Discrete GPU vs Dedicated GPU – What’s The Difference?
Graphics units in PCs and laptops can be separated into one of two camps depending on whether they are (I) integrated into the core processing unit (CPU) or (II) independent from the CPU.
- Integrated GPUs (also sometimes called onboard GPUs) are types of GPUs that share the same chip as the CPU. Integrated GPUs are known for their weaker performance when applied to graphic intensive tasks.
- Independent GPUs (commonly known as graphics cards, standalone GPUs, discrete GPUs or dedicated GPUs) are separate from the CPU and are able to achieve higher performance thanks to having their own power and data connection.
As illustrated above, there are often multiple names given to the same computer component, and that is the case for discrete and dedicated GPUs.
There is no difference between a discrete GPU and a dedicated GPU as they are both terms for the same component, namely a standalone graphics card which is independent from the CPU.
Whilst there might be no difference between a discrete and dedicated GPU, there most certainly is between an integrated GPU and one which is independent. If you are sizing up the specs of a new computer this is something which you should most definitely be aware of.
This article will go through the most important parameters in which integrated GPUs and independent GPUs differ from each other, and provide some brief guidance on how to choose the best option for your intended use.
Dedicated GPUs perform better than integrated GPUs
Performance is the most important parameter to explore when choosing between a dedicated GPU and an integrated GPU. In fact, for the majority of the people looking to decide between the two types of GPU, performance will be the final deciding factor.
Generally speaking, a dedicated GPU will always beat an integrated GPU in graphic-intensive tasks. This is because, as the name suggests, a dedicated GPU is designed to maximize performance. A dedicated GPU can have an independent power draw, a dedicated cooling solution, and it can overall not be limited by the CPU.
And this is the major drawback of an integrated GPU.
Since it shares the chip with a CPU, it shares the power and cooling with the CPU.
Simply put, if you place a CPU and a GPU on the same chip, you will be limiting the potential processing strength of one of those components. And since the CPU is more important for the smooth functioning of a computer, the GPU is the one that is underpowered.
Independent power supply and cooling solutions aren’t the only things that hold back integrated GPUs.
Since dedicated GPUs are separate cards with their own PCB. It also has its own memory. This is called Video RAM or VRAM.
It works the same as regular RAM, but instead of storing all kinds of data, VRAM only handles graphics and video-related data to increase performance, and is crucial for gaming.
While it is an undisputed fact that dedicated GPUs are better performing, it doesn’t mean integrated GPUs are completely useless.
In fact, in the past few years, we have seen integrated GPUs getting better and better. So not only are they perfect for basic computing needs, some integrated GPUs by the likes of Intel and AMD are capable of light gaming as well.
This makes integrated GPUs an excellent choice for laptops where space and power are often limited. But integrated GPUs aren’t just limited to portable computing.
Almost all CPUs these days come with integrated GPUs so that out of the box every computer has graphic processing ability, and the user has a choice to upgrade to a more dedicated option if the need arises.
Dedicated GPUs draw more power than integrated GPUs
When talking about GPUs and performance, we can’t go far without talking about power consumption.
Just like with any device, the more power it draws, the more performance it can output. In exchange for that extra power consumed, we see results in terms of better performance across graphics-intensive tasks.
But just because an integrated GPU uses less power doesn’t mean it is more efficient. In fact, most of the time, if we look at it in terms of a trade off between power consumed vs. performance, dedicated GPUs come out on top. This shouldn’t come as a surprise since dedicated GPUs are optimized to perform well in all aspects of the graphics department.
Dedicated GPUs also draw more power because they have cooling solutions to operate. If you have a dedicated graphics card installed on your PC, you will see that it has its own heatsink and multiple cooling fans.
This allows the dedicated GPU to draw more power without overheating.
Integrated and dedicated GPUs have totally different cooling solutions
Cooling is a pretty huge limiting factor for power draw and performance. If you want better graphical performance, you need more power draw. And if you want more power draw, your chipset needs to be adequately cooled. Otherwise, you will thermally throttle your system.
Thermal throttling is when a component gets so hot that the computer automatically turns it down to save the component from damaging itself and the computer. Electronics generate heat; there is no way around that. This is why high-performance computers and components have a better cooling solution. So, they can perform better without heating up.
Integrated and dedicated GPUs have totally different cooling solutions.
As a dedicated GPU has a separate PCB, it has its own heat sinks and its own cooling fans. This allows the GPU to work at higher power levels.
Since an integrated GPU shares the chip with the CPU, that means it also shares the CPU cooler. But since CPUs are more important for the proper functioning of a computer, CPU load is given precedence over GPU load. This is why most of the time, integrated GPUs are underpowered, so that they don’t use a lot of power and thus generate a lot of heat that the CPU cooler will eventually have to dissipate.
Dedicated GPUs are larger than integrated GPUs
In terms of size, it should be pretty obvious by now that dedicated GPUs are larger (since they are separate chips on separate PCB with separate cooling) and take up more space vs integrated GPUs.
For desktops, the GPU plugs into the PCIe slot on the motherboard and depending on your exact model, the GPUs can be large and heavy when compared to other PC components.
Case in point being the latest RTX 3000 series GPUs by Nvidia.
This isn’t a cause for worry however as all components including the latest motherboards and PC cases are designed with the latest GPUs in mind. So, they too have adequate space and support for dedicated GPUs.
Integrated GPUs, on the other hand, are built right on the same chips as the CPU so they don’t take extra space. This is why as previously mentioned they are an excellent choice for laptops.
Speaking about laptops, another important thing to keep in mind is that dedicated GPUs inside laptops and the ones that are sold for desktops aren’t the same in terms of performance.
Since laptops have space, weight, cooling, and power limitations. Dedicated GPUs inside laptops are smaller and often underpowered as compared to their desktop counterparts.
So, a desktop dedicated RTX 3070 will always be more powerful as compared to a laptop with a dedicated RTX 3070. And due to size limitations, the GPU will be soldered on the motherboard. So you won’t have the freedom to change out your dedicated GPU card inside laptops.
Integrated GPUs are the lower cost option
The last aspect where an integrated and dedicated GPU differ is in price.
It’s a simple equation to solve however as an integrated GPU comes as a combo package with a CPU on the same chip. This makes integrated GPUs the cheaper option.
A dedicated GPU on the other hand is a separate chip on a standalone PCB with its own cooling. So, more materials and processing go into making a dedicated GPU. In fact, for a desktop system, the GPU is the single most expensive component you can buy!
So now that we have demonstrated that there is no difference between a discrete GPU and a dedicated GPU how do you as a buyer figure out whether a discrete or a less powerful integrated GPU would work best in your setup? Well, it is pretty simple.
List what particular tasks you will be using your GPU for. Based on that, you will easily be able to decide what type of GPU works best for you.
Let’s say you just need a laptop for office work – so need to run things like Microsoft Office, e-mail and a web browser. In this case, an integrated GPU will work just fine. It will generally be cheaper, consume less power, and likely also a longer shelf life.
But suppose you are someone that is buying a laptop or a desktop to play video games, render videos, and run machine learning models.
In that case, a dedicated GPU is what you need. This is because all of these tasks require lots of graphical processing power and integrated GPUs simply won’t have the grunt to handle these tasks efficiently.