Troubleshooting
Problem
When using multiple TI processes in parallel (TM1RunTI / RunProcess), there are a number of factors that could hinder or improve performance of your overall runtime.
Diagnosing The Problem
When attempting to achieve maximum performance for parallel TI processes, you must also consider the other work the TM1 Server is doing. Changes to improve parallel TI performance can affect performance of other activities if not planned or tested thoroughly.
Resolving The Problem
When building a job or chore using multiple TI processes in parallel it is recommended to:
-Avoid hyperthreading
-Achieve an even workload distribution across all TI processes
-Avoid running a number of processes in parallel that exceeds the number of cores on a socket
*For example, a 4 socket x 8 core per socket configuration should not exceed 8 parallel threads
-Achieve an even workload distribution across all TI processes
-Avoid running a number of processes in parallel that exceeds the number of cores on a socket
*For example, a 4 socket x 8 core per socket configuration should not exceed 8 parallel threads
**A 2 socket * 16 core per socket configuration would allow for 16 parallel threads
Document Location
Worldwide
[{"Type":"none","Line of Business":{"code":"LOB10","label":"Data and AI"},"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSCTEW","label":"IBM Planning Analytics Local"},"ARM Category":[{"code":"a8m50000000KzIpAAK","label":"Performance"}],"ARM Case Number":"TS004950957","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"All Version(s)"}]
Was this topic helpful?
Document Information
Modified date:
10 May 2021
UID
ibm16451169