[Phase2 RelVal] ~12 million allocations from MTDGeometricTimingDetESModule::produce()

[makortel]

From #45854 (comment)

The MTDGeometricTimingDetESModule::produce() makes ~12 million memory allocations, of which nearly all come from string construction in DDNameFilter::accept() (IgProf profile)

cmssw/Geometry/MTDNumberingBuilder/plugins/DDCmsMTDConstruction.cc

Lines 28 to 29 in f71ff4c

	bool accept(const DDExpandedView& ev) const final {
	std::string currentName(ev.logicalPart().name().fullname());

The ev.logicalPart().name().fullname() returns a newly constructed string

cmssw/DetectorDescription/Core/interface/DDName.h

Lines 40 to 43 in f71ff4c

	/** Returns a string complete of the \a namespace and \a name separated by ":".
	Most likely you want to use ns() and / or name() methods instead.
	*/
	const std::string fullname() const { return ns() + ":" + name(); }

Especially how the DDNameFilter is used

cmssw/Geometry/MTDNumberingBuilder/plugins/DDCmsMTDConstruction.cc

Lines 48 to 51 in f71ff4c

	DDNameFilter filter;
	filter.add("mtd:");
	filter.add("btl:");
	filter.add("etl:");

I'd suggest to compare the namespace and name parts of DDName separately to avoid the string construction.

[makortel]

assign Geometry/MTDNumberingBuilder

[cmsbuild]

New categories assigned: geometry,upgrade

@bsunanda,@civanch,@Dr15Jones,@kpedro88,@makortel,@mdhildreth,@Moanwar,@srimanob,@subirsarkar you have been requested to review this Pull request/Issue and eventually sign? Thanks

[cmsbuild]

cms-bot internal usage

[cmsbuild]

A new Issue was created by @makortel.

@Dr15Jones, @antoniovilela, @makortel, @mandrenguyen, @rappoccio, @sextonkennedy, @smuzaffar can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

[makortel]

@cms-sw/mtd-dpg-l2

[makortel]

type performance-improvements

[makortel]

Addressed in #46531

[makortel]

The PR was merged