facebook rss twitter

Review: GPU-accelerated video-encoding. Who is better, AMD or NVIDIA?

by James Smith on 28 October 2009, 12:09

Tags: AMD (NYSE:AMD), NVIDIA (NASDAQ:NVDA), HiS Graphics, BFG Technologies

Quick Link: HEXUS.net/qaukz

Add to My Vault: x

GPU encoding comparisons, plus conclusion

Here are the same 720p YouTube outputs, but now encoded using the GPU.

GPU encoding on HIS Radeon HD 5870 (click to enlarge)


GPU encoding on BFG GeForce GTX 295 (click to enlarge)

The difference in image quality between the output from the CUDA-accelerated encode and the ATI Stream-accelerated encode is substantial. Not only does the CUDA-accelerated output maintain greater fine-detail quality in the video but it also displays much better contrast. Indeed, the NVIDIA output appears to maintain so much more detail than the CPU-encoded output that it leads us to believe the CUDA codepath is not using the same coding parameters - and may even be using a more compute-intensive encoding engine.

It should be noted, however, that the final image quality of the GPU-accelerated output file varies hugely, dependent on the source footage fed into MediaShow Espresso. This was made blatantly obvious when we analysed the image quality of the output files created from a different HD feed. AMD came up trumps by offering a much cleaner (albeit slightly softer) output than the NVIDIA's.

Conclusion

Although this is just one example of a GPGPU application from the increasing number that are now available, it’s clear to see from the results that not only is there a marked performance difference, but also a marked difference in image quality: and both are dependent upon the actual HD footage used.

Sometimes AMD-produced video is better, other times, as per the screenshots, NVIDIA's is better. Additionally, CyberLink’s MediaShow Espresso doesn’t allow enough flexibility to achieve the best compromise between file-size and image quality. To make matters worse, at least with both of the MPEG-2 1080p video clips we fed it, it would only output with an incorrect 29fps frame-rate, instead of the input files 25 fps frame-rate.

GPGPU encoding is a nascent but burgeoning sector that will develop over time. The intrinsic benefits of massively parallel processing are obvious, and the results show that high-end cards can beat out any consumer CPU with consummate ease. Taken with CyberLink's MediaShow Espresso software as a base, we'd recommend that users opt for a program with significantly better encoding quality and speed. 



HEXUS Forums :: 13 Comments

Login with Forum Account

Don't have an account? Register today!
I thought GPGPU acceleration was supposed to be 10-50x faster than a CPU?
I noticed that the nVidia card was only about 25% faster and the ATi, although being a lot faster than nVidia, was also only about 40% faster than the CPU at the same task.
What's the deal with that?

I understand they were using one of the fastest CPUs available right now, but they're also using 2 very very fast GPUs.
matty-hodgson
I thought GPGPU acceleration was supposed to be 10-50x faster than a CPU?
I noticed that the nVidia card was only about 25% faster and the ATi, although being a lot faster than nVidia, was also only about 40% faster than the CPU at the same task.
What's the deal with that?

I understand they were using one of the fastest CPUs available right now, but they're also using 2 very very fast GPUs.

I've done some of my own testing with Badaboom and a 9600GT, the speed wasn't great - no faster than a my Q6600, sometimes slower and the quality appeared worse than using Handbrake on the CPU.

EDIT: No doubt a 9800 or GTX would be faster than my Q6600, but the quality and tweaking in Badaboom is way below Handbrake.

Tried running it for curiosity on my 8200 IGP too, it was an utter joke - Athlon II 240 much faster (5-10x) - so to tout CUDA support on the IGP is a bit of a stretch, as doing the work on your CPU would probably be faster anyway.
Crap code. A common problem with cyberfail. Overall it's going to take some time for GPGPU code quality to catch up.
I had the impression Nvidia's CUDA was better, but that appears not to be the case (at least speed wise).

It's a pity they're using video encoding as the sole comparison. OK, so it's one of the key areas you might consider, but I'd also prefer a direct comparison that can be objectively measured. Video quality by its nature is subjective, but the output of e.g. fluid dynamics/something else compute bound with a definite correct solution would be nice to see.

PK
Syllopsium
I had the impression Nvidia's CUDA was better, but that appears not to be the case (at least speed wise).

It's a pity they're using video encoding as the sole comparison. OK, so it's one of the key areas you might consider, but I'd also prefer a direct comparison that can be objectively measured. Video quality by its nature is subjective, but the output of e.g. fluid dynamics/something else compute bound with a definite correct solution would be nice to see.

PK

This is the first part of the GPGPU look. We'll be investigating other areas in due course.