MaxMemoryPreload shows variations of tens of MB in some cases

[makortel]

E.g. in #46955 (comment), whose tests recompiled pretty much the universe, resulted e.g.

[makortel]

assign core

[cmsbuild]

New categories assigned: core

@Dr15Jones,@makortel,@smuzaffar 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]

Seems to be easily reproducible in 15_0_0_pre2 in an empty developer area (e.g. RECO step of workflow 9.0)

[makortel]

I managed to catch the cause with IgProf by comparing the MEM_LIVE after 9th event of wf 9.0 step 3. The main culprit seems to be in ONNXRuntime constructor via BoostedJetONNXJetTagsProducer: test1 vs test2 shows 66 MB difference.