If you’ve been shopping for lab hardware lately, you’ve probably experienced some serious sticker shock. 😱 The AI boom has fundamentally shifted manufacturer priorities, and traditional computing hardware prices have skyrocketed as a result. But what if I told you that the solution to affordable lab computing might be sitting in your desk drawer or more specifically, in that old flagship smartphone you replaced a couple of years ago? 📱✨
🌍 The Shifting Landscape of Computing Hardware Link to heading
The AI revolution has been incredible for advancing technology, but it’s come with an unexpected side effect, traditional compute hardware for hobbyists and enthusiasts is getting squeezed out. Manufacturers are prioritizing high-margin AI accelerators and data center GPUs, while the modest computing needs of builders are increasingly an afterthought. The result? Higher prices and limited availability for the very hardware that powers our learning labs and self-hosted services.
But here’s the thing, while we’ve been chasing traditional ARM SBCs (Single Board Computers) like the Raspberry Pi, the mobile industry has been quietly churning out incredibly efficient, powerful System-on-Chips (SoCs) for years. And many of these devices are now readily available in the second-hand market at compelling prices.
🔬 The Experiment: Snapdragon 835 vs. Raspberry Pi 5 Link to heading
To test this hypothesis, I decided to pit a mobile SoC against one of the most
popular boards: the Raspberry Pi 5 Model B rev 1.0. But instead of using the
latest and greatest mobile chip, I went with something you might find in a used
phone from 2016-2017 the Qualcomm Snapdragon 835.
Let’s set the stage with the specs:
Raspberry Pi 5 (BCM2712)
- Release Year:
2023 - Architecture:
Quad-core ARM 2.4 GHz Cortex-A76
Snapdragon 835
- Release Year:
2016 - Architecture:
Octa-core Kryo 280 (4x2.45 GHz performance cores + 4x1.9 GHz efficiency cores)
On paper, you’d expect the Raspberry Pi 5 being seven years newer—to
completely dominate. After all, it’s purpose-built for the maker community and
single-board computing. But the reality is far more interesting.
📊 The Benchmarking Results: Surprisingly Close Link to heading
I ran openssl speed benchmarking on both platforms to get a real-world
performance comparison for cryptographic operations—something that’s critical
for many lab workloads like web servers and application servers.
The results? 🎉
The Snapdragon 835 lagged behind the Raspberry Pi 5 by only 13-15%, for real
workloads which generally uses 8 Kilobytes (8192 bytes) block size for data
processing.
OpenSSL Speed Benchmark Results Link to heading
🥧 Raspberry Pi 5 (BCM2712) Link to heading
openssl speed -multi 8 sha256
# #16 bytes #64 bytes #256 bytes #1024 bytes #8192 bytes #16384 bytes
sha256 207364.00k 771244.97k 2233883.46k 4243494.35k 5793920.34k 5845172.22k
📱 Snapdragon 835 Link to heading
openssl speed -multi 8 sha256
# #16 bytes #64 bytes #256 bytes #1024 bytes #8192 bytes #16384 bytes
sha256 96938.25k 364744.94k 1209499.86k 2967597.06k 5170452.89k 5445429.93k
Let that sink in for a moment. A mobile SoC from 2016, came within striking
distance of a 2023 single-board computer. And remember, the Snapdragon 835
was optimized for battery life and thermal constraints of a smartphone, not
sustained compute workloads.
💡 Why This Matters for Your Lab Link to heading
1. 💰 Cost Efficiency is King Link to heading
Used smartphones with Snapdragon 835 (and similar generation chips) are
abundant in the second-hand market. You can often find them for a fraction of
what you’d pay for a new Raspberry Pi 5 and they come with built-in batteries
(UPS, anyone? 🔌), storage, display, and networking already integrated.
2. ⚡ Performance Per Dollar Link to heading
When you’re only giving up 13-15% performance for potentially 50-70% cost savings, the math starts to look really compelling. For many lab workloads such as DNS servers, lightweight web applications, monitoring tools, this performance difference is negligible.
3. 🔋 Power Efficiency Link to heading
Mobile SoCs are designed from the ground up to be power-sipping beasts. The
Snapdragon 835’s 10nm process was cutting-edge for its time, delivering
excellent performance per watt. In a homelab running 24/7, those electricity
savings add up quickly.
4. 🌡️ Thermal Characteristics Link to heading
These chips are designed to run in passively cooled environments (your pocket). This makes them ideal for compact, quiet lab deployments where active cooling is undesirable.
🌐 The Broader Implications Link to heading
This isn’t just about Snapdragon 835 vs. Raspberry Pi 5. It’s about
recognizing that we have an entire ecosystem of powerful, efficient, and
affordable computing hardware that’s been hiding in plain sight. As flagship
smartphones are upgraded every 1-2 years, millions of perfectly capable
computing devices are relegated to drawers or recycling centers.
🏗️ Real-World Usage: CNCF Components on Mobile SoCs Link to heading
Theory is great, but what about practical applications? I’ve been successfully
running popular
CNCF (Cloud Native Computing Foundation) components
on the Snapdragon 835 SoC, including:
- Prometheus - The industry-standard monitoring and alerting toolkit
- Node Exporter - Hardware and OS metrics exporter for Prometheus
These aren’t lightweight toy applications—they’re battle-tested cloud-native tools used by organizations worldwide. Running them on a 2016-era mobile chip demonstrates that these devices are more than capable of handling serious lab workloads.
The setup has been stable, performant, and power-efficient. Prometheus scrapes metrics, stores time-series data, and serves queries without breaking a sweat. Node Exporter reliably collects and exposes system metrics. All of this on a chip originally designed to power a smartphone.
Want to see it in action?
Dashboard Snapshot
of node exporter running on Snapdragon 835. (Depending on your network, it
might take a few seconds to load the dashboard). I will be sharing more details
about the setup and configuration in an upcoming post, so stay tuned!
This real-world deployment proves that mobile SoCs aren’t just a theoretical alternative but they’re a practical, viable option for modern lab infrastructure running cloud-native workloads.
⚠️ The Challenges (Because Nothing’s Perfect) Link to heading
Let’s be honest, using mobile hardware for computing isn’t without challenges:
- Software Ecosystem: While there are solutions like termux which allow you to install various Linux packages on Android, the packaging and support for certain applications may not be as robust as on traditional ARM SBCs.
- Vendor restrictions: Most vendors have recently cracked down on unlocking bootloaders (for e.g. you can only unlock bootloader on a single device in a calendar year), which can restrict ability to install certain softwares.
- Documentation: Not as friendly as Raspberry Pi.
- GPIO and expansion: Limited (Serial communication via USB port) or None compared to traditional SBCs.
🎯 The Bottom Line Link to heading
As computing costs continue to climb due to AI-driven market dynamics, we need to get creative about our lab infrastructure. Mobile SoCs represent an untapped reservoir of affordable, efficient computing power that deserves serious consideration.
The next time you’re about to drop $80-100 on a Raspberry Pi 5, consider
this, that old flagship phone might just be your next lab server. 🚀 It’s not a
bad idea—it might actually be brilliant. 💪