trixi-framework/Trixi.jl

Benchmark pre-computing `weights2d`, `weights3d`

Open

#279 建立於 2020年11月2日

在 GitHub 查看
 (0 留言) (0 反應) (0 負責人)Julia (686 star) (153 fork)batch import
good first issueperformance

描述

As suggested by @gregorgassner in https://github.com/trixi-framework/Trixi.jl/pull/274#discussion_r515895414.

Something like weights2d, weights3d could be added to the cache of the semidiscretization.

貢獻者指南

技術棧
無
領域
performancebackend
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-2 days
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
mostly clear
前置要求
Julia programmingTrixi.jl cache internals
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
20
研究方向
Start by reviewing the linked PR #274 discussion, especially the comment by @gregorgassner. Examine the current implementation of `weights2d` and `weights3d` in the source code to understand how they are computed. Then, explore the semidiscretization cache structure to identify where to add these precomputed weights. The goal is to benchmark the performance improvement of caching these weights versus computing them on the fly. Consider writing a small script to test both approaches.