TamoSoft Throughput Test Results
The first test was conducted at a distance of 10 feet, which is slightly more than double the minimum recommended separation distance of 1.5 meters. ASUS suggests reducing the transmitting power of the PCIe Adapter and the adapter if they are located any closer than 1.5 meters. The default transmitting power is set at the factory to 80mW, and is easily adjusted with the included software. In this test, the PCIe Adapter and router were located at approximately the same height and there was a clear line-of-sight between the two sets of 3x antennas, with no obstructions.
The first thing I noticed with the TamoSoft results is that the UDP performance of all three PCIe Adapters was more uniform. The results with the Passmark test suite seemed like more of an On-Off situation, with the best performance six or seven times higher than the best of the rest. I’m glad I had more than one test suite in house, to gain a better perspective. All that pain I went through when I got the latest 10GbE networking hardware on the test bench paid off…
With only 10 feet separating the PCIe Adapter and routers, and no obstructions or reflecting surface between the two, radio signals were clear and stable during this test. The ASUS combination of PCE-AC66 adapter and RT-AC66U IEEE 802.11ac router pulls first place in both TCP and UDP tests. The average result for the ASUS pair was 121 Mbps with TCP and 611 Mbps with UDP. Both of those results are for the “Downstream” test, as I explained in the Testing Methodology section. That’s a 50% improvement over the n-based routers in TCP, and a whopping 2.6x performance advantage in UDP. Even in less challenging situations, the new 802.11ac standard provides a significant performance benefit.
Moving the PC and Wi-Fi adapter into an adjacent room, with double the distance and multiple obstructions between the antenna arrays changed the results slightly. In TCP, the ASUS 802.11ac pair went from an average throughput of 121 Mbps down to 106 Mbps. That’s a 12 percent reduction, but in real life, you probably wouldn’t notice it. The Linksys actually gained about ten megabits per second in this test, while the TRENDnet lost about the same amount. The UDP benchmarks followed a similar trend, with the ASUS throughput reduced to 447 Mbps, about a 27% loss. The Linksys gained about 50 Mbps, and the TRENDnet also gained about 20 Mbps of throughput at the longer distance, complete with obstructions. Anyone who says they can make sense out of radio waves is a magician, a god, or a liar. These are all still good, usable throughputs for a home environment, especially since they are real data rates, not theoretical numbers.
Next, I dragged the test PC downstairs to the room that’s furthest away from the router’s stationary location in the library upstairs. The furthest downstairs room happens to be the pantry, right next to the kitchen. Both rooms have lots of dense wooden and metal items on the surrounding walls to block and reflect radio waves. The only thing more challenging would be to go over to my neighbor’s house and set up shop in his kitchen! The ASUS PCE-AC66 and RT-AC66u pair held on to most of their line-of-sight performance levels, even in the toughest location. The TCP throughput was down only slightly, to 113 Mbps for the ASUS, which is less than a 10% loss. The other two PCIe Adapters both dropped back to 74 and 66 Mbps, which are still decent results. In UDP mode, the ASUS team stayed close to the rate achieved in the adjacent room, just dropping back about 5% to a 425 Mbps rate. The two 802.11n routers lost quite a bit more performance in this location and achieved 183 and 117 Mbps of throughput.
The results with the TamoSoft software test application were generally consistent with the ones I obtained with the Passmark Advanced Network Test. The UDP benchmark in the TamoSoft app is probably more realistic than what Passmark provides. There were too many packet losses with the Passmark test, and the wild oscillations that occurred at the beginning of the test were not present when I used the TamoSoft throughput test. Clearly, none of these results are anywhere near the typical wired data rates of 1 Gbps. They never will be, until we have to start encrypting ALL our data transmissions, in a vain attempt to avoid having our Netflix data being monitored.