[Legup-bugs] [Bug 129] Loop pipelining does not work in conjunction with loop unroll flag

bugzilla-daemon at legup.org bugzilla-daemon at legup.org
Wed Mar 15 15:24:44 EDT 2017


http://legup.eecg.utoronto.ca/bugs/show_bug.cgi?id=129

--- Comment #19 from Andrew Canis <andrewcanis at gmail.com>  ---
I'm guessing you are correct, that the reason you aren't seeing any more
speedup is due to the number of memory ports. Right now, with local rams
enabled, each C array will be put in a separate dual port memory. So that makes
sense that an unroll factor of 2 sees a speedup (using both ports) but doesn't
improve beyond that.
You can also look in the pipelining.legup.rpt file for the lines near the top
saying "Calculating resource MII using IMS technique". This will give the
resource minimum initiation, (resMII), based on memory ports.

To fix this, as you mentioned, you need array partitioning. I don't believe
this is in the legup 4.0 release as this is a very new feature, you may need to
email Jason. Right now this would not happen automatically without explicitly
enabling a pass to do array partitioning.

-- 
Configure bugmail: http://legup.eecg.utoronto.ca/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.


More information about the Legup-bugs mailing list