Some interresting tests with the new and the old wireless packages, by using 2.9rc7 + RB230 and RB500
Hi, I downloaded package, but unable to install - it is unvisible for installer 
Probably belongs to i386 platform
Yes it is sure - it is not package for routerboards but for I386, damned 
Im running 2 pc routers.First P4-1.4 256 drr 333 v 2.8.28 second p2-233 32 mb sdram v. 2.8.26 with cm9 65 mW cards.I have real 36.6 Mbps bandwitch :smiley: but after few hours speed is only 20 mbps :frowning: Why? sorry for frong section, i cant find post delete option…
Anybody tryed 2.9 with wireless?how about bandwitch?
We did all our wireless bandwidth tests with 2.9RC9.
On the 600MHz platform, you can get 65Mbps through a bridge.
2.9 (release) does work, I have it on my AP now at home, and it’s very nice - but haven’t had time to do bandwidth tests between routers yet. We’ll be doing that soon.
Regards
Just been testing with 2.9.2.
Nstreme performance on slower CPUs is hugely improved - well done MT.
Even on a slow 133MHz Geode box we now get ~24Mbps UDP throughput using bandwidth test - that’s up from 10Mbps or about that with older builds.
Impressed - keep up the good work MT.
Regards
hi stephen,
what about fast processors like your mobile Pentium platform ? These results are far more interesting than ~20Mbps 
thnx!
Hi there,
Just now most of our 600MHz+ boxes are “out” and our lab gear “in a state of flux”. As soon as I get a chance to configure a fast P2P pair I’ll post up results.
Also, we have setup a test network to invite “selected guests” (potential or existing customers) access via VNC to see the performance of our boxes via the net. Let me know if you’re interested -
Regards
speculation : could this higher performance be due to PROPERLY WORKING hardware compression of atheros cards? We all know that bandwidth test uses compressible data, at least to some extent.
somebody from mikrotik wants to shred some light here?
You might be right on some routers, however:
The “better performance” I reported above was with compression “off”, bandwidth-test traffic (which doesn’t compress I notice) and the CPU was limiting.
i.e. the improvement I reported is due to better coding of the “new nstreme” not other factors - IMHO
Regards
good to know! I don’t expect such boost on full-powered beasts, but seems like the code is getting optimized what is excellent sign.
Thanks for sharing info!
hello,
i have done the upgrade of the firmware and routerOS but i now realised that the routerboards hangs during bootup, and when i removed the RouterBouard 71 from the PCI slot of the R230, the routeboard boots successfully. i ready through the manual,and i could deduced that one should disable “pci-backoff” should in case this kind of problem is occurs.
i have tried everything i could. Can anybody be of help
with rb500, with a level of signal of -60, what are the trhouput you can acheive and with which configuration ?