Intel hardware Acceleration VS cuda outright video transcoding experience

Compare trial Platform Introduction

After experiencing the outstanding performance of the core I5 2300 in 3D games and multitasking, HD playback and more, today I used my free time to compare my old platform against the core I5 2300 in HD transcoding, and the results were shocking. Good gossip not much to say, first to give their own new and old PC to do an introduction:

Overview of the new platform

(The NVIDIA GeForce 9800 GT is intentionally configured in the legacy system to compensate for the shortcomings of the old platform CPU and GPU, as compared to the HD decoding aspect)

The above is a summary of two platforms, and then say that the software used mediacoder This free software everyone is not unfamiliar with it, the main reason for the multi-core CPU is optimized, and for NVIDIA graphics card has cuda acceleration and sandy Intel hardware acceleration in the bridge platform.

Accelerated mode of the new platform in MediaCoder selection accelerated mode of old platform in MediaCoder selection

performance efficiency under Intel hardware acceleration

The following are the specific steps:

Target: The 1G size, resolution 1080P, encoded as AVC HD video file transcoding for better clarity, small capacity, encoded as H. MP4 file. Compare the conversion time of the old and new platforms.

File information to be converted

Step One

First general set to output file resolution of 640*368 frame number is 12.500, the file encoding format is H. 2000Kbps, the average bitrate is as follows:

Step Two:

Transcoding on the new platform turns on Intel hardware acceleration, recording CPU usage and time spent in the conversion process.

First turn on core I5 2300 all cores

CPU usage is 47.6% because Intel hardware acceleration is turned on

Convert to a MP4 file encoded in H. 159, Spents. (Here is a simple bit rate; Bitstream x time = Total capacity here to note is the unit of the bitstream, 1 byte (b) = 8 bits (b), the capacity of the files on our computer k/m, all refers to b;1 kilobyte (k/kb) =2^10 bytes=1,024 bytes Kbytes; 1 megabyte (m/mb) =2^20 bytes=1,048,576 bytes megabytes; So if you use BITS/S's bitstream compute capacity, remember to remove 8. Rate relationship This file size and clarity, if the average bitrate is set to 1000Kbps then the text will be reduced by almost half of the conversion time has also decreased, but the clarity is greatly discounted. So in order to ensure the clarity of 2000KBP)

Cuda hardware accelerated execution efficiency

Step three: Do video transcoding on the old platform, enable Cuda acceleration for NVIDIA graphics cards, record CPU usage, and time spent in the conversion process. See figure below:

Open the old platform dual core

The Cuda acceleration of Nvidia graphics cards in the old platform is not performing satisfactorily.

CPU usage is slightly higher, and the Nvidia graphics card does play a big role in terms of CPU usage. (Because the time is limited, no power measurement of the n card, so the specific n card in the transcoding of the full operation remains to be verified.) )

The old platform has a transcoding time of 9 minutes and 19 seconds Intel accelerated VS Cuda accelerated outright

Comparison: the first into a larger or transcoding time, both from the above figure can be seen in the contrast set (except for hardware acceleration) are the same, but the speed varies greatly. While the old platform transcoding file occupies more than the new platform occupies 8MB, it can be said that the core I5 2300 with the independent transcoding engine unit greatly improve the efficiency, this can be seen in the figure of the CPU utilization.

The CPU usage from the lower right corner is 100% encoder occupies 66% (after opening Intel hardware Acceleration Encoder cpu% only 1.5%) visible independent transcoding engine units share most of the work of the encoder. With this, you can see that the core I5 2300 is the most powerful in transcoding.

Some friends may say that this video source file encoding format is not very good to support hardware acceleration, and in the setup of the problem, in fact, there are two reasons, one is because my purpose is to see the new platform and the old platform difference how much other less attention, and two is I also busy anxious contrast, MediaCoder This set of software professional is also relatively strong, so there is no time to look closely please forgive us. Finally, say a little more about the single display after the installation of the set is masked problem. It is heard that lucid Virtu software can solve this problem, so in the purchase of the core I5 2300 and have a separate video card friends can search as if there are 30 days free trial.