Re: [uCsimm] eth circuits: are there 100mbit circuits too?

From: Glenn West (gwest@krdl.org.sg)
Date: Thu Oct 05 2000 - 20:50:39 EDT


I Agree.

Really the point of the current SIMM is connectivity. The other thing to
keep in mind, I have a "Broadband" gateway at home, its a off-the-shelf
product, a ARM7Thumb and two ethernet 10baseT for a couple hundred
dollars. Its works great. Dont forget that "physical" specs, and logical
specs are different. The cable modem will only give me 1 to 2 megabits
anyway, so why waste money on 100megabit ports. The only reason
the 100baseT is nice, is sometimes newer hubs dont have 10baseT
ports. Your doing it just for "easy" plug and play, but your not using
the bandwith. The ARM is about the same performance, maybe
slightly better than the process or the uCsimm. Also dont forget
that not that many years ago, the equiv of the UcSIMM was a
SUN workstation. So it can do something very useful with what
you have. But you cant do "wire-speed" processing.

Some ethernet chips do have buffer memory, but non will
send a "ack" for your TCP packet. That would require processing
the whole protocol stack in hardware. While there are a
few, that would add significant cost to the board.

At 07:07 AM 10/6/2000, you wrote:
>Well, my uCsimm can't even push maxing out the 10bt interface so there is
>really no point in adding anything faster. Choosing a 100bt chip over a
>10bt would only add extra cost not extra speed.
>
>Patrick Stingley wrote:
>
>>This is a good question and one that I have debated at length
>>elsewhere. I would have to think that a 100Mbps ethernet controller
>>would have to deal with a blocked state if the CPU is busy doing
>>something else. Similarly, I would think that the CPU would deal with
>>traffic from an ethernet interface in an interrupt-driven fashion.
>>So, it shouldn't matter if the CPU isn't very fast, the ethernet
>>interface should go into blocked mode and either buffer packets or NAK
>>them and wait for the processor.
>>I'm not a EE, but merely a lowly computer scientist, so hopefully one of
>>the wire twisters in our midst will step up to the plate and shed some
>>light on this subject.
>>Best regards,
>>Pat Stingley
>>At 03:50 PM 10/5/00 +0200, Alex Beregszaszi wrote:
>>
>>>Hy :)
>>>Are they 100mbit/s ethernet circuits too?
>>>Or this CPU too slow to handle (route) so big bandwith?
>>>Alex
>>>This message resent by the ucsimm@uclinux.com list server
>>>http://www.uClinux.com/
>>
>>This message resent by the ucsimm@uclinux.com list server
>>http://www.uClinux.com/
>
>
>
>This message resent by the ucsimm@uclinux.com list server
>http://www.uClinux.com/

This message resent by the ucsimm@uclinux.com list server http://www.uClinux.com/



This archive was generated by hypermail 2b30 : Sun Apr 07 2002 - 00:01:38 EST