Skip to content

Optimization_4x4_14

Jianyu Huang edited this page Aug 2, 2016 · 4 revisions

Copy the contents of file MMult_4x4_13.c into a file named MMult_4x4_14.c and change the contents:

|| from || to || ||<^> -Include(HowToOptimizeGemm/Details/MMult_4x4_13)- ||<^> -Include(HowToOptimizeGemm/Details/MMult_4x4_14)- ||

Change the first lines in the makefile to {{{ OLD := MMult_4x4_13 NEW := MMult_4x4_14 }}}

  • make run {{{ octave:3> PlotAll % this will create the plot }}}

This time the performance graph will look something like

ImageLink(http://www.cs.utexas.edu/users/rvdg/HowToOptimizeGemm/Graphs/compare_MMult-4x4-13_MMult-4x4-14.png,http://www.cs.utexas.edu/users/rvdg/HowToOptimizeGemm/Graphs/compare_MMult-4x4-13_MMult-4x4-14.png,width=40%)

We now pack the kx4 blocks of B. Notice that in this version, the panels are packed repeatedly, adversely affecting performance.

Clone this wiki locally