does anyone know to which degree video encoding speed
can be improved by involving graphics card api if I understood
correctly is possible in ffmpeg. As encoding
for example dvd resolution with ffmpeg at 10 mbit/s is hardly better than 1.0x speed with x264 codec using no special settings
on an old intel core i5. Is a some entry level amd card worth
trying to gain significant improvement. Depending on the level of ffmpeg integration for iOS an iPhone could potentially be faster for this use case.
I’m definitely not an expert, but I think it depends a lot on the use case, the hardware, the targeted codecs etc.
It’s very difficult to find universal benchmarks on this topic. However with an old CPU it’s very likely that GPU encoding will bring significant improvements in encoding time.
Ffmpeg can use APIs from both AMD and Nvidia cards so there are endless combinations.
@sonyveijo Well Leap 15.6 is on the decline… I’ve had good OTB experiences with Intel entry level GPU’s A310 and A380 on Leap 16.0 and Tumbleweed best experience is with Tumbleweed as the likes on Intel (esp Xe driver over i915) and Mesa packages are rolling along with improvements for the Xe driver.
I’m using the Xe driver at present on Tumbleweed driving three screens;