Rendering 3D Graphics on a 32×32 RGB LED Matrix Display with a Raspberry Pi and GeeXLab
0 Members and 1 Guest are viewing this topic.
When writing an article that evaluates multiple applications, it’s challenging to resist the temptation to compare each program to one another. The focus of this article was to evaluate how each application utilized CUDA to accelerate transcoding times and decrease or match CPU usage, so that is what I’ll attempt to focus my comments on.Each application we reviewed today had their own separate pros and cons, but most of them utilized CUDA quite efficiently and we noticed considerable drops in CPU usage during most of our benchmarks. The only program we didn’t receive large decreases in CPU usage was CyberLink PowerDirector 7. The CPU was pegged at 100 percent during every benchmark we threw at PowerDirector 7. It's not a bad thing for the CPU to be utilized at 100 percent with the GPU still engaged, but it would have been nice to see some more CPU efficiency so we could do more multi-tasking during transcoding.All our benchmarks showed that enabling and disabling CUDA had a huge impact on video transcoding times. Some applications mainly focused on better transcoding times and not on offloading work from the CPU to the GPU, which is just fine if users need an application that takes advantage of every processing option available. Overall, all of these applications had something unique for consumers to enjoy.