Intelligent　reflecting　surfaces　(IRSs)　have　been　proposed　in　recent　years　as　a　promising　technology　to　enhance　signal　quality　at　high　frequencies　and　save　energy.　In　this　paper,　a　Poisson　bipolar　network　model　with　line　segments　is　used　to　analyze　the　energy　efficiency　(EE)　of　an　IRS-assisted,　large-scale　network.　Specifically,　we　investigate　the　performance　impact　of　the　IRS　configuration,　in　particular,　the　number　of　IRS　elements　and　the　phase-shifting　resolution　of　each　element.　Using　customized　energy　consumption　and　channel　estimation　models,　we　obtain　the　theoretical　trade-off　between　signal　quality　and　energy　consumption　as　a　function　of　these　IRS　configurations.　The　optimal　number　of　elements　and　phase-shifting　resolution　of　the　IRS　are　also　derived.　Our　results　show　that　IRS　technology　has　great　potential　for　improving　the　EE　of　dense　networks　if　their　static　energy　consumption　is　small　enough.　Simulation　results　verify　the　accuracy　of　the　obtained　theoretical　results.　IEEE