In a nutshell
- After years of Intel’s dominance, the resurgence of AMD has revitalized competition in the video editing CPU market, narrowing the performance difference and driving innovation.
- Both Intel and AMD’s latest processors have faced challenges with thermal management and voltage, requiring innovative cooling solutions and voltage regulations to maintain performance and safety.
- Factors such as core counts, processing speed, editing software and peripherals play key roles in video editing performance.
After much hype and anticipation, the third quarter of 2022 finally saw the release of AMD’s new Ryzen 7000 and Intel’s 13th-generation Raptor Lake CPUs. In addition to higher clock speeds and increased cache sizes, AMD’s AM5 platform ushers support for DDR5 memory and PCI Express 5. Now that were entering the later half of 2023, are there new CPU options worth considering? Is there a new best CPU for video editing? Let’s have a look. But first, let’s discuss the current state of the CPU market to help put all this into context.
It’s nice to have competition again in the video editing CPU market
A decade of a weakened AMD coupled with Intel’s dominance led to year after year of Intel iterating on its quad-core designs. Intel’s high-end desktop (HEDT) offerings were the only non-Xeon processors pushing the envelope regarding core counts. If you wanted performance, it was not going to be cheap.
Skipping ahead to 2017, we finally saw a change. AMD released the first generation of Zen processors. For the first time in a decade, there was finally a performant alternative to Intel. AMD released Threadripper shortly after. AMD would go on to release Zen 2 and 3 over the next three years, effectively closing the gap—even surpassing Intel in some product categories. The Ryzen processors were no longer just the cheaper alternative.
While not directly aimed at consumers, the Zen architecture shares a lot with AMD’s EPYC server processors, an area that AMD has been gaining significant market share from Intel.
CPU State of the Union 2023
This brings us to 2023. This year, we’ve seen Intel release the Raptor Lake and AMD roll out the Ryzen 7000. At this point, the performance difference between Intel and AMD products is within a hair’s width. Competition has spurred incredible innovation. The main differentiator between the two mainly comes down to price—and processor design to some extent.
Intel’s 13th Gen Raptor Lake processors are an evolution of the technologies introduced in Alder Lake, including Thread Director, which directs data to either the performance core (P-core) for maximizing single-thread performance or the efficient core (E-core) for multithreading.
Their respective flagships are the AMD Ryzen 9 7950X and Intel Core i9-13900K. The 7950X sports 16 cores, 32 threads with a 4.5 GHz base clock, and the 13900K featuring 8 P-cores and 16 E-cores with 3.0 GHz and 2.2 GHz base clocks.
Keeping a cool head
Thermal throttling is when a CPU reduces its clock frequency and power draw because it has exceeded a temperature threshold and is a means of protecting itself from overheating. A video render is basically a timed CPU torture test. There is little sense in leaving performance on the table because of poor cooling.
The Intel i9-12900K released with a whopping maximum of 240 Watt power draw, and that’s before overclocking. Unsurprisingly, many CPU coolers were brought to their knees, as the processor would routinely reach above and up to 100 C under load. A popular mod in the PC building community uses a CPU contact frame to mount the cooler closer to the CPU to increase and improve contact pressure.
AMD has followed a similar trajectory with the Zen 4 processors with increased power draw and higher operating temperatures, albeit with the increased thermal overhead being for different reasons. To maintain CPU cooler compatibility with the previous AM4 socket, AMD increased the height of the integrated heat spreader to compensate for the new LGA (pinless) CPU design. The result was a thicker heat spreader with increased thermal resistance.
Another key difference with Zen 4 is the location of the chiplets no longer being dead center. This results in an uneven heat load on the CPU cooler, with one side having to dissipate more. Fortunately, several manufacturers have started offering offset mounting kits to better align the cooler with the lower-down hotspots.
Death by overvoltage
Recently, there has been some drama and controversy around some third-party AMD motherboards pushing higher than recommended voltages on the latest Ryzen 7000 X3D CPU models, resulting in burned-out processors. The bulk of the problems centered around Asus motherboards, Gamers Nexus produced a great video diving deeper into the issue.
However, this issue should be fully addressed by now, with BIOS patches being correctly issued to supply lower CPU voltage to X3D models.
Core counts
With more cores comes less speed. Higher core counts are great for enabling more simultaneous workloads but at the expense of outright grunt per workload. This is where the 7950X comes into its own with a new base clock of 4.5 GHz—versus the 3.4 GHz of the 5950X. Not only are clock speeds faster, but so is the overall efficiency of the processor. It can do more with less compared to the previous generation. Put all this together, and you have 16 physical cores and 32 threads capable of running flat out at 4.5 GHz. If that isn’t prime video editing and rendering performance, we I don’t know what is.
Price-wise, the AMD 7950X and the Intel i9-13900K both are priced in the $570 range.
There is also a sadder side to this; the world of high-end desktops (HEDT) is no more. As of July 2023, Intel has officially discontinued the X299 platform. Thought, we all knew it was coming. AMD’s Threadripper suffered a similar fate in 2022, with the discontinuation of enthusiast models and new models only available through OEM workstations.
It’s not all bad news, though, with the 7950X beating out the older Threadripper 5955WX by over 20 percent. The biggest restriction is felt around PCIe lanes, which the inclusion of PCIe 5 does help get around.
Factors in processing speed
Editing video taxes computer technologies to their fullest potential. Without the right components, video editing can cost you significant losses in time and effort. Consider rendering and exporting a video that utilizes almost all aspects of a computer. Processing video also depends on the effects, resolutions and codecs. Processing a 4K video vs. a 1080p video will have significant performance differences. A 4k video has four times the pixels as a 1080p video, so this could potentially slow the processing of the video. If you have the right number of cores, motherboard, RAM and a fast SSD, you shouldn’t expect a huge difference in processing times. Alternatively, processing an 8K video compared to a 1080p will significantly increase processing times no matter your setup—as 8k has four times as much information as 4k.
Type of editing software
A lot depends on the NLE software (non-linear editing software) editors use. The peripherals the computer comes built with are also a factor. For instance, Adobe Premier Pro can efficiently utilize up to 24 cores—as seen in Puget Systems’ performance review. However, testing shows diminishing performance among 32 and 64-core CPUs. This is most likely due to software limitations and cannot be changed by the end user. Adobe recently updated its Premier Pro software to use more CPU cores effectively. The update also uses the GPU far more than previous releases to encode and decode H.264/HEVC media. Earlier versions of Adobe Premiere Pro used Intel’s Quick Sync as the only option to achieve hardware-accelerated encoding and decoding. Now editors don’t need to spend $4,000 on a huge multicore CPU to reap the benefits of one.
Peripherals
AMD and Intel are battling for your gaming and video editing dollars. Therefore, end-users can take advantage of enhancements, new technologies and higher-performance peripherals in their day-to-day computer builds. A strong motherboard that matches the front-side bus (FSB) CPU speed is a mandatory purchase to take full advantage of a new CPU. Think of the CPU as a water tank full of data. You have either a fire hose or a garden hose to extract that data. The FSB is that hose. Video editors will want to match CPU, motherboard, RAM and GPU to make the most effective video editing computer system.
That’s all great, but what’s the best CPU for video editing right now?
With the Threadripper series resigned to workstations, consumers are left with only two top-tier CPU choices for video editing. The Intel i9-13900K has a slight edge over AMD 7950X in single-core performance and near enough tied-in multicore performance.
With so many aspects being equal, we need to focus on a few key differentiators: power draw and platform life.
This is where the 7950X emerges as the victor. It has lower peak and average power consumption and greater overall efficiency for the same performance. Then there is the AM5 platform that was released alongside the new processors. If the six-year lifespan of AM4 is any indicator, users can hope to see at least a few years of support. In contrast to Intel’s two-year platform cycle, the current LGA 1700 is now in its last year and leaves Alder and Raptor Lake users with no upgrade path.
Contributing editors to this article include: Blag Ivanov and Brian Teal.
