Blog

UP Squared/ Raspberry Pi 4/ Jetson Nano: performance and platform comparison We have ran a set of benchmarks to evaluate CPU and Memory performance of UP Squared, Raspberry Pi 4 and Jetson Nano.   Upcoming benchmark testing: GPU and AI. Stay tuned! Are you curious to know how UP Squared performed? Let’s have a look at the results:

Test Systems:

• UP Squared N4200 4GB RAM
• Jetson Nano 4GB RAM
• RPi4 4GB RAM

The test is performed by keeping the system aligned as much as possible, realizing the hardware differences and considering the following baseline:

• All systems are Quad Core
• Same RAM: 4GB
• Same Storage amount: 32GB
• Same OS type: Ubuntu 18.04 64bit
• Same version of Phoronix-Test-Suite and test version

Comparison Overview:

The results with the greatest spread from best to worst included:
✔ Stream (Type: Copy) at 3.92x
✔ OpenSSL (RSA 4096-bit Performance) at 2.68x
✔ PHPBench (PHP Benchmark Suite) at 2.3x
✔ RAMspeed SMP (Type: Copy – Benchmark: Integer) at 2.25x
✔ perf-bench (Benchmark: Sched Pipe) at 2.21x
✔ Parallel BZIP2 Compression (256MB File Compression) at 2.13x
✔ Java SciMark (Computational Test: Composite) at 2.03x
✔ x264 (H.264 Video Encoding) at 1.94x
✔ PyBench (Total For Average Test Times) at 1.89x

Results:

UP Squared N4200 4GB had the most wins, coming in first place for 90% of the tests.
▶ Comparison of UP Squared with the RPi4 and Jetson system’s performance shows,

• superior performance for applications written with the most used programming language, compression tests and network/server applications
• improved performance in cryptography, encoding and imaging
• better results for generic CPU tests and memory tests

▶ UP Squared sits on a higher price range compared to Jetson and RPi4 systems, and while it shares some of the I/O available with the other systems, such as the 40pin General Purpose Header and multiple Video Out, it offers users much more power and expandability with,

• Multiple and Higher Speed USB ports: 3x USB3.0 (Type A), 1x USB 3.0 OTG (Micro B) ), 2 x USB2.0 (internal header)
• Faster Embedded Storage with multiple options: 32 GB / 64 GB / 128 GB eMMC, SATA3
• Multiple Network Interfaces: 2x Gb Ethernet ports, optional WiFi/BT via M.2 2230 (including WiFi 6 modules), optional 3G/4G Connectivity via MiniPCIe
• Expansion: 40 pin General Purpose bus + 4-channel 12-bit A/D converter 60 pin
• AI Vision Kit: AI Core-X module (based on Intel Movidius Myriad-X) optional via MiniPCIe
• Gateways: RED certification including supported Wireless interfaces (WiFi/BT, 3G/4G, LoRa)