Wireless Preformance less on Linux than Win98

[David Gibson]

On Sun, Mar 07, 2004 at 04:37:10PM -0700, Craig Bell wrote:
> I've got a dual booted Athlon 550, running Win98 and SuSE 8.1.  My eth0 
> connection is a Netgear MA-311.  The wireless lan on both systems works fine, 
> but Konquerer is much slower on the SuSE to load pages, etc, than IE on the 
> windows side.
> 
> The Netgear Configuration GUI on Windows shows that my Link Quality is around 
> 90%, and the Signal Level a little less.  When I compare that to iwconfig 
> under linux I get much less, about 64/92:
> 
> eth0      IEEE 802.11-DS  ESSID:"ourwirelesslan"  Nickname:"Prism  I"
>           Mode:Managed  Frequency:2.462GHz  Access Point: 00:30:AB:17:54:E2
>           Bit Rate:11Mb/s   Tx-Power=15 dBm   Sensitivity:1/3
>           Retry min limit:8   RTS thr:off   Fragment thr:off
>           Power Management:off
>           Link Quality:64/92  Signal level:-51 dBm  Noise level:-149 dBm
>           Rx invalid nwid:0  Rx invalid crypt:0  Rx invalid frag:0
>           Tx excessive retries:0  Invalid misc:0   Missed beacon:0
> 
> I'm running the orinoco-0.13e driver on Linux, and the netgear supplied driver 
> on windows.
> 
> I've spent the afternoon trying to find what I can on the web about tuning 
> wireless performance, but aside from being in over my head on most of it, I 
> haven't seen anything that directly compares how to get the link quality up 
> or why it would be less on SuSE than on Windows in the same box in the same 
> room.  I would sure appreciate some kind help to increase my throughput, or 
> at least some easier to read doc on the subject of performance.
> 
> Is there any chance that I'm barking up the wrong tree, and that it's not in 
> the wireless config at all?  Can the lower link quality from iwconfig be 
> explained some other way?

Yes.

I don't really know how to properly interpret the signal level values
that the firmware returns.  So the signal level reported by iwconfig
is quite possibly bogus, or at least on a different scale to that
reported by Windows (it also seems to be reported slightly differently
for different firmware versions, which makes life tricky).

Now if you had a markedly different measured throughput, that would be
different question.

-- 
David Gibson			| For every complex problem there is a
david AT gibson.dropbear.id.au	| solution which is simple, neat and
				| wrong.
http://www.ozlabs.org/people/dgibson