filmov
tv
Is GPU Important for 4k Video Editing? | Rendering and Playback Benchmarks
Показать описание
Is GPU important for video editing? When it comes to rendering video footage the GPU works together with the CPU to provide smooth playback and fast rendering. ►⏩►⏩ Check the description below for my video editing laptop recommendations.
Is GPU important for video editing? In this video, I am going to discuss the role of the graphics processing unit in video editing as well as look at some benchmarks from Premiere Pro, video editing 4k footage.
The reason the GPU can handle this amount of visual information is due to the number of cores it contains:
GTX 1050 has 640 cores (768 cores for 1050 Ti)
GTX 1060 3GB has 1152 cores (1280 cores for 6GB version)
GTX 1070 has 1920 cores
GTX 1080 has 2560 cores (1080 Ti has 3584 cores)
So with a lower-end GPU containing 640 -770 cores. This allows the GPU to execute what is referred to as parallel processing through a technique NVIDIA developed called CUDA cores. So rather than a CPU having 6 main powerful cores that attack operations one at a time the GPU attacks many many many more computational tasks at once. More cores = more workers.
So basically the CPU is like the general contractor planning the building of a large Dam and the GPU is the 750 workers all working at the same time to get the Dam up in a fraction of the time it would take the G.C. if he tried to do it by himself. GPUs are best used on visual tasks which is why we see there effect so prominent in video editing.
Ok now that we have a better understanding of what the GPU is, what it is made of, and how it works let’s look at how it performs during video editing.
Before we dive into the tests if you are curious about my recommendations I have listed a few in the description below.
I am going to be using a Gigabyte Aero 15 (i7-9750H, 16GB RAM, RTX 2070 Max-Q)
I have a 9-minute 4k Premiere Pro project with talking head and B-Roll footage. With CUDA cores optimization enabled you can see that the 4k Playback is really smooth. As the timeline begins to play the CPU is sending a task to the GPU to render and display the 4k footage on the timeline. It is handling it exceptionally.
Ok, now I am going to show you what happens when we switch off the GPU. As you see the timeline nearly comes to a halt. With the CPU running at nearly full bore it can’t keep up with the task. There are too many mathematical calculations for the CPU to keep up with, something that the parallel CUDA cores in the GPU can handle with great ease.
Now let’s turn on the CUDA cores again and do a quick render test:
3:25 to render out 7240 frames with GPU support (CUDA cores) turned on
28:06 to render the entire 19-minute project with no GPU support
I also ran these same tests with my 2017 Dell XPS 15 (i7-7700HQ, 32GB RAM, GTX 1050) to provide an example of what an entry-level GPU can accomplish regarding 4k timeline playback and rendering.
With the CUDA cores turned on the 4k playback in the timeline is smooth, but as the CPU begins to heat up it begins to throttle. This causes the timeline to become jumpy and inconsistent. The throttle of the CPU decreases the CPUs ability to quickly send the GPU the information it needs to accomplish its task. As you see the GPU becomes underutilized as the CPU starts to throttle. When we first started the playback, before the CPU heated up, the GPU was bouncing around 65% - 80% utilization. But now that the CPU is throttle it sits much lower.
Ok, now let’s take a look at how this system handles rendering,
5:53 to render out 7240 frames with GPU support (CUDA cores) turned on
1 Hour + 15 minutes to render the entire 19-minute project with no GPU support
As you can see it is important to consider both components when purchasing a laptop or desktop computer for video editing. It is not enough to get either a powerful GPU or a powerful CPU. Having an equally powerful CPU and GPU will yield optimal performance out of each component.
------
DISCLAIMER: ALL LINKS IN DESCRIPTION ARE AFFILIATE LINKS
Thanks for Supporting Our Channel!
Is GPU important for video editing? In this video, I am going to discuss the role of the graphics processing unit in video editing as well as look at some benchmarks from Premiere Pro, video editing 4k footage.
The reason the GPU can handle this amount of visual information is due to the number of cores it contains:
GTX 1050 has 640 cores (768 cores for 1050 Ti)
GTX 1060 3GB has 1152 cores (1280 cores for 6GB version)
GTX 1070 has 1920 cores
GTX 1080 has 2560 cores (1080 Ti has 3584 cores)
So with a lower-end GPU containing 640 -770 cores. This allows the GPU to execute what is referred to as parallel processing through a technique NVIDIA developed called CUDA cores. So rather than a CPU having 6 main powerful cores that attack operations one at a time the GPU attacks many many many more computational tasks at once. More cores = more workers.
So basically the CPU is like the general contractor planning the building of a large Dam and the GPU is the 750 workers all working at the same time to get the Dam up in a fraction of the time it would take the G.C. if he tried to do it by himself. GPUs are best used on visual tasks which is why we see there effect so prominent in video editing.
Ok now that we have a better understanding of what the GPU is, what it is made of, and how it works let’s look at how it performs during video editing.
Before we dive into the tests if you are curious about my recommendations I have listed a few in the description below.
I am going to be using a Gigabyte Aero 15 (i7-9750H, 16GB RAM, RTX 2070 Max-Q)
I have a 9-minute 4k Premiere Pro project with talking head and B-Roll footage. With CUDA cores optimization enabled you can see that the 4k Playback is really smooth. As the timeline begins to play the CPU is sending a task to the GPU to render and display the 4k footage on the timeline. It is handling it exceptionally.
Ok, now I am going to show you what happens when we switch off the GPU. As you see the timeline nearly comes to a halt. With the CPU running at nearly full bore it can’t keep up with the task. There are too many mathematical calculations for the CPU to keep up with, something that the parallel CUDA cores in the GPU can handle with great ease.
Now let’s turn on the CUDA cores again and do a quick render test:
3:25 to render out 7240 frames with GPU support (CUDA cores) turned on
28:06 to render the entire 19-minute project with no GPU support
I also ran these same tests with my 2017 Dell XPS 15 (i7-7700HQ, 32GB RAM, GTX 1050) to provide an example of what an entry-level GPU can accomplish regarding 4k timeline playback and rendering.
With the CUDA cores turned on the 4k playback in the timeline is smooth, but as the CPU begins to heat up it begins to throttle. This causes the timeline to become jumpy and inconsistent. The throttle of the CPU decreases the CPUs ability to quickly send the GPU the information it needs to accomplish its task. As you see the GPU becomes underutilized as the CPU starts to throttle. When we first started the playback, before the CPU heated up, the GPU was bouncing around 65% - 80% utilization. But now that the CPU is throttle it sits much lower.
Ok, now let’s take a look at how this system handles rendering,
5:53 to render out 7240 frames with GPU support (CUDA cores) turned on
1 Hour + 15 minutes to render the entire 19-minute project with no GPU support
As you can see it is important to consider both components when purchasing a laptop or desktop computer for video editing. It is not enough to get either a powerful GPU or a powerful CPU. Having an equally powerful CPU and GPU will yield optimal performance out of each component.
------
DISCLAIMER: ALL LINKS IN DESCRIPTION ARE AFFILIATE LINKS
Thanks for Supporting Our Channel!
Комментарии