Graphics cards aka GPUs play a massive role in computing. For video editing, graphics cards are often overlooked as it’s seen that the CPU does all the work. However, you may find that this isn’t the case.
If you’re looking for a graphics card for video editing, check out the article: Best Graphics Card For Video Editing 2020!
- 1 What Is A Graphics Card?
- 2 GPU VS CPU For Video Editing
- 3 Can SLI Boost Performance
- 4 AMD VS NVIDIA Video Editing
- 5 How Does GPU Acceleration Benefit Me In Video Editing
- 6 What Are Cuda Cores?
- 7 What are Tensor Cores?
- 8 So What GPUs Are Good For Video Editing?
- 9 Can The CPU Bottleneck The GPU?
- 10 What Are The Quadro GPUs?
- 11 Does This All Mean I Need A GPU?
What Is A Graphics Card?
A graphics card or graphics processing unit is responsible for producing the image you see on your screen. The faster the graphics card, the smoother the image on the screen can be produced.
Graphics cards are popular among graphically demanding applications, and are used as an enhancing module.
Graphics cards serve many purposes, not just gaming, but for 3D Modelling, and Video Editing. When working on high resolution projects (2K & 4k), the extra horsepower a GPU can offer could make a difference. This is because editing at high resolutions can partially use a GPU for faster results.
GPU VS CPU For Video Editing
It will be best to purchase a better CPU over a better GPU, this is because Video editing is heavily multi-threaded meaning more CPU cores is king.
Multi-threaded applications can benefit a lot from hyper-threading, this is because Multi-Threaded applications are coded with the intent to use multiple threads at the same time for different tasks. This post will provide further detail on why CPU is king.
However, GPUs are very useful when it comes to graphically demanding tasks such as: After Effects, 3D Modelling, and Video editing from Premiere pro. The CPU does the heavy lifting whilst the GPU comes in handy when the graphically intensive tasks arise.
Can SLI Boost Performance
SLI is when two NVIDIA GPUs work simultaneously. This may lead you to believe that SLI means double performance for everything, This is not true. SLI will boost performance if the application being used is SLI Compatible.
If the application is SLI compatible, the graphically demanding tasks may benefit from two GPUs, but you will not get double the performance.
Truthfully, it would be less risky to invest in one powerful CPU and one powerful GPU. SLI has always been inefficient and quite expensive, so I do not recommend it for Video Editing.
Remember, SLI is mostly for gaming, not editing so you will expect little to no difference with an extra graphics card.
AMD VS NVIDIA Video Editing
NVIDIA has the advantage of having Cuda cores which AMD does not have, AMD has Stream Processors. The use of these Cores will take some of the burden of the CPU when GPU intensive tasks are being operated. This is called GPU acceleration.
However, what about OpenCL? OpenCL could be seen as a Cuda core alternative, keep in mind NVIDIA also has OpenGL support.
AMD cards are typically better than NVIDIA cards at OpenGL, however, NVIDIA’s Cuda is faster and OpenGL doesn’t seem to provide the same performance boost Cuda seems to offer.
What if OpenGL is my only option? OpenGL is better than no OpenGL. Even though it isn’t as powerful as Cuda, it still is a viable method of GPU acceleration. It will offer performance boosts, and the OpenGL framework is very viable.
If you’re looking for a decent video editing GPU, Check out this article about the best RTX 2070 SUPERS.
How Does GPU Acceleration Benefit Me In Video Editing
GPU Acceleration is a technique where the GPU aids the CPU by moving CPU intensive applications to the GPU. This allows the GPU to be a useful addition to the computer and overall boosting speed and efficiency of an application.
GPU Acceleration has been used extensively in video editing to improve the performance of rendering at higher resolutions. It is always best to pair a good GPU with a good CPU for a balance and optimized performance.
To enable GPU acceleration, you are required to activate it in the video editing software. Be wary, not all GPUs are able to be accelerated depending on the software being used.
This video explains how to enable GPU acceleration in Premiere Pro 2019. If you’re interested, definitely take a look.
What Are Cuda Cores?
Just like a CPU has cores: Dual, Quad, Hexa, well so do GPUs, except they have several thousand cores. Cuda cores proprietary to NVIDIA and they are similar to stream processors which is proprietary to AMD.
When picking a GPU, the more Cuda cores usually indicate a faster card, but this is not always true. For example, 1000 Cuda cores clocked at 1000 MHz is no different from 2000 Cuda cores clocked at 500 MHz.
You may have noticed that GPUs are clocked slower than CPUs. This is because CPUs and GPUs are fed different data. CPUs are built to perform more complex tasks and operations, whereas the GPU is built to perform simple tasks.
Cuda cores and Stream processors are similar, but they’re not equal. Cuda cores and stream processors and not equivalent in terms of power and numbers. For example, 1000 Cuda cores is not equal to 1000 Stream processors. This is due to a different GPU architecture, and we cannot possibly compare the two.
What are Tensor Cores?
Tensor cores are also proprietary to NVIDIA. These cores are different from Cuda cores. They both work alongside each other, but both have very different purposes. Tensor cores are exclusive to the new RTX cards.
Tensor cores are new to the GPU market, and the purpose they serve is to do one type of task and that is ray tracing. Since this is a fairly new technology, we are yet to see it be used in video editing.
So What GPUs Are Good For Video Editing?
What you should look for when buying a GPU for Video Editing is its “Cuda” Cores. The Cuda cores are the Cores of an NVIDIA GPU, and the more you have, usually the better performance you can get out of it.
You should also pay attention to the clock speed of the GPU. This tells us how fast the GPU cores are, and is an overall indication of how fast a GPU is.
When the workload is divided between the GPU and the CPU, the CPU will have more free resources to focus on more demanding tasks it needs to complete
Only GPU intensive tasks can benefit from the use of Cuda cores. The GPU cannot help the CPU Encoding And Decoding. This task is extremely CPU intensive and a powerful GPU has no effect on the time it takes to complete this task.
Can The CPU Bottleneck The GPU?
It is important to maintain a clever balance between CPU power and GPU power. If one is vastly more powerful than the other, Then it will get limited by the weaker hardware. For example, if you pair an extremely powerful CPU with a slow GPU, then the CPU will be limited by how much the GPU can push.
Bottlenecking can happen in video editing just as much as it happens in gaming, and the consequences are not pretty. Imagine purchasing a super-powerful CPU for it to be limited by a really slow GPU, that would be a waste of money.
Keep in mind, a bottleneck like this will only occur when a GPU intensive task arises and there is an imbalance.
What Are The Quadro GPUs?
Quadro Graphics cards are workstation graphics cards used for professional needs such as Video Editing and Computer-generated imagery.
These graphics cards are actually more precise than the well-known GTX graphics cards. This is because GTX graphics cards are made for gaming, and they use integer precision engines. Quadro GPUs on the other hand use floating-point precision engines which is more precise than integer-based engines.
GTX graphics cards are built with the intent to produce frames as fast as possible, This is what makes it such a great gaming graphics card.
3.7435893405843 is more precise than 3.7. This is what Quadro specializes in. It needs to be super precise so it can perform calculations, and draw images to a higher precision.
In certain circumstances, a slight discrepancy could cause a lot of problems and cost a lot of money, that is why Quadro cards are popular among big industries.
Additionally, Quadro GPUs offer boosted Cuda cores, this is because they need to offer as much power as possible to make 3D modeling and video editing a much faster process.
These graphics cards can be extremely beneficial for Video editing, however, they come at a hefty price tag ($5000-$10000). These cards target large industries such as large content production studios.
Does This All Mean I Need A GPU?
We have learned that GPUs are actually important for video editing when it comes to GPU intensive tasks. This is because of GPU cores (Cuda cores or Stream processors), they distribute the workload between the CPU and the GPU to free up resources.
It is probably best to invest in a mid-tier GPU, you don’t need anything super powerful. But if you don’t have a GPU, remember most video editing is heavily CPU based. Remember to not cause a bottleneck, and to always maintain a smart balance between CPU power and GPU power.
For entry-level video editing @1080P, a (GTX 1060 x i5 8600K) will do the job. @2K&4K you should consider purchasing a (GTX 1070 x i7 8700K) or a (GTX 1080TI x i9 9900K).