Date: Sat, 31 Jan 2004 00:09:27 +0800 From: Raymond <[EMAIL PROTECTED]> Subject: Re: [LIB] Networking problems
Firstly, firewire cards (and, to a lesser extent, USB2.0 devices) have onboard controllers.
That explains a whole lot right there Raymond. Thanks for the run down. I don't suppose two PCs can be connected together via Firewire or USB2 to transfer files, can they? Where there's a will, there's probably software.
There are USB (and I presume USB2.0) PC to PC links but AFAIK they use a lot of CPU time anyway (USB/USB2 controllers offload a lot more processing to the CPU than firewire ones do).
So between the right hardware, and the right drivers as Raymond, higher transfer speeds on LANs >are< possible. But..... aaaccckkk... this reminds me of the headaches I had trying to find a modem for a typical modern desktop a few months that would connect and transfer data consisitantly at 4+KBPS. Don't know why, but 4 modems from Fry's in CA wouldn't connect/transfer faster than 3.3KBPS. Puttoing the generic modem I had in a system into the problem PC saw 4+KBPS connections/tranfers on ebvery connection. And I found few people who could explain much.
The signal processing chips in the modems make a huge difference. Some are able to filter out certain types of line noise better than others and filtering out more noise, loosely speaking (signal processing gurus don't jump on me here!) means they can transfer at a higher rate through different modulation techniques.
NIC brands do make a big difference, as some take 60-80 % CPU time, especially in 100Mbps mode, while other brands do a lot by themselves and leave more CPU time for other processes. Cheap NIC cards rarely perform as well as e.g. 3Com, Xircom, .... It is really a matter of value for money here.
So are there NIC cards with onboard controllers like the Firewire amd USB2 cards Raymond mentioned?
All NICs have onboard controllers. However, even the best ones really only give the CPU network packets (anything at a hardware or data link level in the OSI model for those who have done network theory). The CPU still needs to do a lot of checking, error correction, retransmission requests, data translations and so-on before it goes anywhere near the disk. In contrast, with my Firewire example, the data is already almost in the right format, all the CPU needs to do is shove it through the disk (be it through PIO or DMA) ... so I guess the comparison between NIC transfers and firewire isn't really all that fair. Case in point, my friend has a Dell laptop (recent model P4) which (AFAIK) does DMA off the hard drive. Yet he can load programs faster off a 7200RPM ATA133 hard drive in an external firewire case than off the laptop's internal hard drive (and disk-to-RAM speed tests confirm this)!
- Raymond
---
/~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\
| | "Does fuzzy logic tickle?" |
| ___ | "My HDD has no reverse. How do I backup?" |
| /__/ +-------------------------------------------|
| / \ a y b o t | [EMAIL PROTECTED] |
| | Need help? Visit #Windows98 on DALNet! |
| ICQ: 31756092 | Libretto IRC channel #Libretto on DALNet! |
\~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~/
************************************************************** http://libretto.basiclink.com - Libretto mailing list http://www.silverace.com/libretto/ - Archives
-------TO UNSUBSCRIBE-------
Reply to any of the list messages. The reply mail should be
addressed to: [EMAIL PROTECTED] - Then replace any text
on the message's subject line: cmd:unsubscribe
--------TO UNSUBSCRIBE DIGEST------
Do above but with this on subject line: cmd:unsubscribe digest
**************************************************************