People use laptops in all sorts of weird places. Part of a laptop’s appeal is portability and versatility, but you might notice that after a bedtime binge session, your sheets, or your lap, is pretty warm.
Modern computers, especially laptops, are built with a number of stopgaps between firing up a TV show and shutting down. You’d likely give up using your laptop because programs aren’t working long before it crashes. It keeps the temperature in check through a combination of fan speed control, power management, and throttling.
These techniques do a decent job keeping your laptop from melting, but how do they affect the user experience? When your next evening HBO GO session gets steamy, you’ll know exactly when to crack a window and cool off.
The test procedure
The first test we used was 7-Zip, a common utility for packing and unpacking .zip files. It has a built-in benchmarking utility that compresses and decompresses data which resembles an assortment of files. It places constant load on the processor, forcing it to fire up and run as quickly as it can.
The second utility is Futuremark’s 3DMark program. It’s a series of four graphical tests – short sequences that are increasingly difficult to render. These strain most components of a laptop, and of course place heavy load on the GPU, which is often the hottest chip in a modern PC.
To see how these tests impacted performance in different environments we ran them first with each laptop on a flat, solid desk, with no obstructions around it to prevent venting hot air or pulling in cool air.
Then we used a sleeping bag as a proxy for a situation where your laptop might not be able to vent well – like a bed, or your lap. We spread the sleeping bag out flat on a table, then set the laptop on top.
We used RealTemp to keep a time-stamped log of each core’s temperature, the GPU’s temperature, and the overall clock speed. In some cases, it was able to read the processor’s reported maximum temperature, and would also track if a core was OK or Hot. We tracked all of these statistics throughout both tests so we could chart the data and identify trends.
What happened, in a nutshell
Three laptops were tested; the Alienware 15, the Lenovo ThinkPad T450s, and the Asus UX305. Detailed results can be found in the links below, but if you want the summary, here’s what happened.
Surprisingly, the laptops did not experience significant slowdown when placed on the sleeping bag. Performance was identical more often than not. None of the laptops crashed or turned off automatically, not even when we ran the 3DMark benchmark to engage both the processor and the graphics chip.
That means system temperature does not significantly impact speed. Buying a laptop cooler under the theory that it may improve performance or stability seems a waste of money, and you can comfortably use a modern laptop in sub-optimal conditions without concern that it’ll overheat and reboot (or worse).
With that said, using the laptops on a sleeping bag did drastically increase external temperatures. At the worst, as with the Alienware 15, they shot high enough to be potentially hazardous. At best, they were high enough to be very uncomfortable.
Using a laptop on a surface that retains heat, or blocks the intake of cool air, will give you a case of sweaty palms. Some might argue it will decrease long-term durability — though that’s outside the scope of this test. But it won’t make your system slow to a crawl.
Next page: Alienware 15 testing
The Alienware 15 is a fast, heavy gaming laptop that’s more of a desktop replacement than a portable rig. Our test unit came with an Intel Core i5-4210H, which is a dual-core processor with a base clock of 2.9GHz, 16GB of RAM, and a GTX 970M.
While sitting on a desk, 7-Zip caused the processor initially to spike to about 75 degrees, but then the fans kicked in and the system kept a cooler 68-degree average for extended periods. True to expectations, the clock speed stayed constant throughout the test, and spiked for short periods when a little extra horse power was needed.
Nestled warmly in our sleeping bag, the cores hit their hot point of 80 and never dropped below 75. Despite that, the effects of heat on the Alienware’s processor weren’t striking. In fact, it didn’t appear that during our testing, the heat had any effect on the processor’s overall clock speed. The extra heat didn’t slow down the average clock speed of the processors as much as it limited its use of extra features like Turbo Boost.
The 3DMark tests told a similar story of the effects of heat on performance. On the desk, the system was able to frequently boost the total clock speed up to 3.4GHz, which it was able to do because it had room in its thermal overhead.
With the system on the bag, the processor held firm during the 3DMark test at its base clock of 2.9GHz after an initial high point – the Turbo Boost feature basically turns off. When it was cooler, the laptop could gun it to load a scene, then back off to handle the smaller pieces of the test without overheating or needing Turbo Boost.
Placing the Alienware on the bag had an adverse impact on external temperatures. The maximum external reading shot up to 143 degrees Celsius, more than hot enough to burn a user through prolonged contact. Presumably, you’d move the laptop if that became an issue, but it’s worth note. Sitting a hot laptop on your legs or your hands might become a health issue should it remain in place long enough.
While sitting on the desk, the GPU never broke 70 degrees Celsius, a very comfortable operating range for the chip. While on the bag, the core’s temperature only went below 70 once, and very briefly. It hovered around 75 or 80, pushing over 85 a few times.
The 3DMark results showed a direct correlation between heat and gaming performance. With proper cooling, the Alienware scored 5,768 on the First Strike test. On the sleeping bag, the score dropped to 5,106. In the 3DMark Sky Diver test, the Alienware scored a 14,160 on the desk, which came down to 12,043 while hot. Cloud Gate dropped from 11,647 to 9,342 without room to breathe, and Ice Storm went from 63,914 to 61,645.
That’s an average 12.5 percent drop in performance across all four tests. Even if you don’t use your laptop for gaming, that’s a noticeable dip in the stability and speed of your system. If you do use it for gaming, you’ll see that change in lower framerates and longer load times.
Next page: ThinkPad T450s testing