Concerning the used Quark Scheduler i have two questions.
- Is the schedular mainly lock-free, such as intels TBB, to allow
'real' fine grain parallelism? Acctually i roughly debugged the code, and
saw alot of mutexes, so I was just wondering....
- What was the objective, not to use other available solutions, such as
OpenMP or TBB, which relies on a work-stealing concept?
thanks for any help,