As a TMPGEnc 3.0 user I was recently looking into the latest and greatest that TMPGEnc 4.0 has to offer and I saw on their "new features" page that they have implemented CUDA in some of their filters and show some benchmarks/speedups with those filters.
I haven't used it yet, so I can't comment on the value this brings or whether the IQ of the resultant CUDA encoded images are as good as CPU based encoding. But I thought some of you might be intrigued and maybe didn't know about this.
Of course since these benches are provided by the same people trying to sell you the product they have to be taken as absolute best-case examples of potential performance speedup as well. Would be great if Anandtech checked them out in their next CUDA badaboom review.
I haven't used it yet, so I can't comment on the value this brings or whether the IQ of the resultant CUDA encoded images are as good as CPU based encoding. But I thought some of you might be intrigued and maybe didn't know about this.
New Features in Ver. 4.6.2.266
NVIDIA CUDA technology is now supported for processing the video filters and decoding. The multiple cores of the GPU can divide the workload and run the processes in parallel for a huge boost in processing speed over your computer's CPU*. This lets you apply multiple filters such as video noise removal (time), smart sharpen, color correction, and more without having to add hours to your output time!
TMPGEnc 4.0 XPress
Of course since these benches are provided by the same people trying to sell you the product they have to be taken as absolute best-case examples of potential performance speedup as well. Would be great if Anandtech checked them out in their next CUDA badaboom review.