▷ Gpu or graphics card? when should we use each term

The graphics processing unit (GPU) is one of the indispensable parts for the correct operation and use of a computer, especially if this use is focused on tasks based on the rendering of graphics, either for work or for simple interactive leisure.

Being such a required part in all equipment, these processors are in various formats and a large number of models, which in turn have different capacities and power.

Of the available formats, one of the most popular among users is dedicated or discrete graphics processors (Discrete GPU), also known as graphics cards. A component that we traditionally find on PC and among which are the most powerful graphic processors aimed at the general public.

Given its popularity and its obligatory use, it is common in a colloquial conversation to use the term GPU interchangeably to refer to the graphics processor from which the acronyms arise or to the graphics card that constitutes a component by itself, but as we can imagine this use is not the everything correct. Today we want to define in a more technical way what these terms refer to and why we use them incorrectly despite the fact that this use is accepted by the entire community.

Index of contents

The importance of defining each term

Language within the hardware world is highly technical due to the need to correctly identify and differentiate the many elements on a computer. This moves these terms away from the normal use of language that we can have in a colloquial conversation, so to know the true difference between two terms as twinned as GPU and graphics card, it is essential to delve into different names.

GPU Vega 20 (Image: Fritzchens Fritz)

Thus, the graphics processing unit (GPU) is a processor specially designed for operations and calculations related to rendering 3D graphics. Their specialization makes them very efficient in solving operations and algorithms that the team's main processor cannot optimally handle. An example of these operations could be the floating point calculation.

They are used in a wide variety of systems and usually come in two formats: discrete, which has already been mentioned, and integrated. Depending on whether they are part of some of the components of the equipment (usually the processor) or if they have hardware dedicated exclusively to the graphics processor.

Asus graphics card

On the other hand, the graphics card, or graphics accelerator card, is an independent piece of hardware that integrates the graphic processing unit on the same PCB together with elements necessary for its operation, such as memories (GRAM), converters (RAMDAC) and connectors and drivers.

It is therefore in charge of generating the video output of the equipment in which they are connected, so they are an expansion of the motherboard, to which they are connected through an interface to communicate with the rest of the elements of the equipment. Currently, the most widespread interface is PCI-express and it is the one that you probably use on your computer to connect the card.

Why do we use both terms interchangeably?

Graphics cards are not the most widely used format in which we find graphic processing units. According to the JPR information agency, more than 70% of graphic processors used in PCs are integrated GPUs, with the PC being the platform where the vast majority of dedicated graphics proliferate.

WE RECOMMEND YOU what this is and what it is used for in computing

Knowing this, we can ask ourselves why both terms are used without attending to their real denomination. The answer to this can be found in the attention aroused by graphics cards and the names with which they baptize graphic processors.

As the most powerful GPUs in the consumer market are normally found in graphics cards, the user community (Mainly gamers) who acquire these components address them as GPUs or graphics cards, since one encompasses the other.

GF100 processor of a GTX 470 (Image: Fritzchens Fritz)

Added to this is the cumbersome name of the graphics chips themselves. For example, the GPU of an RTX 2080 is TU104 (Specifically TU104-400A-A1) which in turn is the same as the RTX 2070 Super (In this case, TU104-410A-A1), so calling the GPU the RTX 2080 we would actually be referring also to the card of the directly lower range belonging to a different series, but obviously a name as obtuse as "TU104" and its variants is much less fluid than naming them by their commercial name.

Final words and conclusion about GPU or graphics card

Therefore, while the interchangeable use of the terms GPU and graphics card to refer to integrated and dedicated processors is not technically correct because the graphics card is a separate component that integrates a GPU.

Ter we recommend reading the best graphics cards on the market

It is so settled in the community and it is so inconvenient to learn the real names of these products that at the moment of truth, any user will understand you use the term you use to refer to our beloved graphics co-processors.