GPU Efficiency Calculator: FPS per Watt and Energy Cost

A faster graphics card isn’t necessarily a less efficient one. A powerful GPU may finish each frame so quickly that it delivers more performance for every watt it consumes. At the same time, an efficient-looking card can waste electricity if it runs at an unnecessarily high frame rate for hours every day.

The two calculators below help you look beyond headline wattages. The first estimates how much a graphics card costs to run. The second calculates its gaming performance per watt, allowing you to compare GPUs tested under the same conditions.

You’ll get the most accurate result by entering measured gaming power from a reputable review or your own monitoring software. Manufacturer specifications are useful when planning a build, but they don’t tell you exactly how much electricity a particular game will use.

What Does FPS per Watt Mean?

FPS per watt measures how much gaming performance a graphics card produces for each watt of power it uses. The formula is straightforward:

Average frame rate ÷ average gaming power = FPS per watt

Suppose one graphics card averages 120 FPS while consuming 220 watts:

120 ÷ 220 = 0.545 FPS per watt

A second card averages 105 FPS at 170 watts:

105 ÷ 170 = 0.618 FPS per watt

The first card is faster, but the second card is more energy efficient. Whether the second card is the better purchase depends on its price, the games you play and whether 105 FPS is already enough for your monitor.

This distinction matters because raw performance and efficiency answer different questions:

• Average FPS tells you which GPU is faster.
• FPS per watt tells you which GPU turns electricity into frames more efficiently.
• Price per FPS tells you how much performance you receive for the purchase price.
• Total cost per FPS includes both the purchase price and estimated electricity cost.

No single measurement identifies the best graphics card for everyone. A useful comparison considers all four.

How Much Does a GPU Cost to Run?

To estimate a graphics card’s electricity cost, convert its average power consumption from watts to kilowatts and multiply it by the time used and your electricity rate:

Watts ÷ 1,000 × hours used × electricity price per kWh

For example, imagine a GPU averages 220 watts during gameplay. You use it for 14 hours per week and pay 0.30 per kilowatt-hour.

• 220 watts equals 0.22 kilowatts.
• Fourteen weekly hours equals 728 hours per year.
• 0.22 × 728 equals 160.16 kWh per year.
• 160.16 × 0.30 produces an annual GPU energy cost of 48.05.

Over four years, the GPU portion of your gaming electricity use would cost approximately 192.19, assuming your habits and electricity rate remained unchanged.

That isn’t the cost of running the whole computer. Your CPU, monitor, cooling fans, motherboard, storage, lighting and power-supply losses also consume electricity. Our guide to gaming PC power consumption explains how to estimate the complete system rather than the graphics card alone.

Don’t Use the PSU Rating as Power Consumption

A computer with an 850-watt power supply doesn’t continuously use 850 watts. The number printed on the PSU describes the amount of output it’s designed to provide, not the amount the system constantly pulls from the wall.

A gaming PC may draw only a small fraction of that capacity while browsing the web, considerably more while gaming and briefly approach a higher level during demanding workloads or transient power spikes.

The power supply also draws slightly more electricity from the wall than it delivers to the components because no conversion process is perfectly efficient. If a system needs 500 watts of DC output, its wall consumption will be somewhat higher.

This is why whole-system measurements taken at the wall and GPU-only measurements shouldn’t be placed in the same comparison table. They measure different things.

Advertised GPU Power vs Real Gaming Power

Graphics card manufacturers publish figures such as Total Graphics Power, Total Board Power or Typical Board Power. These specifications are useful for choosing a suitable power supply, planning cooling and comparing the general power classes of different products.

They aren’t the same as measured consumption in every game.

For example, Nvidia lists reference Total Graphics Power figures of 145 watts for the GeForce RTX 5060, 180 watts for the RTX 5060 Ti, 250 watts for the RTX 5070, 300 watts for the RTX 5070 Ti and 360 watts for the RTX 5080. AMD lists a Typical Board Power of 220 watts for the Radeon RX 9070 and a Total Board Power of 160 watts for the 16GB RX 9060 XT.

Those figures offer useful context, but the actual draw can change according to:

• The particular game and graphics engine
• Resolution and image-quality settings
• Ray tracing
• Frame-rate caps
• Upscaling and frame generation
• The card manufacturer’s factory settings
• Cooling and operating temperature
• Voltage, frequency and power-limit settings
• Whether the system is CPU-limited or GPU-limited

Even two cards carrying the same GPU name may have different limits and factory overclocks. Nvidia notes that specifications can vary between add-in-card manufacturers, so check the exact model rather than assuming every version behaves like the reference design.

For current specifications, consult the official Nvidia GeForce RTX 50 Series and AMD Radeon RX graphics pages.

How to Compare GPU Efficiency Properly

FPS-per-watt calculations are only meaningful when the GPUs were tested under equivalent conditions. You can’t fairly divide a 1080p result from one review by a 4K power measurement from another.

For a defensible comparison, use the same:

• Games and benchmark sequences
• Resolution
• Graphics-quality preset
• Ray-tracing settings
• Upscaling mode
• Frame-generation setting
• CPU, memory and operating system
• Driver generation
• Power-measurement method

A multi-game average is usually more informative than one favourite title. One architecture may be exceptionally efficient in a particular game and comparatively weak in another.

The frequently updated Tom’s Hardware GPU benchmark hierarchy is one place to find consistently tested performance data. Other original hardware reviews can also work, provided their methodology and power measurements are clearly documented.

Don’t combine benchmark numbers from unrelated reviews unless the test conditions genuinely match.

Should You Measure GPU Power or Whole-System Power?

Use GPU-only power when comparing the efficiency of graphics cards. This isolates the component you’re investigating and reduces the effect of differences elsewhere in the test system.

Use whole-system power at the wall when estimating your electricity bill. Your energy provider charges for everything entering the computer, not just the GPU.

A wall-power meter is the most direct option for checking total consumption. Software tools can also report GPU power, although the available sensors and their accuracy differ by hardware.

The important thing is to label the measurement honestly. A result should specify whether it represents:

• GPU chip power
• Graphics board power
• Complete PC power
• Complete setup power, including displays and accessories

Without that label, a wattage number can be badly misleading.

Why a Frame-Rate Cap Can Improve Efficiency

An uncapped GPU may render far more frames than your display or your preferences require. If your system is producing 240 FPS in a game where you’re perfectly happy at 120 FPS, the extra work can increase power consumption, heat and fan noise without meaningfully improving your experience.

A sensible frame cap can be one of the easiest efficiency improvements available. It’s particularly useful in menus, loading screens and less demanding games where frame rates can climb unnecessarily high.

Choose a target that suits:

• Your monitor’s refresh rate
• The game’s responsiveness requirements
• Variable refresh rate support
• The minimum frame rate your GPU can sustain consistently

A competitive player may reasonably value every reduction in latency. Someone playing a turn-based strategy or narrative adventure probably won’t benefit from pushing the GPU to its limit at all times.

Efficiency is about meeting your actual performance target with the least unnecessary work—not forcing every game into the lowest possible power state.

Do DLSS and FSR Reduce GPU Power?

Upscaling technologies render a game internally at a lower resolution and reconstruct a higher-resolution output. This can raise frame rates or reduce the rendering workload required to reach a chosen performance target.

However, simply enabling upscaling doesn’t guarantee lower power consumption. If the GPU uses the lighter workload to produce as many additional frames as possible, it may continue operating near its power limit.

The largest efficiency benefit often appears when you combine upscaling with a frame cap. The GPU can then reach the target frame rate while doing less rendering work.

The visual and performance trade-offs vary by game, resolution and mode. Our FSR versus native resolution guide examines those differences, while the guide to DLSS settings for competitive games focuses on high-frame-rate play.

Does Frame Generation Count in FPS per Watt?

You can calculate FPS per watt with frame generation enabled, but the result needs a clear label.

Generated frames aren’t equivalent to conventionally rendered frames. They can make motion look smoother, but they don’t update game logic or player input in the same way as newly rendered frames.

That means a card producing 160 displayed FPS with frame generation shouldn’t be presented as directly equivalent to one rendering 160 native frames per second.

For transparent comparisons, publish separate figures such as:

• Native raster FPS per watt
• Upscaled FPS per watt
• Displayed FPS per watt with frame generation
• Base rendered frame rate before generation

This avoids giving an AI-assisted result an unfair advantage without explaining what the number represents.

Can Undervolting Make a GPU More Efficient?

Undervolting attempts to run a graphics card at a lower voltage while retaining most or all of its performance. When successful, it can reduce power use, temperature and noise.

The result depends on the individual GPU, however. Two chips of the same model may not remain stable at identical settings. Driver changes, demanding games and long sessions can also expose instability that a short benchmark misses.

Anyone experimenting with an undervolt should:

• Change settings gradually
• Test several demanding games
• Watch for crashes, visual corruption and driver resets
• Record the original settings
• Avoid copying an aggressive value without testing it

A lower power limit is often simpler. It may reduce peak performance slightly while improving performance per watt, especially near the upper end of a GPU’s voltage and frequency range.

Neither technique is mandatory. Choosing efficient components from the beginning is safer and easier for most people. Our guide to energy-efficient gaming PC parts looks at the entire build rather than the graphics card in isolation.

Does a More Efficient GPU Always Save Money?

No. Electricity is only one part of the ownership cost, and usually not the largest.

Imagine one card costs 100 more but saves 15 per year in electricity. It would take more than six years for the energy savings to recover the higher purchase price, assuming everything else remained equal.

That doesn’t make efficiency irrelevant. Lower consumption may also mean:

• Less heat entering the room
• Quieter cooling
• A less expensive power supply
• Better suitability for compact cases
• Longer operation on a battery backup or portable power system
• Less demand on household electricity during long sessions

But it does mean that “uses fewer watts” and “costs less overall” aren’t interchangeable claims.

Use the calculator to compare the purchase-price difference with the estimated lifetime electricity saving. In many cases, buying the less expensive card and applying a frame cap will be more economical than paying a large premium for a modest efficiency improvement.

Efficiency Depends on the Performance You Need

The most efficient card on a benchmark chart may not be the most appropriate one for your system.

A GPU that performs brilliantly at 1080p may lack enough memory or processing power for your preferred 4K settings. A high-end model may produce excellent performance per watt while consuming far more total electricity than a modest card that already meets your needs.

Start with a practical target:

• Resolution
• Desired graphics quality
• Minimum acceptable frame rate
• Monitor refresh rate
• Games you play most often
• Features such as ray tracing or frame generation

Then look for the card that reaches that target efficiently and at a sensible purchase price.

Buying performance you won’t use is rarely efficient, regardless of how impressive the benchmark result looks.

How to Build Your Own GPU Efficiency Comparison

You can create a useful personal shortlist in a spreadsheet with the following columns:

1. GPU model
2. Purchase price in your region
3. Average FPS at your chosen settings
4. Average measured gaming power
5. FPS per watt
6. Price per FPS
7. Estimated annual electricity cost
8. Total ownership cost
9. Total ownership cost per FPS

Use current local prices rather than launch prices. A technically efficient card can be poor value when stock is scarce or retailer pricing is inflated.

Also record the benchmark source and date. GPU drivers, game patches and pricing change, so a transparent comparison should show when its inputs were collected.

GPU Efficiency Calculator Methodology

The electricity calculator uses the following assumptions:

• One year contains 52 weeks.
• One kilowatt equals 1,000 watts.
• The entered power figure is an average while gaming.
• Gaming time and electricity prices remain constant.
• The result covers the GPU only unless whole-system power is entered.

The FPS-per-watt calculator divides average frame rate by average measured power. It doesn’t adjust for image quality, latency, frame-time consistency, ray tracing, upscaling or generated frames. Those factors must be held constant or disclosed separately.

The calculator is an estimator, not a laboratory measurement. Its accuracy depends on the quality of the figures entered.

Frequently Asked Questions

Is a higher FPS-per-watt number better?

Yes. A higher result means the GPU produced more frames for each watt consumed under the stated test conditions. It doesn’t automatically mean the card is faster, cheaper or better suited to your games.

Does a 300-watt GPU always consume 300 watts?

No. A published 300-watt board-power figure describes a design or operating limit, not constant consumption. Actual power changes with the game, settings, frame rate, card model and system conditions.

Should I use TDP in the calculator?

You can use the manufacturer’s power specification for a rough upper-level estimate, but measured average gaming power will produce a more realistic result.

Does the calculator include the CPU and monitor?

Only if you enter whole-system or complete-setup power. If you enter GPU board power, the result represents the graphics card alone.

Does a higher-resolution monitor increase GPU electricity use?

It often increases GPU workload because the card must render more pixels, but the exact effect depends on the game, settings, frame cap and whether the GPU was already operating near its limit.

Can a faster GPU use less electricity?

Yes. A newer or more efficient GPU may finish a fixed workload faster or reach a capped frame rate at lower utilisation. But an uncapped faster card may also use its available power to render additional frames.

Is 60 FPS more efficient than 144 FPS?

Usually, when all other settings remain equal and the frame rate is deliberately capped. Producing fewer frames reduces the required work, although the size of the saving varies between systems and games.

Which GPU currently has the best performance per watt?

There isn’t one universal answer. Rankings change with the resolution, game selection, ray-tracing settings, upscaling, power measurement and new hardware releases. Use a consistent current benchmark suite rather than relying on a permanent winner.

The Bottom Line

GPU efficiency isn’t simply a matter of buying the card with the lowest wattage. The useful question is how much performance a card delivers for its power, purchase price and total running cost.

Measure or source realistic gaming consumption, compare cards under identical conditions and calculate whether any electricity saving is large enough to offset a higher price.

Most importantly, choose a GPU for the performance you’ll actually use. A sensible frame-rate target, efficient settings and a well-balanced PC can save more energy than buying excessive performance and allowing it to run uncapped.