Mulithoming..

[Madison Kelly]

PCI-X = 64bit Bus x 133MHz = 1,064,000,000bytes/sec (~1.014GB/sec)
64bit PCI = 64bit Bus x 66MHz = 528,000,000bytes/sec (~503MB/sec)
32bit PCI = 32bit Bus x 33MHz = 132,000,000bytes/sec (~126MB/sec)

1Gbit = 125,000,000bytes/sec (~119MB/sec) MAX, you never really get it. 
Even full duplex (~238MB/sec) you still only half saturate the PCI bus. 
In reality you are lucky to get ~3/4 of the actual speed and then there 
is the rarity of actually fully saturating both incoming and outgoing 
ports.

HTH!

Madison

Justin Zygmont wrote:
> that's the thing, I read somewhere that 64 bit 66Mhz PCI was only 
> 533Mbps/sec, and AGP 4x was just barely enough for 1000Mbps/sec.  Not sure 
> how true this still is, has anyone ever been lucky enough to use a card 
> like this?
> 
> 
> On Mon, 13 Oct 2003, Ralph Doncaster wrote:
> 
> 
>>Not just a fast PC, but a fast bus too.  Standard 33Mhz/32-bit PCI can't
>>handle full-duplex GigE.
>>
>>Ralph Doncaster, IStop.com president
>>6042147 Canada Inc.
>>
>>On Mon, 13 Oct 2003, Justin Zygmont wrote:
>>
>>
>>>oh ya. and I guess you could never get the full speed of the NIC since
>>>they're using the same cable.  Also, speaking about NICs, I just saw a
>>>1000BaseT ethernet card for only about $40, must take a fast PC to be able
>>>to transfer that fast.
>>
>>--
>>The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
>>TLUG requests: Linux topics, No HTML, wrap text below 80 columns
>>How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>>
> 
> 
> --
> The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
> 


--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml