Hi all:
I have some different algs in dm8168 dsp side, i found that
in mcfw alg_link framework the video datas are passed through one prev link and
the algs run in serial order, so the time delay is total time of every alg costs,
it is seems the framework not so suitable for multi algs,why not these algs are done in parally order,
then time delay will decrease?
another question is that so is the internal mem space which alloc when resources init,
can be shared between algs anymore?
best regards
xavier