[Prism54-users] SMC2802W V2 (prism54 nitro)

Dave Feller dave.feller@conexant.com
Sat, 20 Mar 2004 11:14:41 -0500


Yes - everything ARM based is capable of running XM... Obviously the
compression rate is dependent on how much host "power" is available as
we do compression in the upper MAC (running in the driver) -=20
For Access points, there are two distinct types - those that use an
external host Network Processor and those that use an integrated NPU/MAC
- We call the  later WiSoC (Wireless System on a Chip)... So Client to
client rates are always great (assuming a decent computer is used) and
AP rates are dependent on the power of the NPU. We have the best luck
with Xscale based routers as they have the most available Mips... But
everything is upgradeable with software to get this functionality if it
is g or a/g product.

David Feller
Director Prism Marketing
Conexant
972-562-7379
=20

-----Original Message-----
From: prism54-users-admin@prism54.org
[mailto:prism54-users-admin@prism54.org] On Behalf Of Feyd
Sent: Friday, March 19, 2004 11:51 PM
To: prism54-users@prism54.org
Subject: Re: [Prism54-users] SMC2802W V2 (prism54 nitro)

On Fri, 19 Mar 2004 10:44:13 -0500
"Dave Feller" <dave.feller@conexant.com> wrote:

> There are actually two revisions of Nitro - the first released early
> last year is simply a software addition of packet bursting and
> protection mechanisms for G solutions - this did two things - it helps
G
> radios play better when b radios are around (before the release if a b
> radio entered the network the whole thing would come down to b speeds
> and the whole benefit og G went away) - and by sending more packets
> through the air with less dead time in between sped up the whole
network
> by about 15%...
>=20
> The new Nitro (NitroXM) adds compression (actually much like WinZip
> compression but real time), more packet bursting, concatenation, and a
> new feature we call DirectLink that allows client to "associate"
> directly with each other bypassing the AP if they need to send data
> locally thereby effectively doubling the throughput... We are claiming
a
> max of 140Mbps throughput  - we have shown 90ish using Chariot TCPIP
> client to client. UDP goes even faster... SMC announced XM but have
not
> seem it on the shelves myself - expect it very soon.

Are the present chips capable of the NitroXM? From the product brief it
seems to be software update too, as was the old Nitro. That would be
very nice :)

Feyd
_______________________________________________
Prism54-users mailing list
Prism54-users@prism54.org
http://prism54.org/mailman/listinfo/prism54-users