[ https://issues.apache.org/jira/browse/SYSTEMML-1881?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Matthias Boehm resolved SYSTEMML-1881. -------------------------------------- Resolution: Done Assignee: Matthias Boehm Fix Version/s: SystemML 1.0 > Tuning parfor degree of parallelism for operations > -------------------------------------------------- > > Key: SYSTEMML-1881 > URL: https://issues.apache.org/jira/browse/SYSTEMML-1881 > Project: SystemML > Issue Type: Task > Reporter: Matthias Boehm > Assignee: Matthias Boehm > Fix For: SystemML 1.0 > > > Currently, we assign remaining parfor parallelism conservatively to > operations of the parfor body. Consider, for example, a Kmeans or MSVM > scenario with 10 runs or 10 classes respectively. On a box with 16 HW > threads, we assign k=10 to the parfor and {{floor(16/10)}} to remaining > operations. Since it is usually a good idea to slightly over-provision CPU in > order to get full utilization (due to barriers at the end of each operation), > we should tune this to {{round(16/10)}} which provides performance > improvements of about 15% in above examples. -- This message was sent by Atlassian JIRA (v6.4.14#64029)