Stats

  • Locked Locked
  • Replies 1
  • Subscribers 61
  • Views 741
  • Members are here 0
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

poor performance of LEC compile -parallel

archive

LEC spend many time to compare these "difficult" compare points of my case. And, I always get
some abort points.   I try to use "compare -parallel machine.list"  command to speed up run time.
I add 3 machines in machine.list file.  I am surprise to find the run time of "compare -parallel"
 is 3x -5x slower than "compile" only.  I parpared a small case to local AE. The AE saw  similar
 result but LEC RD said that is case-by-case.  I don't like the excuse. No explain, just case-by-case.  Anyone can tell why ?

Tom Tseng


Originally posted in cdnusers.org by tom.tseng

  • Hi tt,
    Parallel processing is a subject of much discussion in software engineering. Most parallel processing algorithms work through a system called MPI or Message Passing Interface. What it means is that when slave processors take on a computation there is message passing happening between the master and the slave or between multiple slaves. Sometimes the message passing could become the bottleneck and the parallel task could be slower than the standalone processing. Therefore when LEC R&D said that this was case-by-case, it was most likely correct. This is especially so when you have a smaller testcase because the message passing will take more time than the actual processing time. Also there are other factors such as speed of the communications backbone between processors etc. to worry about.

    Nonetheless, are you still facing aborts in the standalone run? If so, file a support request and get their help to solve that abort.
    -- mm


    Originally posted in cdnusers.org by makkarm

Community Guidelines

The Cadence Design Communities support Cadence users and technologists interacting to exchange ideas, news, technical information, and best practices to solve problems and get the most from Cadence technology. The community is open to everyone, and to provide the most value, we require participants to follow our Community Guidelines that facilitate a quality exchange of ideas and information. By accessing, contributing, using or downloading any materials from the site, you agree to be bound by the full Community Guidelines.