Date: Tue, 6 Nov 2001 11:59:53 -0500 (EST) From: "Brandon D. Valentine" <bandix@looksharp.net> To: "Jeroen C. van Gelderen" <jeroen@vangelderen.org> Cc: Marcel Prisi <marcel-lists@virtua.ch>, <stable@FreeBSD.ORG> Subject: Re: What NIC to choose ? Message-ID: <20011106114853.C42904-100000@turtle.looksharp.net> In-Reply-To: <3BE80E27.3080707@vangelderen.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 6 Nov 2001, Jeroen C. van Gelderen wrote: >It would be interesting to know which clone chipset was giving >you trouble. It seems unfair to declare all clone chipsets to >be unreliable unless you have had a bad experience with each >one of them. (The converse it true also, that is why I indicated >the exact make of the tulip clone that I have not had trouble >with.) Indeed that would be unfair. Here is one particular model that I have found to suck. =) Ethernet controller: LiteOn LNE100TX (rev 32). I have both Kingston and Linksys branded examples of this crappy non-standard tulip clone here. They all report pretty much the above line in /proc/pci under linux. >Then there is the issue of the quality of the clone card itself >which -when improperly engineered- may cause failures that have >nothing to do with the chipset, no? That can be true, yes. But in this case I've seen it across multiple vendors' implementations. >My experience with the Linksys LNE100TX 4.1 (ADMtek chipset) >has been nothing but positive, despite the fact that it is >a tulip clone. Which were the exact types of cards that you >had fail? Exactly the same card. You may have had positive experiences with 'em, but I have seen too many keel over dead. >As for the Intel EtherExpress, my previous post was not so >positive. I noted reproducible timing-related errors when they >were depoyed in quality 2U riser cards. No other card (3Com, >LinkSys) had this problem. Using the Intel PRO 10/100 cards >without risers gave no problems but they still have a worse >price/performance ratio for my particular setup. Interestingly enough, the beowulf cluster they were used in contains 2U cases w/ risers. The risers are pretty crappy and they have caused a lot of problems in their own right. However, those seem to have been resolved for a while now and NICs just keep keeling over. All further system purchases are 1Us with onboard NICs and much more reliable 64-bit PCI risers for Myrinet. The Intel eepros in that cluster's infrastructure nodes never have a problem though, and they pass more data than any interface in the cluster. -- "Never put off until tomorrow what you can do today. There might be a law against it by that time." -- /usr/games/fortune, 07/30/2001 Brandon D. Valentine <bandix at looksharp.net> To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011106114853.C42904-100000>