admin管理员组

文章数量:1401849

In the last example of Mark Harris' webinar I don't understand the indexing before the parallel reduction part. In "Reduction #6" the gridSize/number of dispatches was ceil[N (the size of the data) divided by blockSize/WorkgroupSize and by 2 because we accessed two items at once].

Now "it's as many as necessary". But what is the gridSize? It can't be ceil[N/2/workgroupSize] like before because then in the while loop it would exit on first iteration. What's the number of dispatches when calling the kernel then?

本文标签： cudaNVIDIA webinar on parallel reduction gridsizeStack Overflow

版权声明：本文标题：cuda - NVIDIA webinar on parallel reduction gridsize - Stack Overflow 内容由网友自发贡献，该文观点仅代表作者本人， 转载请联系作者并注明出处：http://www.betaflare.com/web/1744311563a2600054.html， 本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如发现本站有涉嫌抄袭侵权/违法违规的内容，一经查实，本站将立刻删除。