No I don't think that is the case. The throughput numbers for this
benchmark at one specific data point on an 8-way look like.
2.4.14 Default Scheduler 39463
2.4.14 MQ Scheduler 385687
2.4.16 Default Scheduler 51364
2.4.16 MQ Scheduler 240667
The MQ numbers are still quite a bit better even on 2.4.16. I can
easily get MQ 2.4.16 back up to MQ 2.4.14 levels by reintroducing the
code to give a slight preference to the currently running task.
However, our MQ scheduler is trying to closely match the behavior of
the default scheudler wherever possible (for better or worse). It
should also be noted that performance of the default scheduler
increased as a result of Ingo's changes.
It is entirely possible that there is some other bug/feature in our
MQ code that is causing this situation. As mentioned by others, the
scheduler code that was removed from 2.4.15 should have little if
any impact on performance. I need to do some more analysis here.
-- Mike - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/