By running applications on the GPU rather than entirely on the CPU, it has been found to reduce the time required to perform computer intensive tasks. Read on to see if either GeForce 9800 GTX+ or a Radeon HD 4770 is quicker for video conversion on Cyberlink's MediaShow Espresso application.
CyberLink MediaShow Espresso software has been designed to work with both AMD’s Unified Video Decoder (UVD) through its ATI Stream technology and NVIDIA's general purpose parallel computing architecture known as CUDA. By using AMD Stream and NVIDIA CUDA technology consumers can quickly convert digital video files for use on portable devices like the Apple iPhone, Sony PS3, and YouTube. Since CyberLink MediaShow Espresso works on both NVIDIA and AMD graphics cards it gives us a rare chance to benchmark both technologies to see which is better.
After seeing this I decided to see how TMPGEnc was using CUDA and the CPU together. When you tell the program to use CUDA it states that if you have a fast processor that using CUDA for filtering may actually increase the encode time. I really didn't think much of this until you showed how the ATi card wasn't fully utilizing the CPU. I recorded Battle of the Bulge from AMC a few days ago so I decided to use this file to test with. Rather than use TMPGEnc to remove the commercials and recode the entire file I used MPEG-VCR to remove the commercials and TMPGEnc to deinterlace, resize and crop the original recording. Attached is a picture of the encoding, or rather recoding.
I have 4 History Channel recordings that I need to remove the commercials from and deinterlace. The problem there is I don't know exactly how to use TMPGEnc's batch tool.
Is the Espresso software pretty limited to what you can do with the file? It seems like Badaboom where you can only tell it the format and size you want the file. Is that correct? With TMPGEnc you can do all kinds of adjustments to the file and really fine-tune the result.
Well, I thought I would chime in. And please take what I say with a grain of salt as I know next to nothing about encoding. But I thought I would give it a try. In a nutshell all I did was take a 6GB uncompressed AVI video I took of NFS Pro Street using Xfire then converted it to multiple formats/resolutions in both 16:9 and 16:10 ratios.
Long story short using AVIVO and Espresso it took ~4 minuites and 44 seconds to encode. Oh, the input resolution was 1920x1200. I don't know if that is good, bad or otherwise. Personally I thought it was screaming fast but I noticed a max GPU usage was 18% and about 70% CPU. Wildly fluctuating and GPU usage was almost non existent according to CCC. That is, the GPU was idle most of the time.
-Rampage-Motherboard: Asus Rampage Formula CPU: Q6600 @ 3.2Ghz Cooling: OCZ Vendetta 2 Ram: 4GB OCZ Reaper HPC PC2-8500 5-5-5-15 @ 540Mhz Video Card: Sapphire Radeon HD 5870 Storage: Seagate Barracuda 7200.10 320GB/4x 500GB Barracuda 7200.11 Raid 0 Power Supply: PC P&C Silencer 750 Quad Case: Antec P180B Monitor: Syncmaster 245T OS: Vista 64bit
I downloaded the trial version of Espresso and fired up the 4 History Channel recordings. When I started the recoding the CUDA activity animation went away. I stopped the process and only started one and it was back.
