Ok, we recently got some of the new R52N cards and wanted to try them out. First, I tried one R52N card in a RB133 and another R52N in another RB133. I made on an AP, and one a client. It took me awhile to figure out why it could only do around 13 Mbps, but I finally came to the conclusion that it was CPU limited.
So I thought, Ok I'll try it out on RB433AH's. First I tried two 802.11a cards on version 3 RouterOS. When the cards were put into turbo mode, I was able to achieve 40 Mbps without the CPU usage being maxed. Then I upgraded to version 4 beta 4, and could only achieve around 34 Mbps with the 802.11a card and the R52n. One of the RB433AH's had it's CPU maxed at 100%. I have both RB433AH's in bridge mode. The one that had it's CPU maxed is connected to the internet through eth2. eth1, eth2, and wlan1 are all Bridged together on it. So I'm guessing because it's connected to the internet through eth2, it causes that RB433AH to work harder than the other one.
I was really surprised at these results and do not see how those boards at 680 MHz cannot do better. What is the R52N card for if it's maximum throughput cannot be obtained because of board limitations. It seems like version 4 is wayyy slower than version 3. I realize a RB600 might do a better job, but the 433AH is a fairly new board. Does anyone have any suggestions, or will the boards just not handle any more throughput than that?
(update)
Apparently one reason I'm having trouble with the CPU usage is because I'm using the btest tool and i'm testing to the RB433's. However, even testing from a computer to a computer, I am having difficulty getting very high speeds.