{var%20f='http://v.t.sina.com.cn/share/share.php?appkey=1515056452',u=z||d.location,p=['&url=',e(u),'&title=',e(t||d.title),'&source=',e(r),'&sourceUrl=',e(l),'&content=',c||'gb2312','&pic=',e(p||'')].join('');function%20a(){if(!window.open([f,p].join(''),'mb',['toolbar=0,status=0,resizable=1,width=440,height=430,left=',(s.width-440)/2,',top=',(s.height-430)/2].join('')))u.href=[f,p].join('');};if(/Firefox/.test(navigator.userAgent))setTimeout(a,0);else%20a();})(screen,document,encodeURIComponent,'','','https://www.manongdao.com/data/attach/logo/logo.png', '推荐 再贱就再见 的问题《What is Segmented Ring Allreduce in OpenMPI (V2.0.》','https://www.manongdao.com/q-754683.html','页面编码gb2312|utf-8默认gb2312'));){kind=link}
I have been trying to understand the Segemented Ring Allreduce in OpenMPI (V2.0.2). But I failed to figure out this pipelined ring allreduce, especially how the phases are pipelined. (i.e. COMPUTATION PHASE 1 (b) seems to perform the two phases concurrently instead of "pipelinely".) Could MPI experts provide the motivation behind this Segmented Ring Allreduce and details about the pipeline?
Really appreciated, Leo
i think this has been asked and answered at https://github.com/open-mpi/ompi/issues/4067
very specific questions regarding the internals of Open MPI, are better asked directly to the Open MPI mailing list(s) or github repository