How we test desktops

Desktops have certainly been around the block, but that hasn’t made them easier to understand for the average consumer. Here at Digital Trends we write reviews that de-mystify the PC and enlighten anyone who cares to read them. This requires subjective impressions and objective testing.

Reviewing a desktop is a unique process because there are concerns relevant to them that aren’t relevant to devices like laptops, tablets or smartphones. Let’s dive in to our testing procedure.

Also make sure to check out the best desktops as rated by our reviewers. 

Look and feel

A desktop computer usually is placed in a designated spot and stays there, but this does not mean build quality is irrelevant. Consumers long ago grew tired of black and tan mid-towers. Today’s desktops come in all shapes and sizes for all sorts of purposes.

We take a careful look at the dimensions of the systems we review so that you know how the PC will fit into your office, home theater or living room. Basic functionality and build quality are key, as well. A good desktop needs to have easily accessible ports. Some products come with panels that cover their optical drives or flip-up front-facing ports. If these exist, they must have a robust feel and simple operation.

Aesthetics are particularly subjective, but we do provide some commentary on them — and a nice selection of high-quality photos. There’s nothing wrong with a little vanity. Desktops are often placed in public areas of a home, such as the living room, and they need to at least look as if they were worth the price paid.

What’s inside counts, too

There’s more to a desktop than its external appearance and features, however. Much more. Desktops, unlike other modern electronics, are still partially or entirely user serviceable. Processors can be replaced, graphics cards upgraded, and hard drives added. But this is only an advantage if a system is easy to work on.

maingear vybe interior fan

We open up all the desktops we review and report on the guts. A spacious layout that is free of clutter from various cords is ideal. Small desktops have difficulty because not enough space available, but we like to see that the manufacturer gave at least a second of thought to making the internals easy to work with.

Tool-free design features and readily accessible expandability are also a boon. We prefer case panels that pop off with a latch system, hard drives that can be exchanged without hassling with screws and RAM slots that are free for an upgrade. Even small systems can make hard drive swaps a breeze — if they’re properly designed.

The extra mile

Most computers see success because of a combination of great hardware and reasonable pricing, but there is sometimes more to it than that. Software can be both a hazard and a boon, depending on how it’s implemented.

Bloatware can detract from the user’s experience, but systems such as the HP TouchSmart include free apps that make the computer easier to use. Whatever the case, we never remove pre-installed software (no matter how annoying) during our reviews. If we can’t live with it during our testing we doubt you can live with it every day.

Warranties and peripherals can also be a defining trait. If all other things are equal a system with a 3-year warranty will be a better buy than one with a 1-year warranty. Manufacturers are notoriously stingy with warranty periods, so we always comment when a system offers exceptional service at no additional cost.

The test bench

Performance is obviously a concern for anyone buying a computer. We put all desktops through a number of benchmarks that provide a way to objectively gauge performance.

Our general CPU benchmarks include SiSoft Sandra’s Processor Arithmetic test and 7-Zip, a compression program that’s highly optimized for multiple threads. These tests put the processor on maximum attack and give us an idea how a system performs in demanding applications.

After this is completed we run PCMark 7, a comprehensive test suite designed to task every component including the memory and the hard drive. Because this benchmark is so comprehensive it sometimes provides a perspective you can’t find elsewhere. For example, computers with solid state drives do well because they absolutely demolish mechanical drives in the system storage portion of the benchmark.

We wrap up our suite with 3DMark’s Cloud Gate and Fire Strike tests, two synthetic benchmarks that offer a good window into overall graphics performance. If a system is built for gaming we also conduct real-world game tests with titles like SkyrimDeus Ex: Human Revolution and Civilization V. We report the average frames-per-second, as recorded with FRAPS.


Desktops usually do not become as warm as other, smaller PCs, but cooling must still be considered. Some systems are quieter than others, and some offer better cooling of internal components. This is of particular concern for gaming systems, which can become extremely loud and hot during intense use.

digital storm bolt ii cpu fans

To judge how a system is to live with we use a decibel meter and take a measurement of noise from two feet away from the front of the desktop. A reading is taken at idle, during the 7-Zip processor test, and after a ten-minute run of the Furmark graphical stress test. This provides a broad idea of how loud the desktop is under various conditions.

We also take a look at CPU and GPU temperatures at load. While we don’t normally report these numbers, we will list them in a review if either figure seems abnormally high. We consider a CPU temperature above 70 degrees Celsius, or a GPU temperature above 95 degrees Celsius, to be unusually warm.

Putting the pieces together

Desktops are complex products that consist of numerous components. This makes the verdict important, but also difficult to reach. There’s no simple 1+1 formula that can be applied. Every component has an effect on every other.

With that said, we usually reach our verdict through the lens of the desktop’s intended purpose. A budget PC meant for basic computing will not severely penalized if it doesn’t ace our benchmarks, but a high-end workstation with the same failing would be nearly worthless. We try to put ourselves in the shoes of the person who is likely to be buying the system and ask if the computer does what is expected of it — and more.

Pricing and competition are obvious concerns, as well. Our verdicts usually include some market research into the price of similar desktops and their value compared to the system being reviewed. A low price can forgive some — but not all — shortcomings.

Our review process is meant to provide a broad platform on which any given desktop can be judged. We don’t expect every reader to agree with every verdict, but we do hope that you leave more informed than when you arrived.

Get our Top Stories delivered to your inbox: