Do You Need A GPU? – 6 Real Uses

A graphics card/GPU is a powerful component and often seen as absolutely necessary with many PC builders, this is because it accelerates the graphical potential of your PC build. PC builds such as gaming PCs require graphics cards to process the graphical information you see on screen, you will often see very powerful GPUs in these PCs.

Many people will associate having a GPU with a gaming PC, so if you’re not building a gaming PC would you really need a graphics card? If you’re going to build a PC, you can save quite a bit of money by not purchasing a graphics card, so we’re going to go over whether a graphics card is a necessary component.

Answer: Having a dedicated GPU in many scenarios is actually optional, such scenarios include having a CPU with integrated graphics. If you’re just building a PC for work, then you can get away with an Intel or Ryzen CPU with integrated graphics and leave the dedicated GPU out. You will always need some sort of graphics processor or else you won’t see anything on the screen.

Certain CPUs May Require A Dedicated GPU

dedicated gpu

If you’re building a PC with an Intel process that isn’t an “F” variant, then technically you do not need a graphics card. Most Intel processors will actually come with an iGPU which can be useful for just displaying simple information on the screen. However, “F” variant Intel CPUs and many Ryzen CPUs do not come with integrated graphics, hence requiring the use of a dedicated GPU.

If your PC does not have an iGPU or a dedicated GPU, then your PC has no way of processing graphical information and displaying it to the screen. Everything you see on your display is the work of a graphics processing unit whether it’s integrated or not. So whether you’re doing simple work or gaming, a GPU is necessary.

If you do not have a PC with an integrated graphics card, you can pick up a cheap dedicated GPU with the purpose of just outputting a display to the screen. This is a cost-effective way of getting a graphics processing unit, you do not need the top-of-the-line GPU to display information to the screen. We recommend not spending over $200 for a budget dedicated GPU, if there’s no other option, you can always purchase second-hand.

GPUs Will Improve Graphical Performance

gpu

Having dedicated graphics will massively improve the performance of your PC from a graphical perspective, this will open many possibilities for your PC. This is due to many applications actually requiring the parallel processing the GPU can offer, a few applications are gaming, video editing, streaming, crypto mining, and machine learning.

PCs without dedicated GPUs will be limited to the performance of an iGPU which severely hinders the capabilities of your PC. Integrated graphics can never compare to a decent dedicated GPU as integrated graphics share resources with the CPU itself. A dedicated GPU actually has its own memory and own processing unit which makes it many times more powerful and an iGPU.

Integrated graphics cards aren’t completely helpless, they’re able to watch videos in up to 4K resolution, but it dwarfs in comparison to even a budget-end dedicated GPU. The strongest integrated GPU which is the Intel® UHD Graphics 770 competes with pretty weak GPUs such as the GT 1030 and still loses in graphical situations.

Video Editors May Require Graphics Cards

gpu for video editing

Video editing is seen as a CPU-intensive task often requiring many processing cores to be effective, but there are some video editing applications that require a dedicated GPU. Graphics cards aren’t there just for gaming, they’re used for high-resolution video editing with DaVinci Resolve for faster results. Graphics cards are also great for Adobe After Effects.

Software such as After Effects and Davinci Resolve will use the GPU to improve rendering times, this software will often still rely on the CPU, but the GPU is there to accelerate the performance. Many professional video editors will have decent graphics cards in their builds for this exact reason, not having a GPU is leaving a lot of performance on the table.

The free version of DaVinci Resolve will not use the graphics card to improve upon rendering times, so if you’re wondering why your GPU isn’t being used in DaVinci Resolve, this may be the issue. The paid version of DaVinci Resolve will offload a lot of the image processing onto the GPU which will help immensely with the rendering times. When picking up a graphics card for video editing, NVIDIA graphics cards are a pretty good option due to Cuda Cores.

Streaming PCs May Require A Graphics Card

streaming gpu

Most of the time, streaming is actually CPU intensive, you’ll often find this is the case if you’re just getting into streaming. Software such as OBS will use the x264 encoder which is actually CPU intensive, and it’ll use as many CPU cores as possible. X264 is a software encoder and you should see practically 0% GPU usage when it comes to streaming.

However, saying streaming is purely CPU intensive isn’t telling the whole truth, this is because hardware encoders for streaming will use the graphics card instead. Hardware encoders will purely utilize GPU power, and such hardware encoders are NVenc for NVIDIA, AMD VCE for AMD graphics cards, and QuickSync for integrated graphics cards.

Dedicated streaming PCs will also require some sort of graphics processing power whether it is integrated or dedicated, this is because it will require this power for compositing and displaying your webcam. You shouldn’t require a super-powerful graphics card for a dedicated streaming PC, you can probably get away with an integrated GPU, but a dedicated GPU will be able to do things much easier.

AI/Machine Learning Need A Dedicated GPU

gpu machine learning

When it comes to applications such as AI machine learning and deep learning, graphics cards are an extremely important component. This is because a discrete/dedicated GPU will be able to crush massive amounts of data in parallel faster than a CPU/Integrated GPU will be able to. The same graphics card you use in your gaming PC can make for a competent machine learning GPU.

GPUs are preferred over the CPU because of the GPU’s ability to perform multiple computations simultaneously, this speeds up processes such as AI training. CPUs are great pieces of hardware, but they’re not optimized in the way GPUs are when it comes to bandwidth. A good way to view this is that the CPU can move little pieces of data really fast, and the GPU can move massive amounts of data really slowly, this type of optimization is what makes GPUs best for machine learning.

The GPU can also perform matrix multiplications far faster than the CPU can, this is the main limit of machine learning. Essentially, machine learning has a bunch of tasks that can be parallelized, and the thousands of processing cores the GPU has can help with that. CPUs don’t come with thousands of cores, they usually have up to 16 cores max for top-of-the-line consumer CPUs.

Gamers Will Heavily Benefit From A GPU

The most obvious use for a graphics card is for gaming, the whole premise of gaming is being able to display graphical information on the screen whether it’s high quality or not. While it is technically possible to game on integrated graphics, the performance will be subpar and a dedicated GPU, even a budget one can make the experience many times better.

If you’re looking to play extremely graphically intensive games such as assassin’s creed, at max settings 4K, then you will require top-of-the-line graphics cards, NVIDIA and AMD often provide such graphics cards. The goal is to achieve a playable frame rate such as 60FPS, integrated graphics cards will struggle to get anywhere close to this number.

A comparable dedicated GPU to the Intel HD Graphics 770 is the GT 1030 and they will often show pretty similar average frame rates in games, especially Forza Horizon 5. But the dedicated graphics card will still provide a better gaming experience, this has something to do with 1% and 0.1% lows. The average frame rate isn’t good enough when gauging playability as 1% and 0.1% lows play a huge part in smoothness, and when they’re bad it’ll appear that the frame rate is much lower.

Conclusion

The verdict is that the GPU is an extremely versatile component that can help with many tasks such as machine learning, video editing, streaming, and gaming. Without a dedicated GPU, you limit your PC’s ability to process graphical intensive applications and tasks that benefit from parallel computing.

If you’re not going to use a dedicated graphics card, then you will require a processor that comes with an integrated GPU. Integrated GPUs are far slower than dedicated GPUs because they have to share resources with the processor. Integrated GPUs are best for processing simple graphical data, but they can also do just fine with light gaming. If your PC does not have an integrated or dedicated GPU, then your PC will have no way of displaying information on the screen.