CLI dashboard showing real-time and historical power draw readings from a Kasa smart plug
The Challenge: Performance vs. Power Costs
High-performance computing often means high electricity bills, especially in regions with time-of-use pricing. Finding the balance between computational power and energy efficiency becomes critical when running resource-intensive workloads.
Current Workstation Specs
- CPUs: Dual EPYC 7C13 (base 2GHz, boost 3.7GHz) – 128 cores total (256 threads)
- Motherboard: Gigabyte MZ72-HB2
- RAM: 512GB DDR4
- GPUs: 2x A4000 16GB
- GPUs: 2x 4090 48GB (Upcoming)
- Storage: 10TB NVMe
- PSU: Coolmaster 1600W Platinum
Understanding Power Usage
I have been setting up a workstation for off-loading compute-intensive LLM workflows from my desktop. The build includes dual Epyc CPUs with plans to add 4 GPUs for up to 128GB VRAM to run multiple agents in parallel. A key constraint was ensuring the system could run comfortably on a standard household 120V outlet.
Since I already use smart plugs throughout my home, I added one to monitor the workstation’s power consumption. However, checking power statistics using the kasa phone app or Home Assistant dashboard proved cumbersome, especially when I already maintain a terminal window with monitoring tools like htop, nvtop, and nload in a 2×2 grid on my secondary display.
TP-link Kasa EP25 Smart Plug
Building a Terminal-Based Solution
After sketching a simple wireframe for what I wanted: a clean, terminal-based UI that would display power consumption data already being collected by my Home Assistant instance through the TP-Link integration. After getting the data access sorted, I spent some time with ‘Claude 3.7 Sonnet Thinking’ iterating on the CLI structure and features.
The result: WattWise, a lightweight CLI tool that pulls power usage data from smart plugs (either directly or through Home Assistant) and presents it in a clean, information-dense dashboard right in the terminal.
WattWise Demo
Key Features
Monitoring Features
- Real-time power monitoring with wattage and current display
- Color-coded power values (green for low usage, yellow for medium, red for high)
- Historical consumption charts directly in the terminal
- While smart plug measurements aren’t lab-grade (typically ±1-3% accuracy), they’re more than sufficient for practical usage monitoring
Po
14 Comments
ajsnigrutin
Why all this instead of a simple cronjob switching from performance to powersave profiles depeding on current time (=electricity price)?
throwaway3231
It's usually more energy-efficient to finish a task quickly with a higher power draw, also known as race-to-idle.
PeterStuer
Nice project, but would it not be more rational to have your system running underclocked/undervolted at the optimal perf/watt at all times, with an optional boost to max performance for a time critical task? Running it away from the optimum might save on instant consumption but increase your aggregate consumption.
gtirloni
Could you share how much you have saved in $?
whalesalad
Wonder if a big UPS/power bank would be better? Charge it during periods where power is cheaper, and utilize it when power is more expensive. Then again if you do not need full performance all the time – this is a cool solution.
russdill
If your computer is still doing bursty jobs during that period, it will use less power but still as much energy. Sure, you can reduce the power but if you aren't also reducing what you ask it to do, it'll just use that max amount of allowed power for a longer period of time.
PeterStuer
Had a quick look through the code but I can't find where he actually throttles the PC. Anyone can point me to it?
https://github.com/naveenkul/WattWise
Havoc
From what I’ve seen price per token make home generation uncompetitive in most countries. And that’s just on elec – never mind cost of gear
Only really makes sense for learning or super confidential info
naveen_k
Quick update: Definitely wasn't expecting this to end up on the front page. I was more focused on publishing the dashboard than the power optimizer service I'm running. I'll take all the feedback into account and will open source an improved version of it soon. Appreciate all the comments!
gitpusher
People have made valid criticisms about the basic effectiveness of your strategy. But in any case, this is a pretty awesome hacker project – nicely done! Love the appearance of your CLI tool. I am definitely bookmarking for future inspo
gorbypark
Pretty neat! I’m currently working on a project that uses an ESP-C6 that just exposes a “switch” over matter/thread thats based off the results from the Spanish electricity prices API.
The idea is have the switch be on when it’s one of the cheapest hours of the day, and off otherwise. Then other automations can be based on it.
This was pretty trivial to do in home assistant but I want something that’s ultra low power and can just be completely independent of anything for less technical users.
My end goal is to have a small battery powered device that wakes up from deep sleep once a day to check the day ahead prices via WiFi. The C6 might be overkill for this, but once I have a proof of concept working I’ll try and pick something that’s ultra low super ultra low power. Something that needs charging once or twice a year would be ideal.
The ideal form factor might be a smart plug itself, but I can’t find any with hackable firmware and also matter/thread/wifi.
black_13
[dead]
Symbiote
Within the next year or two, I'm going to look at implementing something similar at my work.
We don't pay for electricity directly (it's included in the rackspace rental), but we could reduce our carbon footprint by adjusting the timing of batch processing, perhaps based on the carbon intensity APIs from https://app.electricitymaps.com/
Though, the first step will be to quantify the savings. I have the impression from being in the datacentre while batch jobs have started that they cause a significant increase in power use, but no numbers.
mythrwy
This looks cool but I feel it should notify the user with a snip from the song "You Suffer" by "Napalm Death" when throttling occurs.