Comparison of the Top 5 Single Board Computers

This blog post will compare the best 5 known Single Board computers: Raspberry Pi 4, BeagleBone Black, Nvidia Jetson Nano, Google Coral Board, and the Asus Tinker board 2. From basic specs to OS and performance, here is what you need to know.

Time to read:

15

minutes

Single board computers (SBCs) are small computers that can be used for a variety of purposes from experimentation, learning to code, building a home entertainment system, to robotics and home automation. Most of these devices are also capable substitutes for desktop computers and offer a perfect way to do some web browsing, word processing, or office work.

So, for people who like to tinker and experiment, a single board computer may be the perfect option. But here's the thing, there are so many options available that it's often difficult to know which model to choose or where to begin looking for the right one. 

Now, the Raspberry Pi 4 is without a doubt the most famous of all single board computers. Since its launch, it has sold over 30 million units and it’s built up a loyal following of devotees. 

But, apart from the Raspberry Pi 4, there are many other models to choose from which all offer their own advantages and disadvantages. Sifting through all this information to find the right one is a challenge, to say the least. 

Fortunately, we’re here to help and make the process of choosing the right single board computer a lot easier. With this post, we’ll look at the top five single board computers in more detail.

Devices

Before comparing the different devices, we must look at the devices and their specs in a bit more detail.

Raspberry Pi 4

Since 2012, Raspberry Pi has been offering developers, makers, learners, and tinkerers a single board computing platform that’s easy to use in a small form factor. As a bonus, it’s also affordable. 

The Raspberry Pi 4 Model B is their latest model and has taken this device to the next level. It’s a significant upgrade from the previous model and offers a ground-breaking increase in processor speed, memory, connectivity, and multimedia performance.

Despite all these improvements, it still offers full backward compatibility with and the same power consumption as the Raspberry Pi 3 Model 3+.

Raspberry Pi 4 Specifications:

The specifications of the Raspberry Pi 4 are the following:

  • Broadcom BCM2711, Quad-core Cortex-A72 (ARM v8) 64-bit SoC @ 1.5GHz
  • 2GB, 4GB or 8GB LPDDR4-3200 SDRAM
  • 2.4 GHz and 5.0 GHz IEEE 802.11ac wireless, Bluetooth 5.0, BLE
  • Gigabit Ethernet
  • 2 USB 3.0 ports; 2 USB 2.0 ports.
  • Raspberry Pi standard 40 pin GPIO header (fully backward compatible with previous boards)
  • 2 × micro-HDMI ports (up to 4kp60 supported)
  • 2-lane MIPI DSI display port
  • 2-lane MIPI CSI camera port
  • 4-pole stereo audio and composite video port
  • H.265 (4kp60 decode), H264 (1080p60 decode, 1080p30 encode)
  • OpenGL ES 3.1, Vulkan 1.0
  • Micro-SD card slot for loading operating system and data storage
  • 5V DC via USB-C connector 
  • 5V DC via GPIO header 
  • Power over Ethernet (PoE) enabled (requires separate PoE HAT)
  • Operating temperature: 0 – 50 degrees C ambient

Beaglebone Black

The Beaglebone Black is a low-cost, community-supported development platform. Although it’s certainly usable for hobbyists, this board has more of an engineering focus. It features dual 46-pin headers, 4GB of storage, and a NEON floating-point accelerator. It also has two PRU 32-bit microcontrollers.

One thing to keep in mind is that it doesn't come standard with any Wi-Fi or Bluetooth connectivity. Users who need this will need to go for the Beaglebone Black Wireless which has Bluetooth and 802.11 b/g/n Wi-Fi as standard. 

Beaglebone Black Specifications:

The specifications of the Beaglebone Black are the following:

  • AM335x 1GHz ARM Cortex-A8
  • 512MB DDR3 RAM
  • 4GB 8-bit eMMC onboard flash storage
  • 3D graphics accelerator
  • NEON floating-point accelerator
  • 2x PRU 32-bit microcontrollers
  • USB client for power & communications
  • USB host
  • Ethernet
  • HDMI
  • 2x 46 pin headers


NVIDIA Jetson Nano

The NVIDIA Jetson Nano is the latest board in NVIDIA’s Jetson range which also includes the Xavier NX and AGX Xavier models. It aims to make the power of AI available to all makers, learners, and developers. 

The Jetson Nano Developer Kit is a small, powerful computer that’s designed to run multiple neural networks in parallel for applications like image classification, object detection, segmentation, and speech processing. 

The kit contains all the inputs and connections required for prototyping IoT solutions, and also a compute module that can be integrated into industrial solutions. 

NVIDIA Jetson Nano Specifications: 

The specifications of the NVIDIA Jetson Nano Developer Kit are the following:

  • 128-core Maxwell GPU
  • Quad-core ARM A57 @ 1.43 GHz
  • 4 GB 64-bit LPDDR4 25.6 GB/s
  • microSD 
  • Video Encode: 4K @ 30 | 4x 1080p @ 30 | 9x 720p @ 30 (H.264/H.265)
  • Video Decode: 4K @ 60 | 2x 4K @ 30 | 8x 1080p @ 30 | 18x 720p @ 30 (H.264/H.265)
  • 2x MIPI CSI-2 DPHY lanes
  • Gigabit Ethernet, M.2 Key E
  • HDMI and display port
  • 4 x USB 3.0, USB 2.0 Micro-B
  • GPIO, I2C, I2S, SPI, UART
  • 69 mm x 45 mm, 260-pin edge connector

Google Coral Dev Board

The Google Coral Dev Board is a single board computer with Google’s custom Mendel operating system. It’s designed for use with the TensorFlow Lite neural network and also has a full complement of GPIO pins. 

The first thing that makes it unique is the Edge TPU Module which is known as a System On Module (SOM). This basically means that it connects to the baseboard and provides everything needed to make the board run. So, it contains the CPU, GPU, RAM, Wi-Fi, and flash memory. The baseboard then has the necessary connectors for USB, LAN, HDMI, SD Card, Audio, and power.

The on-board Edge TPU is a small Application Specific Integrated Chip (ASIC) that’s capable of performing 4 trillion operations per second, making it ideal for machine learning implementations. 

Google Coral Dev Board Specifications:

The specifications of the Google Coral Dev Board are:

  • NXP i.MX 8M SoC (quad Cortex-A53, Cortex-M4F)
  • Integrated GC7000 Lite Graphics
  • Google Edge TPU coprocessor: 4 TOPS (int8); 2 TOPS per watt
  • 1 GB LPDDR4 (option for 2 GB or 4 GB coming soon)
  • 8 GB eMMC, MicroSD slot
  • Wi-Fi 2x2 MIMO (802.11b/g/n/ac 2.4/5GHz) and Bluetooth 4.2
  • Type-C OTG; Type-C power; Type-A 3.0 host; Micro-B serial console
  • Gigabit Ethernet port
  • 3.5mm audio jack (CTIA compliant); Digital PDM microphone (x2); 2.54mm 4-pin terminal for stereo speakers
  • HDMI 2.0a (full size); 39-pin FFC connector for MIPI-DSI display (4-lane); 24-pin FFC connector for MIPI-CSI2 camera (4-lane)
  • 3.3V power rail; 40 - 255 ohms programmable impedance; ~82 mA max current
  • 5V DC (USB Type-C)
  • 88 mm x 60 mm x 24mm

Asus Tinker Board 2

The Asus Tinker Board 2 is a single board computer from Asus. It’s the second generation of Asus’s Tinker Board and comes with more computing power compared to the first version.  

At its core lies the 64-bit Rockchip RK3399 system-on-a-chip, which has a dual-core ARM Cortex-A72 running at 2.0 GHz and a quad-core ARM Cortex-A53 running at 1.5 GHz. It also features a Mali-T860 MP4 GPU running at 800 MHz and either 2GB or 4GB dual-channel LPDDR4 memory.

It’s important to keep in mind that only the Tinker Board 2S comes with onboard storage in the form of 16 GB eMMC flash storage. 

Asus Tinker Board 2 Specifications:

The specifications of the Asus Tinker Board 2 are the following:

  • Rockchip RK3399
  • Dual-core ARM Cortex-A72 @ 2.0 GHz and Quad-core ARM Cortex-A53 @ 1.5 GHz
  • Arm Mali-T860 MP4 GPU @ 800 MHz
  • 1 x HDMI™ with CEC hardware ready
  • 1 x USB Type-C (DP Alt Mode)
  • 1 x 22-pin MIPI DSI (4 lane)
  • Dual-channel LPDDR4 2GB or 4GB
  • Micro SD(TF) card slot (push/pull)
  • 1 x RTL8211F-CG GbE LAN
  • 1 x M.2 - 802.11 a/b/g/n/ac wireless & BT 5.0 (2T2R)
  • 1 x HDMI audio output
  • 1 x S/PDIF TX pin (from GPIO)
  • 1 x PCM/I2S pins (from GPIO)
  • 3 x USB 3.2 Gen1 Type-A ports
  • 1 x USB 3.2 Gen1 Type-C® OTG port
  • 1 x 15-pin MIPI CSI-2 (2 lane)
  • 1 x 40-pin headers
  • 1 x 12~19V DC Power Input Jack (5.5/2.5 mm)
  • 3.37 inch x 2.125 inch (85 x 56 mm)

Operating System Compatibility

Of all these single board computers, the Raspberry Pi is probably the one that can run the most operating systems. This includes Raspberry Pi OS, previously known as Raspbian, the official supported operating system for the Raspberry Pi. 

It comes in different versions that include either only the desktop or the desktop with recommended software, depending on the version downloaded. 

In addition, it's also compatible with Debian Buster, LibreElec, Ubuntu, RetroPie, and TLXOS amongst others. Apart from these, it’s also capable of running various other Linux-based distributions and Android.

The Asus Tinker Board can run Debian 9 or Android 10 operating systems. In fact, according to Asus, it was engineered to run Android 10. This promises to give users some features that aren't available on some other single board computers. These include things like better 3D computer performance, Android Neural Network support, and increased security. 

Like the Raspberry Pi, the Beaglebone Black can run several operating systems. It comes preinstalled with a Debian distribution, but other compatible operating systems include Ubuntu, Angstrom, Android, and others.

Unlike the computers above, the NVIDIA Jetson Nano and Google's Coral Dev Board are limited when it comes to operating systems. The NVIDIA comes preinstalled with Linux4Tegra, which is based on Ubuntu 18.04. It should therefore be able to run a normal Ubuntu distribution, but Linux4Tegra was specifically designed to run on NVIDIA hardware.

Google’s Coral Dev Board was designed for and comes as standard with Google's custom Mendel operating system which is designed for use with the TensorFlow Lite neural network. It’s important to keep in mind that Mendel isn’t a desktop operating system. 

Performance

When it comes to performance, the Asus Tinker Board 2 appears to offer the most impressive specs on paper with its Rockchip RK3399 six-core system on a chip. This gives the Tinker Board 2 impressive performance compared to other single board computers.

Second in line, and because it's designed for machine learning applications with its Edge TPU, is Google's Coral Dev Board. It simply offers power that many other single board computers can't and is one of the most powerful single board computers on the market.

Next in line is the Raspberry Pi 4. Its Cortex A72 processor offers higher performance and faster clocking speed compared to previous versions. Also, keep in mind that the Raspberry Pi 4 can be easily overclocked which results in much higher scores in performance testing.

Just under the Raspberry Pi is the NVIDIA Jetson Nano. Its processor is a generation behind the processor found in the Raspberry Pi 4 and may therefore lack the punch of the Raspberry Pi. 

At the bottom of the pack is the Beaglebone Black. Its 1 GHz ARM Cortex-A8, while not necessarily slow, just can't compete on processing power in this company.

Video

The NVIDIA Jetson Nano comes with an extremely powerful GPU in the form of the 128 core Maxwell GPU. This makes it suitable for machine learning and artificial intelligence applications where a powerful GPU is necessary. 

However, the problem is that it comes only with one video output, so it isn't possible to use dual displays with the Jetson Nano. As a result, for users who are looking for a home entertainment system, the Jetson Nano may, in some cases, not be the right choice.

That's where the Asus Tinker Board 2 comes in. Its GPU, the Mali T-860 running at 800 MHz, is a multicore GPU which is the highest performance model built on Arm’s Midgard architecture. 

As such, it is specifically designed for complex graphics use cases and offers full support for many frameworks for next-generation and legacy 2D or 3D graphics applications. In addition, it also offers full dual display support at 4K UHD resolution.

Like the Asus Tinker Board 2, the Raspberry Pi offers dual-display support. This makes both of these computers ideal for home entertainment system use. Keep in mind, though, that the GPU in the Asus Tinker Board 2 is far superior to the GPU found in the Raspberry Pi 4. 

Bringing up the rear is the Beaglebone Black. Its PowerVR SGX530 GPU, running at 200 MHz, although capable of some 3D hardware acceleration, can't match the specifications of the other single board computers in this comparison.

So, now the question is whether this leaves Google's Coral Dev Board. As mentioned earlier, the Coral Dev Board is an immensely powerful machine, and it has the necessary GPU capabilities to run TensorFlow and be used in machine learning and artificial intelligence applications. 

The problem, however, with the Coral Dev Board is that it lacks a desktop operating system. In fact, according to the official Coral Dev Board documentation, it's not advised to connect a monitor and keyboard to the board and users should only use network connections. So, although it may be powerful, its main purpose is not to display video.

Memory

When it comes to memory, the Raspberry Pi 4 offers the most options. It offers 2 GB, 4 GB, or 8 GB LPDDR4 memory options. Next up is the Asus Tinker Board 2 which has options for 2 GB or 4 GB LPDDR dual-channel memory.

The NVIDIA Jetson Nano offers only one option with 4 GB LPDDR memory. Google's Coral Dev Board offers 1 GB of memory, but, according to them, options for 2 GB and 4 GB are coming soon. Once again bringing up the rear, is the Beaglebone Black with only 512 MB of memory.

Apart from the Beaglebone Black, all these computers have enough memory for the tasks they’re designed for. Still, the Beaglebone Black is suitable for a variety of uses despite its low memory.

Connectivity

The Raspberry Pi 4 offers two USB 3.0 ports, two USB 2.0 ports, a Raspberry Pi standard 40 pin GPIO header which is fully backward compatible with previous models, two micro-HDMI ports, a 2-lane MIPI display port, a two-lane MIPI CSI camera port, stereo audio and composite video port, Gigabit Ethernet, Wi-Fi, and Bluetooth 5.0.

The NVIDIA Jetson Nano offers four USB 3.0 ports, one USB 2.0 port, an HDMI and display port, two MIPI CSI camera ports, Gigabit Ethernet, and Wi-Fi.

The Asus Tinker Board 2 offers one USB 3.2 type-C, three USB 3.2 type-A ports, one MIPI CSI camera port, one 40 pin header, one HDMI audio output, one HDMI port, one USB type-C display port, one MIPI DSI display port, Gigabit Ethernet, Wi-Fi, and Bluetooth 5.0. 

Google's Coral Dev Board offers two USB type-C ports, one USB 3.0 type-A port, and a USB micro-B port. It also has an HDMI 2.0 port, a MIPI DSI display port, a MIPI CSI camera port, a 3.5 mm audio jack, two digital PDM microphones, a four-pin terminal for stereo speakers, Gigabit Ethernet, Wi-Fi, and Bluetooth 4.2.

The Beaglebone Black has a USB port for power and communications, a USB host port, ethernet, HDMI, and two 46 pin headers. It's important to keep in mind that the Beaglebone Black does not come with Wi-Fi or Bluetooth and if it’s needed, users should look at the Beaglebone Black Wireless which has both Bluetooth and Wi-Fi as standard.

Considering the above, all these single board computers are evenly matched when it comes to connectivity with all featuring a wide array of connectivity options. 

Storage

When it comes to storage, all these single board computers come with micro-SD card slots for storage. In addition, the Beaglebone Black comes with 4 GB 8-bit eMMC onboard flash storage while the Coral Dev Board comes with an 8 GB eMMC onboard storage. Also, the Asus Tinker Board 2S comes with 16 GB eMMC onboard storage.

This, ultimately, means that they all offer users the ability to customize the amount of storage based on their needs and requirements by using the appropriate size micro-SD card. Keep in mind, though, that with the Raspberry Pi, the Asus Tinker Board 2, and the NVIDIA Jetson Nano, the operating system will take some space on the SD card. 

Pricing

Now that we've looked at the features of the different single board computers and how they compare, let's look at the pricing of each. Keep in mind, though, that these are approximate prices and pricing may differ depending on the vendor. 

Pricing for the Raspberry Pi 4 Model B starts at $35 for the 2 GB model. This price goes up to $55 for the 4 GB model and $75 for the 8 GB model. 

Also, keep in mind that Raspberry Pi offers the Raspberry Pie 4 Desktop Kit which includes everything needed to use the device as a desktop computer. Pricing for this model starts at $99.95 for the 2 GB model. This price goes up to $119.95 for the 4 GB model and $139.95 for the 8 GB model.

Pricing for the NVIDIA Jetson Nano Developer Kit is $99. It's important to note that NVIDIA has also released the NVIDIA Jetson Nano 2 GB Developer Kit. Apart from having lower memory, it also comes with fewer features than the Jetson Nano Developer Kit. Pricing for this model is $59.

Pricing for the Beaglebone Black is about $60.00. Pricing for the Coral Dev Board is $129.99. As referred to above, there are plans to launch 2 GB and 4 GB models of the Coral Dev Board. At the time of writing this post, prices for these models were not available as yet.

At the time of writing this post, pricing and availability for the Asus Tinker Board 2 were not available yet but it's expected that Asus will try to offer it to the market at a price similar to the previous model. This model started at a price of $55.

The Bottom Line

So, after all the above information, which one is best? Well, the answer is it depends on a specific user’s needs and requirements. So, for example, for someone who focuses only on machine learning and artificial intelligence, the Coral Dev Board or NVIDIA Jetson Nano may be the right choice.

Likewise, for people who want to build a generic solution, the Raspberry Pi or Asus Tinker Board 2 may be the one to go with. And somewhere in between, the Beaglebone Black also has its place with its engineering focus.

Ultimately, depending on a specific user’s needs or requirements, they should choose the model that’s right for them.