FX 8350 shines with h264/x264, it beats i7 3770k in h264/x264...
Anything else to add besides this which is just going to start a flame war?
OpenCL is not AMD's. It's a standard by the Khronos Group, which handles OpenGL, who's members include AMD, nVidia, Intel, Apple, and ARM, among others.Neither of these utilizes GPU encoding. The nVidia version is called CUDA and the AMD version is called OpenCL.
Anything else to add besides this which is just going to start a flame war?
I'm pretty sure there are softwares out there that can use a GPU to accelerate rendering video to a file...Rendering a video 3D file uses the CPU.
Why would that start a flame war? Video coding is one reason to "need" a lot of CPU power, and one area where AMD is not behind. I'd definitely recommend AMD as a viable option to anyone looking to do a lot of video coding.
OpenCL is not AMD's. It's a standard by the Khronos Group, which handles OpenGL, who's members include AMD, nVidia, Intel, Apple, and ARM, among others.
nVidia supports OpenCL about as well as AMD does.
I'm pretty sure there are softwares out there that can use a GPU to accelerate rendering video to a file...
Ok I've got some early data.yes I am talking about a video.
12 minute HD video at the best settings. I don't know the best settings so post them once the video is done.
upload it to youtube, which will compress it a bit but then I can compare it to my HD video.
Then tell me how long it takes to render.
thanks
Ok I've got some early data.
I took a 12 minute clip of a high quality 1920 by 798 MKV video and converted it from color to black and white (just for a test). The output file was the same resolution. The original file size is 1.23GB. Audio is 6 channel DTS, it was set to just pass through (didn't mess with the sound).
Using Avidemux with Linux it took 8 minutes to transcode the 22 minute video at around 33fps using an i7-3770s. This was all done in RAM so IO was not a big issue. The OS loads to RAM and both the input and output files were also in RAM.
Using Avidemux with Windows 7 was slower, not sure why. I have an SSD so IO should not have been a big issue... Anyhow it took the Windows version of Avidumux 13 minutes at around 23fps to do the same job with all 8 virtual cores banging away on the i7-3770s.
Screen:
I'll run the same tests on the i5-2500k and the 1055t as soon as I get a chance.
even an 8 core 8350 can't touch a 3930k (6 cores, 12 threads) at the same speed, and yes, most 3930k's will do 4.2 at least with good cooling.
Were all 8 threads loaded? Just wondering how well threaded Avidumux is compared to something like Handbrake which does load all 8 threads when I downsample my raw M2TS files to M4P.
Shephard said:do you have any of your own footage in raw format that you can then render into whatever the best codec is, h264 or whatever it's called?
Using Avidemux with Windows 7 was slower, not sure why. I have an SSD so IO should not have been a big issue...
Your saying that the hard drive will make the render take longer compared to the cpu?
When I just cut the file to length (to 12 minutes) it took only seconds because there was no need for decoding and recoding. The copy was exactly the same as the original except for length.But you just changed color it was already an mkv you didn't render as anything else?
Not sure but the Windows version is definitely only 32bit. I did not compile either version from scratch.Was that Avidemux built entirely with 64bit in mind? x264 64bit is faster than 32bit. In case you didn't compile it yourself, certain optimizations might have also been not turned on for the Windows binaries.
what do you mean potential bottleneck? I just want to see a raw file rendered in Sony Vegas or whatever program he choses into the best HD quality uploaded to youtube.
Your saying that the hard drive will make the render take longer compared to the cpu?