Experimental Physics and Industrial Control System
Ken,
I am getting a better handle on how to get meaningful data out of
quantify. One of my quantify runs this afternoon reported that the
example portable server was spending a whopping 55% of its total CPU
time in the CRTL pool management routines. Closer inspection revealed
that this was nearly all related to new and delete (malloc and free) for
the data structure that gets queued for each monitor update. In R3.14 I
have a template for creating free lists in C++, but didn't when I wrote
the portable server and therefore none of its data structures were
originally obtained from free lists. Clearly this needs to change. Note
however that this is fixing an issue that existed also in the R3.13
portable ca server, and all previous versions. So this does not explain
the performance slow down of the gateway in R3.14, but still it is an
encouraging result and I expect that we can make a big improvement based
on this alone.
Jeff
__________________________________________________________
Jeff Hill Internet [email protected]
LANL MS H820 Voice 505 665 1831
Los Alamos NM 87545 USA FAX 505 665 5107
- Navigate by Date:
- Prev:
Re: base max thread priority Eric Norum
- Next:
Re: base max thread priority Marty Kraimer
- Index:
<2002>
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
- Navigate by Thread:
- Prev:
Re: Gateway Ned D. Arnold
- Next:
CROSS_COMPILER_TARGET_ARCHS default change Janet Anderson
- Index:
<2002>
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024