適応型投機的並列メッシュ生成のためのタスク処理フレームワーク
メッシュ操作の機能と性能の分離を可能にするタスク処理フレームワークを提案する。このフレームワークにより、スレッド管理とスケジューリングの決定を一般的で再利用可能なタスク処理フレームワークに引き上げることができる。
Efficient Parallel Mesh Generation with Adaptive Speculative Tasking Framework
The presented tasking framework enables separation of concerns between functionality and performance aspects of parallel mesh generation, leading to improved performance and portability.
Accelerating Photoacoustic Image Reconstruction through GPU-Based Parallel Computing
GPU parallel computing can significantly accelerate the computationally intensive iterative reconstruction process in photoacoustic imaging, enabling faster image processing and broader adoption of this promising medical imaging technology.
Optimizing Heterogeneous Task Scheduling through Adaptive Priority in Task-based Runtime Systems
INSPIRIT, an efficient and lightweight scheduling framework with adaptive priority, is proposed to optimize task scheduling in task-based runtime systems on heterogeneous hardware. INSPIRIT introduces two novel task attributes - inspiring ability and inspiring efficiency - to determine task priorities, eliminating the need for application domain knowledge. INSPIRIT also jointly considers runtime information such as the number of ready tasks in worker queues to guide task scheduling, exposing more performance opportunities in heterogeneous hardware while reducing overhead.
Efficient Algorithms for Computing Round-Optimal Broadcast Schedules in Logarithmic Time
The article presents optimally fast O(log p) time algorithms for computing round-optimal broadcast schedules for message-passing parallel computing systems, where n indivisible blocks of data need to be broadcast from a root processor to all other processors in a fully connected network of p processors.
Libfork: Portable Continuation-Stealing with Stackless Coroutines
Stackless coroutines in C++20 enable fully-portable continuation stealing, leading to optimal time/memory scaling in parallel computing.
Libfork: Portable Continuation-Stealing with Stackless Coroutines
Libfork enables fully-portable continuation stealing using stackless coroutines, achieving optimal time/memory scaling in parallel computing.
