hat

Recently, I've been working on a Pi router build with multiple 2.5 Gbps Ethernet ports using Radxa's Dual 2.5G Router HAT.

I wanted a simple way to check on total network TCP throughput using both interfaces (or really, as many interfaces as possible) to multiple computers on my network, and I noticed iperf3's --bind option (like --bind [ip address of interface]) was not splitting the traffic on both interfaces—it would just route all traffic through one!

Luckily, I found the issue Failing to bind to interface when multiple interfaces are present, and in it, @bmah888 mentioned the --bind-dev option, which is new as of iperf 3.10+.

Using that option (like --bind-dev [interface name]), you can force an instance of iperf3 to bind to one particular device. For example, assuming I have two servers on my network running iperf3 -s, I can run the following on the Pi to saturate both connections as much as the Pi will allow:

A couple years ago, after I heard about the CaribouLite on CrowdSupply, I pre-ordered one.

I've dabbled in SDR with an RTL-SDR v3 for a few years, even using one with nrsc5 to listen to baseball games OTA because of silly MLB blackout restrictions.

But low-cost SDRs like the RTL-SDR v3 are receive-only, and have a limited frequency range, and lower quality RF filtering, so it can be frustrating if you're trying to work with lower-power RF... or trying to transmit at all!

Today I published a video detailing my testing of three new Raspberry Pi HATs—these HATs all add on PoE+ power and an NVMe SSD slot, though the three go about it in different ways.

You can watch the video for the full story (embedded below), but in this post I'll go through my brief thoughts on all three, and link to a few other options coming on the market as well.

GeeekPi P33 M.2 NVMe M-Key PoE+ HAT

Raspberry Pi today launched the AI Kit, a $70 addon which straps a Hailo-8L on top of a Raspberry Pi 5, using the recently-launched M.2 HAT (the Hailo-8L is of the M.2 M-key variety, and comes preinstalled).

The Hailo-8L's claim to fame is 3-4 TOPS/W efficiency, which, along with the Pi's 3-4W idle power consumption, puts it alongside Nvidia's edge devices like the Jetson Orin in terms of TOPS/$ and TOPS/W for price and efficiency.

Google's Coral TPU has been a popular choice for a machine learning/AI accelerator for the Pi for years now, but Google seems to have left the project on life support, after the Coral hardware was scalped for a couple years about as badly as the Raspberry Pi itself!