Table-A) CPU performance: cmsExp vs cmsExpVG, 11.0.p03; all uses SerialOnly RM 
Legend:
cmsExpVG   = cmsExp + Geant4/VecGeom.1.2.0
11.0.p03   = geant4-11-00-patch-03 - build with -DGEANT4_BUILD_MULTITHREADED=ON (D) 

Average (CPU Time)/Event: cmsExp  vs cmsExpVG
#----------------------------------------------------------------------------
#  cmsExp              cmsExpVG        100*(cmsExpVG-cmsExp)/cmsExp  sample
#
#  mean      error     mean      error    %diff    error
#----------------------------------------------------------------------------

   225.55    1.59      213.01    1.74     -5.5      1.3        higgs.FTFP_BERT.1400.4

#----------------------------------------------------------------------------


Table-B) CPU performance: 11.0.p03 vs 11.0.p02

Legend:
11.0.p02    = geant4-11-00_patch-02 - build with -DGEANT4_BUILD_MULTITHREADED=ON (D), Serial RM in app
11.0.p03    = geant4-11-00-patch-03 - build with -DGEANT4_BUILD_MULTITHREADED=ON (D), Serial RM in app

Average (CPU Time)/Event:  11.0.p02 vs 11.0.p03
#----------------------------------------------------------------------------
#  110p02                  110p03           100*(p03-p02)/p02  sample
#
#  mean      error         mean      error       %diff error
#----------------------------------------------------------------------------

  216.9900   1.6921 	  213.0100   1.7410 	  -1.8  1.1    higgs.FTFP_BERT.1400.4 

#----------------------------------------------------------------------------


Table-C) Total Memory : 11.0.p02 vs  11.0.p03

#-----------------------------------------------------------------------------
# total           memory     100*(p03-p02)/p02 
# 110p02          110p03       %diff           sample
#-----------------------------------------------------------------------------

  10262.5000 	  10148.1000 	-1.1	 higgs.FTFP_BERT.1400.4 
       
#-----------------------------------------------------------------------------