Some time ago we tested the UHD 770 and UHD 730 core display carried by the 12th generation Core processor, at that time we mainly tested the game performance of this core display, video encoding / transcoding is just a simple test, but I see that someone below replied to want to see intel and AMD nuclear display transcoding efficiency comparison, this can indeed have, but since you want to do it to do something complete, the NVIDIA and AMD unique display also added, see how much difference is between the unique display and the nuclear display.
NVIDIA's latest 30 Series graphics cards include the fifth generation NVDEC and the seventh generation NVENC, which can support video encoding in MPGE-1, MPGE-2, VC-1, VP8, VP9, H.264, H.265, AV1 formats, as well as video encoding in H.264 and H.265 formats.
Intel's 11th and 12th generation Core processors are integrated with Xe architecture core display, both support the eighth generation Intel Quick Sync Video, can support MPGE-2, VC-1, JPEG, VP9, H.264, H.265, AV1 format video encoding, support MPGE-2, JPEG, H.264, H.265 format video decoding, including Tiger Lake and Alder Lake's kernel display has two multi-format encoder engines, while Rocket Lake has only one.
AMD's current desktop Ryzen™ 5000G series processors use Vega architecture core display, integrating the VCN 2.2 engine, while the latest RX 6000 series graphics card and Ryzen™ 6000 series mobile processors use TheAvi 2 architecture GPU and use the VCN 3.0 engine, both of which support MPGE-2, VC-1, VP9, JPEG, H.264, H.265 decoding, as well as H.264 and H.265 format video encoding, the difference between the two is that VCN 3.0 supports AV1 format video decoding and VCN 2.2 does not support, it should be noted that the RX 6500 XT using The Navi 24 GPU only has a video decoder without an encoder, and does not support AV1 format, other RX 6000 series graphics cards use a complete VCN 3.0 engine.
Test platform with description
In terms of processor, Intel uses Core i5-12600K, and its integrated UHD 770 is currently the highest nuclear display in Intel, and Core i9/i7 also uses this, that is, the frequency is a little higher, there is no essential difference. The AMD side was supposed to use Ryzen 7 5700G, but at present there is no one on hand, so I have to use Ryzen 5 5600G, the video engine of the two is the same, but the lack of two CPU cores will indeed have an impact. For graphics, NVIDIA uses the RTX 3050 and AMD uses the RX 6600, both of which were tested on the Core i5-12600K platform.
We tested three of the more commonly used software, including video transcoding software HandBrake and two video editing software Premiere Pro and DaVinci Resolve.
HandBrake video transcoding
HandBrake We are to convert a 10 minutes and 02 seconds video into a 4K H.264 and H.265 video, will use different brands of corresponding video encoders, encoder presets are the fastest, output video quality is fixed is 30, the following table records the transcoding frame rate.
The CORE i5-12600K's integrated UHD 770 core display has two multi-format encoder engines, which are quite fast when converting H.265 format video, nearly double the speed of the other three, but the utilization rate cannot be improved when converting H.264 video, the transcoding speed is a little slower than the Ryzen 5600G, and the two unique displays are better than the Ryzen 5600G, and the RX The 6600 is more efficient at transcoding than the RTX 3050 in both H.264 and H.265.
Premiere Pro video export
We used Premiere Pro to export a 10-minute video project, the original material is 4K resolution, exported to 4K H.264 and H.265 format video, recording the time taken to export. Premiere Pro supports Intel's core display with other unique hardware acceleration, we will also test this, in this case UHD 770 core display will load video decoding, and the unique display will load video encoding, AMD's core display can not achieve this function.
In the Pr Intel UHD 770 performance is actually quite bad, we have used the latest Premiere Pro 2022 22.2 version, but the video export when the UHD 770 occupancy rate is still not mentioned, Core i5-12600K than Ryzen 5 5600 export much slower, the two single use alone is the export efficiency are similar, and the export speed with uhd 770 nuclear display has been significantly improved, And the CPU usage rate is the lowest when used in this way, so it is recommended to use the Intel processor with a core display with a dedicated display with a Premium Pro-based one.
DaVinci Resolve video export
DaVinci We are also a 3-minute project to export to 4K H.264 and H.265 format video, recording the time taken to export, but I have to say that the compatibility of this software does have some problems, Intel nuclear display export H.265 format video is wrong, and AMD whether it is a solo display or a nuclear display to export H.264 format video is the CPU at work, the GPU is not loaded at all.
In fact, only the RTX 3050 can run the two export tests normally, but it can be seen that the RX 6600 is faster than the RTX 3050 when exporting H.265 video, the core display of Ryzen 5 5600G is much slower than the two single displays, and the fastest way to export H.264 video is actually processed with the Core i5-12600K CPU, which can be exported in 63 seconds, faster than the RTX 3050 and its own nuclear display. It's just that the DRY CPU usage rate is definitely full, and when accelerated with GPU hardware, it is only about 50% to 60%.
summary
How to say this test, in fact, just look at the transcoding ability, in fact, the two nuclear displays and two unique displays of this test are quite close in most cases, that is, Intel's Xe nuclear display is a little incredible fast on HandBrake's H.265 transcoding.
As for the video editing software, the compatibility of the software may be more important than the performance of the hardware itself, only the Core i5-12600K UHD 770 core display in the Pr export time is even slower than the Ryzen 5 5600G Vega core display, but if used with NVIDIA or AMD alone, there is a significant acceleration effect, and amD's nuclear display does not have this function. DaVinci's words do not have the pr kind of core display and unique display together to accelerate the function, and AMD's VCN engine in the H.264 output completely did not play a role, do not know what the problem is, Intel's Quick Sync Video in the output H.265 also directly reported the error.
Unique display, in fact, NVIDIA and AMD graphics cards in this area of acceleration effect are very good, we tested the RX 6600 slightly faster than the RTX 3050, but the original performance of this card is stronger than the RTX 3050 so this result is normal, if you use Pr, no matter which graphics card does not matter, but it is best to use it with Intel's nuclear display, and with DaVinci as the main word, it is recommended to choose N card handle.