Quaro K2200 H.264 Encoding acceleration: is it any good??

[er_ouz]

Hi!
So I've been looking into buying a Quadro K2200 card, as I'm a video editor (I use Avid Media Composer), and one of my needs is fast H.264 encoding.

Currently encoding a single hour long video can take 4-6 hours. I know of at least one product (the Matrox MXO 2 Mini MAX) that has a dedicated H.264 encoder that allows h.264 encoding to be "faster then real time"- meaning it reduces the 4-6 hours needed to encode an hour long video into about 50 minutes!

I noticed that the new Nvidia Quadro K2200 video card also has a dedicated h.264 encoder- but I couldn't find any information as to how efficient it is anywhere!!! I found a whole bunch of benchmarks, but none pertaining to the h.264 encoding.

So- my question is: how efficent is the Quadro k2200 at encoding h.264 vidoes? Is it a suitable alternative to the Matrox MXO 2 mini MAX in that field? Will it work with Avid Media Composer? Or a different encoding utility?

Please help me out!
Thanks

 

[i7Baby]

Most of the encoding is done by your CPU. What you got?

It also helps having lots of fast ram. What you got?

The help the GPU provides is a function of cuda cores (nVidia) or Streaming Processors (Radeon). Gaming cards are cheaper and have a lot of cores/processors. My editor uses Hardware Acceleration with SLI. So I've got 2 x gtx670s.

My GPUs help with some encoding but not with all codecs. My i7 3930k does most of the work though.

 

[er_ouz]

I have a 16 Gb i7 4790K OC to 4.5 Ghz
The entire point of the dedicated encoder chip is to take the load off the CPU though, and do the encoding on the dedicated chip which is also much faster because it was designed for that task only. Like I said- the On the MXO2 mini MAX this means faster then real time h.264 encoding- regardless of the CPU.

 

[i7Baby]

"and do the encoding on the dedicated chip which is also much faster because it was designed for that task only." what dedicated chip?

 

[er_ouz]

"and do the encoding on the dedicated chip which is also much faster because it was designed for that task only." what dedicated chip?

Dedicated H.264 encoder (requires application support)
( https://www.pny.com/nvidia_quadro_k2200 )

 

[i7Baby]

There's nothing about that in that link.

The cuda cores assist the CPU in rendering some codecs. Cuda cores are not dedicated to just doing this.

The K2200 has 640 cuda cores. A GTX670 has 1344. That's why I have 2 x gtx670s.

 

[er_ouz]

There's nothing about that in that link.

The cuda cores assist the CPU in rendering some codecs. Cuda cores are not dedicated to just doing this.

The K2200 has 640 cuda cores. A GTX670 has 1344. That's why I have 2 x gtx670s.

Under Specifications:
Features & Benefits
3D Graphics Architecture

Scalable geometry architecture
Hardware tessellation engine
FXAA/TSAA dedicated antialiasing engine
Bindless Textures
Shader Model 5.0 (OpenGL 4.4 and DirectX 11)
Up to 16K x 16K texture and render processing
Transparent multisampling and super sampling
16x angle independent anisotropic filtering
32-bit per component floating point texture filtering and blending
Up to 64X full scene antialiasing (FSAA)
Decode acceleration for MPEG-2, MPEG-4 Part 2 Advanced Simple Profile, H.264, MVC, VC1, DivX (version 3.11 and later), and Flash (10.1 and later)
Dedicated H.264 encoder (requires application support)
Blu-ray dual-stream hardware accelerating (supporting HD picture-in-picture playback)
Quadro Boost (Automatically increases GPU engine throughput to maximize application performance).

Look, I appreciate you trying to help, but I know what I'm asking. Dedicated h.264 encoders exist in many forms, and the question I'm asking is how good the Quadro K2200 is, and will buying it be a good alternative to the Matrox MXO2 Mini MAX, which I know offers better then real time h.264 encoding.

 

[i7Baby]

For a start, the K2200 is not a dedicated h.264 encoder. Its a workstation graphics card.

Secondly, its the CPU that does the bulk of the encoding work.

Make sure you have editing software that uses hardware acceleration so that the GPU can be used to help the CPU do the rendering.

The more cuda cores, the better.