Commit ba83689b authored by Ben Newton's avatar Ben Newton
Browse files

Added to readme

parent 83259754
Pipeline #96082 failed with stage
......@@ -14,6 +14,7 @@ Things that should be done (TODO):
6. There are some To-Dos marked to check that Input index is not palced where output index is actually needed. See ThreadIndicesPointNeighborhood.h
7. It may be nice to replace the SimpleIndexPortal in ThreadIndicesTopologyMap.h with a different construct to do the straight across mapping. (see todo comment)
CHANGES by file
--------------------------------------------------------------------------------------------
vtkm/CellShape.h
......@@ -213,6 +214,9 @@ git clone git@gitlab.kitware.com:BenNewton/DotProduct.git
See the README file in the ComputeAirPressure.git repo for more info on running the comparison script.
Run with: python compre.py -k
The -k says keep the build files (if you want to examine the optimization output, or run manually, etc)
##############################################################
Reading the optimization output for various benchmark tests.
Information on whether or not, and how well the main worklet loop has been vectorized can be found in the optimization report when using the intel compiler. If the option is enabled to output the report (the 512 block size outputs the report when using the compare.py script) then the following cat/grep commands will isolate the lines associated with the worklet loop. NOTE, the line numbers in WorkletInvokeFunctorDetail changed since I ran these, so the line numbers will need to be updated. Searching without a line number ....h(' and looking through the output may help you find the right line number.
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment