Performance Testing

A combination of iPerf3 on my desktop PC and laptop and the iPerf3 Android app were used to measure throughput for the router, with a ruler or tape measuring the distance from it and a Wi-Fi analyzer using my phone's or laptop's Wi-Fi antenna and connection to measure signal attenuation as I moved from right on top of the router to further and further away inside my residence. Please note that signal attenuation depends heavily on a number of factors apart from just distance, including any physical barriers and other devices operating in the bands (2.4 GHz, 5 GHz, 6 GHz, etc). With a common test location, the variable being tested is the router and Wi-Fi itself, and it would as such be valid to compare results to other products tested thus.


The plot above shows how the antennas in the MSI RadiX AXE6600 fare with clients connected at various distances. The client was positioned at varying distances in a 3D space, with the shortest distance measurement taken between the two. Given the nature of the supported bands, this test was conducted for wireless-N (WiFi 4) at 2.4 GHz, wireless-AX (WiFi 6) at 5 GHz, and wireless-AX (WiFi 6E) at 6 GHz. Previous testing confirmed the signal attenuation depends almost entirely on the operating band itself. As such, WiFi 6 over 2.4 GHz should be the same here as WiFi 4 on 2.4 GHz, and likewise WiFi 6 at 5 GHz would be within error margins on WiFi 5 (wireless ac) at 5 GHz too. This is why I opted to include only the three primary band/standard combinations that customers of the RadiX AXE6600 would operate in. We can see that signal loss is heavier for the 5 GHz network relative to the 2.4 GHz network, which is as expected. Likewise, a similar trend is seen with 6 GHz attenuation compared to 5 GHz. All three in general have a relatively steep fall compared to WiFi mesh systems we've tested in the past, although I will give MSI props given the signal attenuation here is not as fast compared to some more expensive routers that have internal antennas. Those six external antennas here that can be directed to face the client clearly help, and these six outperformed the eight external ones on the recently tested Reyee RG-E5 router too.


By removing the actual internet speed variable from the equation, a TCP throughput test done at these same spots from the router paints a more useful story, while also helping with a comparison of those numbers across test residences by having the X-axis as the signal's attenuation instead of distance from the router. As you can imagine, these tests took a long time and there is so much going on that it made sense to split the results across the four tested combinations. The first graph has the older, yet more prevalent, wireless-N and wireless-AC networks tested at 2.4 and 5 GHz, respectively. While the latter performed as expected, I noticed the 5 GHz band seemed somewhat nerfed on the RadiX AXE6600 with the throughput numbers not where they should have been. Indeed, the same could be said about wireless-AX (WiFi 6) at 5 GHz too and this is when I noticed MSI went for a 2x2 radio configuration here for a max of 1200 Mbps throughput on the 5 GHz band. It's only the 6 GHz band that gets a 4x4 radio with up to 4800 Mbps throughput to where I have to say you should only get this router if you are going to use it for WiFi 6E operation. The 6 GHz band introduces up to 160 MHz channel width compared to the max 80 MHz on the 5 GHz band and 40 MHz on the 2.4 GHz band. This wider and potentially less crowded band allows for more throughput too, and I ended up with a max of 4651 Mbps right on top of the router here—over four times as much as WiFi 6 on 5 GHz to further cement the fact that MSI has designed the AXE6600 for 6 GHz operation and the RadiX AX6600 should be more up your alley if you are still on WiFi 6. I will once again remind people that I have the European model here with only four channels on the 6 GHz band, so the other versions might perform even better there.


Here is where all the data above comes together for some comparisons, whereby I chose a distance of 5 m from the router and charted the throughput from various routers on wireless-N at 2.4 GHz, wireless-AC and AX where applicable on the 5 GHz band, and wireless-AX on the 6 GHz band. I've finally had enough entries with wireless-AX to merit a third graph now for easier comparison and there will be more WiFi 6E options included here down the line. This makes it all the more obvious that the RadiX AXE6600 is not ideal for those still on WiFi 5 or 6 as you can do better for the money easily—especially for WiFi 6 where it only bests a mobile router! Meanwhile, it effectively bullies the competition in WiFi 6E mode. I am not aware of the RadiX AXE6600 supporting a mesh configuration if you purchase multiple of these units so it's best considered a standalone router only. This does mean you should be aware also that some of the better performing products on the older WiFi standards are mesh systems of two or three units in total.

As for wired performance, there's not a lot to talk about with the four GigE LAN ports on the router, which are the least you can expect on a router in 2023. I would have liked to see at least one multi-GigE LAN port here in addition to the designated WAN port itself, but at least you do get a working USB port. I also briefly tested the router for its QoS modes to see whether the potential privacy nightmare is even worth it. Things are less obvious when you are living by yourself and have an extremely fast, non-throttled, non-data-capped internet connection but even so downloading a large game from Steam shaved off ~3-4% consistently on the "Game Mode" and the default AI-Auto actually did better (we are talking a few seconds) after a week of testing. It appears there's certainly a learning period here, at least while GT Booster gets more data points from AXE6600 users. Latency over WiFi was no different from the others at the same / similar throughput and attenuation, which again is more to do with physics than anything magic MSI can try to claim out of a so-called gaming router. I am not an online gamer by any means and I am plenty happy with the ping performance here. If you need more specific features for gaming, you may have to look at dedicated firmwares such as DumaOS, available for specific routers.


The power-draw comparison chart identifies whether specific routers are vastly different from others, which turns out to be the case here. A Brand Electronics 4-1850 power meter was put between the power adapter for the router unit and wall socket. Simple Kill-A-Watt units are good for basic checks, but not reliable enough for tests in my opinion. The MSI RadiX AXE6600 router was set up for a minimum of 24 hours of use across multiple days, and power consumption was averaged across a period of idle (inactivity at night) and normal operation (during the day). Note also that the Nighthawk M2 and M5 are mobile hotspot routers and have different battery modes, including a battery-only operation, which does throw things off somewhat. I suppose that relatively beefy power supply and the seemingly overkill cooling solution are useful now that we see this is among the more power consuming individual units tested to date. Indeed, during the 6 GHz testing above, I measured the power draw at a max of 38 W although it settled around 36 W normally—on par with the per-unit numbers for the high performance NETGEAR Orbi systems.