Side by Side 68U vs 3100

[doczenith1]

Last week I jumped on the Best Buy sale and picked up an RT-AC3100. I performed some comparisons and logged the data which I'll let y'all draw your own conclusions on (see attached image). One item worth mentioning is the wired WAN-LAN speed is the same for each router with CTF enabled. With no NAT acceleration the 3100 jumps ahead as expected. All speed results are in Mbps. Both routers were reset to factory defaults after upgrading the firmware and the router was "forgot" and reconnected on the phone when changing routers. If anyone has any particular router configuration settings that they would like me to test just let me know.

 

[maxbraketorque]

Nice comparison. Makes me feel good about staying with my RT-AC68U

 

[SoCalReviews]

It looks like there some small advantage with the download speeds with the 3100. I'd be curious to see a test comparison with the RT-ac1900p which has the same 1.4Ghz speed main processor as the 3100.

 

[doczenith1]

I've been messing around with some testing on torrent download speeds and did a reset on both routers. On the 3100 with defaults plus a forwarded port I'm getting 633d/673u with no NAT acceleration so the AiProtection and Traffic Analyzer (forgot to mention that in the OP) have somewhat of an impact. Another interesting item that I discovered was that after resetting both routers they both reported CTF only. On the 68U I disabled NAT acceleration > reboot > enable NAT accel > reboot and it now shows CTF and FA. Same process on the 3100 and still only CTF.

I found a closer speedtest server that gives me higher download speeds but lower upload speeds so I've been gathering some data using that server in the basement location (garage is getting too cold but wanted to focus on one of the weaker signal locations). I'm thinking this will give better data on maximum throughput. I'll add a few test at the personal desk from a benchmark standpoint.

 

[maxbraketorque]

If you are testing wireless speeds, then doesn't it make more sense to copy files from a LAN-based server? That's what I do.

 

[doczenith1]

If you are testing wireless speeds, then doesn't it make more sense to copy files from a LAN-based server? That's what I do.

Yes it does. I actually didn't even think of that. I only have a cell phone and using a speedtest site was the first thing I thought of for testing the wifi speeds. My WAN connection is also much faster than the wifi connection speed so I knew there wouldn't be a bottleneck there. But, in hindsight using an app on my phone for transferring files would be a much more controlled test and would yield more accurate results. Damn you MBT now I have another item to add to the testing list

Any recommendations for Android apps for transferring files? My quick search brought up AirDroid but the description looks likes it includes a lot of bloatware. And is there a particular file type that works best?

 

[maxbraketorque]

Yes it does. I actually didn't even think of that. I only have a cell phone and using a speedtest site was the first thing I thought of for testing the wifi speeds. My WAN connection is also much faster than the wifi connection speed so I knew there wouldn't be a bottleneck there. But, in hindsight using an app on my phone for transferring files would be a much more controlled test and would yield more accurate results. Damn you MBT now I have another item to add to the testing list

Any recommendations for Android apps for transferring files? My quick search brought up AirDroid but the description looks likes it includes a lot of bloatware. And is there a particular file type that works best?

AirDroid is actually essentially free of any bloatware on the phone. Its quite a handy app, and I use it regularly to transfer photos and files.

 

[doczenith1]

Well I finally got around to doing some testing with AirDroid. I didn't really see any clear cut winner. One thing I noticed is that the max speeds with AirDroid are around 360 Mbps which seemed slow when considering link speeds at the two closer locations was 866. I can routinely get around 500 Mbps when using the certain Ookla servers (unfortunately not the servers I used in the earlier testing).

 

[sfx2000]

Not really surprised with those numbers...

 

[doczenith1]

Decided to venture out into VPN land so I have a few more numbers to share. Both routers running Merlin 380.64. Using the VPN client to connect with PIA VPN servers. Results are below.

AC3100
CTF enabled
DL: 61 Mbps with core 1 at 25%, core 2 at 75%
UL: 84 Mbps with core 1 at 35%, core 2 at 100%

AC68U
CTF + FA enabled
DL: 43 Mbps with core 1 at 25%, core 2 at 80%
UL: 60 Mbps with core 1 at 35%, core 2 at 100%

AC68U
CTF enabled
DL: 44 Mbps with core 1 at 30%, core 2 at 80%
UL: 58 Mbps with core 1 at 40%, core 2 at 100%

As a comparison when using the PIA VPN client on my Snapdragon 808 based cell phone I get 77 down and 124 up.

When using the same PIA VPN server with a windows client I'm able to attain 250 Mbps down and 350 Mbps up. The speed tests were conducted over a wired connection from the computer to the router.

All speed test performed using the DSLReports HTML5 test.

Data encryption: AES-128
Data authentication: SHA1
Handshake: RSA-2048