Personal computing is no longer limited to spreadsheets and Word documents, and thanks to the immense processing power of modern GPUs, your computer can even do much more than play games. However, you may be wondering why desktop GPUs cost so much and whether discrete laptop GPUs offer better value for money. Clearly, there is a contrast in size and performance at play when it comes to answering the two questions, but how wide does this gap go?
Let’s take Nvidia’s GeForce RTX 5060 as an example, available in desktop and laptop versions. The laptop version has 13% fewer shader cores than the desktop version. The laptop’s chip also has a lower base clock speed of 1.46 GHz than the desktop 5060’s 2.28 GHz. These specification differences contribute to a 7-15% loss in production on paper, depending on what you look at. In the real world, however, you’ll find deltas of around 20% in graphics test scores. Why this additional loss of performance? In short, physics. A slightly longer answer would be size, power and heat.
Many materials break down when exposed to a certain level of heat. Silicon and other metals in PC components are subjected to high thermal stresses as they consume energy and convert it into operation. Processing inefficiencies generate latent heat that can accelerate failure and impact performance. So how to solve this problem?
Laptop GPUs face more challenges than desktop GPUs
Marvin Samuel Tolentino Pineda/Getty Images
Engineers in the computer industry mitigate heat in several ways, but with office equipment, air circulation comes first. A typical PC case has plenty of space for air to circulate, and it’s equipped with fans and places where you can install even more fans to bring in cool air and exhaust hot air. In a compressed form factor such as a laptop, this task becomes more difficult as parts are crammed close together and heat builds up. While fans play an important role here, you’ll also see insulation and other relief measures such as thermal sealant, silicone pads, and vapor-ducted heatsinks—useful, but less effective than more elaborate water-cooling systems for desktops.
Remember, you’re buying a laptop because you’re looking for a blend of power and portability. These values will require compromises: size matters and that means adjusting components to the new plans. It can also mean working with a lower thermal design power (TDP), where a certain power level produces acceptable operating temperatures. Coming back to the RTX 5060, the desktop chip is expected to consume up to 145 watts, while the laptop version is expected to run at 100 watts. Ultimately, a desktop GPU will almost always beat a discrete laptop GPU for gaming and compute-intensive uses, but it still beats using a laptop with an integrated GPU.
