Equipment Corporation Disclaimer: This message reflects my Littleton, MA own warped views, etc.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199504121816.SAA06094>