YOU SAID:

Multicore Programming Multi-core processors enable multiprocessing in a single physical package. [...] Possible gains are limited by the fraction of the software, which can be split into multiple cores at the same time. At best, so-called embarrassing parallel problems can produce acceleration factors close to the number of cores. The basic steps in designing parallel applications are: partitioning This phase is intended to speed up parallel execution. The focus is on the definition of a large number of small tasks to obtain a fine-grained decomposition of a problem. communication The tasks generated in the partitioning phase are to be executed simultaneously, but can not be executed independently. A task usually requires data that is associated with another task. Data must then be transferred between tasks. This information flow is indicated in the communication phase. agglomeration Parallelism is so fine-grained that the overhead of parallel scheduling or communication raises the useful work. In the third step, the developers assume whether it is useful to combine or agglomerate the tasks identified by the partitioning phase. mapping In the fourth and final phase of the design of parallel algorithms, the developers specify where each task should be executed.

INTO JAPANESE

マルチコアプログラミングマルチコアプロセッサは、単一の物理パッケージでマルチプロセッシングを可能にします。 [...]可能な利益はソフトウェアの一部によって制限され、同時に複数のコアに分割することができます。せいぜい、いわゆる厄介な並行問題は、コア数に近い加速要因を生むことがあります。基本的なステップi