| Index: > A B C D E F G H I J K L M N O P Q R S T U V W X Y Z |
|
|||||
Each thread in the queue has a tag indicating its preferred / kin processor. At allocation time, each thread is allocated to its kin processor. This can effectively reduce cache problems.
However, it does not curb the persistent load-balancing problem.