2009-02-01, 22:23
The problem only really comes out when you decode something which is eating up at least 100% of one core.
Could you try killasample.mkv ? and see the utilizations.
I have already tried Jaunty alpha 2 with latest kernel. It does NOT make a difference. Only the load sharing is better, but the sum of the two cores never goes up to 120-130% ! So the decode is still going sequential ! (in Hardy it reaches 190 or 195% when decoding killasample)
Could you try killasample.mkv ? and see the utilizations.
I have already tried Jaunty alpha 2 with latest kernel. It does NOT make a difference. Only the load sharing is better, but the sum of the two cores never goes up to 120-130% ! So the decode is still going sequential ! (in Hardy it reaches 190 or 195% when decoding killasample)