Hi Abel,
I've been using a script in Linux for FFMpeg. Stumbled onto your program to use in Windows. Works great. Is there any way to use multiple graphics cards to work on the Q? Or a way to run multiple instances of the program? Thanks for any info. Great work
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Abel, Can't believe you updated that so fast. Thanks! I tried multiple instances with multi cards. Just tried specifying which GPU in the command line. Didn't work out. If I set to decode with the GPU and specify a certain GPU in the command line for encoding then it actually used one card for each. One weird thing is if I add -gpu 1 to command line it encodes on GPU 0. So backwards for some reason. I'm not sure what a good solution is. I will have a couple Quadro cards arriving soon and can test more. I haven't seen any solutions out there for using multiple cards yet.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Abel,
I've been using a script in Linux for FFMpeg. Stumbled onto your program to use in Windows. Works great. Is there any way to use multiple graphics cards to work on the Q? Or a way to run multiple instances of the program? Thanks for any info. Great work
Hi, I'look into it for next release, now only one instance is allowed.
Hi Abel, Can't believe you updated that so fast. Thanks! I tried multiple instances with multi cards. Just tried specifying which GPU in the command line. Didn't work out. If I set to decode with the GPU and specify a certain GPU in the command line for encoding then it actually used one card for each. One weird thing is if I add -gpu 1 to command line it encodes on GPU 0. So backwards for some reason. I'm not sure what a good solution is. I will have a couple Quadro cards arriving soon and can test more. I haven't seen any solutions out there for using multiple cards yet.
I found this:
"When transcoding, no matter which gpu I select with -gpu option, ffmpeg always use gpu #0.
I managed to solve it by use -hwaccel_device # instead of -gpu
For example:
ffmpeg -hwaccel_device 0 -hwaccel cuvid -c:v h264_cuvid -i <input> -b:v 2048k -vf scale_npp=1280:-1 -c:v h264_nvenc -y <output>