Baseline 5GHz Tests
Most of our tests were run in a living room, with the main system and router at one end of the room and the client #1 machine in the other. We had line of sight connectivity between the devices over a span of about 30 feet. All of these initial baseline tests were run with the routers applying WPA2 encryption, as this is the most common setting for home router deployments.
Our 2GB test involves taking several hundred small files (MP3s, documents, etc.) and compressing them into a single ZIP. This helps to isolate our results from the added processing overhead of sending hundreds of files rather than one.
There’s nothing fancy about this test. It’s just a straight up file transfer to establish a baseline for comparison. Under fairly ideal conditions, our Asus router surprisingly wins the event, although the Netgear R6300 shows the narrowest performance gap between the two transfer directions.
PassMark’s PerformanceTest 8 offers another look at basic throughput, only now we can assess performance of TCP and UDP traffic separately. UDP traffic almost invariably exhibits higher throughput than TCP since it doesn’t involve the latter’s error-checking (and thus packet resending) routines. If a packet fails to transfer, UDP keeps right on going. Here’s a side-by-side look at the potentially extreme difference between TCP (left) and UDP traffic as seen on our Asus router, albeit in a different environment before we started official testing. Same distance, same test, just different packet traffic protocols.
Now back to our comparative PassMark results:
Since we see four different instances of throughput hitting in the 410 to 413 Mb/s range, it seems clear that we’ve reached the ceiling for performance given this benchmark’s parameters. Note, however, that while three of our four routers hit this level for UDP traffic, only the Nighthawk R7000 reach it with TCP traffic. Amped’s RTA15 is the only router not to hit this ceiling with UDP, although it comes close.
Now let’s see if the more in-depth IxChariot corroborates our earlier tests. Remember when we talked about biasing for average throughput first and then falling back on lowest throughput to decide a close race? This is one of those times. All three of our 3x3 routers show similar average and top-end speeds during 5GHz TCP testing. Asus, however, wrecks both Netgear models in maintaining a better throughput floor. Amped also fares well in this regard, leading us to think that all of that power boosting might be making some amends for only having a 2x2 antenna array.
When we performed testing on the Amped RTA15, IxChariot tossed us warning message CHR0336: A timing record was received with a measured time between 1 and 20ms. While the other three routers go quietly about their business (with the Nighthawk showing only the barest advantage), Amped is all over the map, like so:
Compare this to the much more tightly fluctuating Nighthawk on the same test:
We’ll see more of this in the tests to come. Amped struggles with UDP in general, especially across 5GHz.
A final note here. You will notice that our 5GHz TCP results are significantly faster than our UDP results. This is generally not how it works in the real world, as is reflected in PerformanceTest. However, as we discussed in our prior 11ac round-up, Ixia appears to use a custom protocol implementation that skews the relative performance of TCP and UDP.