Actually, residential energy consumption is expressed in final energy terms. This is the actual consumption by end-users of energy carriers (e.g. electricity, gasoline, diesel, natural gas, etc...). The sum of all these end-use consumption is the Total Final Consumption (or TFC).
This is to be distinguished from primary energy sources (crude oil, coal, natural gas, etc...), which are used to produce the energy carriers that end-users consume (e.g. in a power plant to produce electricity, or in a refinery to produce gasoline). The sum of all primary energy sources consumption is the Total Primary Energy Supply (TPES).
Since residential energy consumption is reported in final terms, the correct denominator to calculate the share you were looking for is TFC. Residential energy consumption thus accounts for 17.9% of TFC, versus 17.7% for industry.
The non-residential sector is denominated in IEA terminology as 'Commercial and Public Services'. This sector encompasses office buildings, hotels & restaurants, education & research, hospital & healthcare, sports facilities and public administrations. In short, buildings that house the tertiary sector. Therefore it can't be identified at all with the industry (btw industrial buildings consumption are accounted for in the 'industry' sector).
In total, residential and non-residential buildings account for 32% of TFC in the US, second only to transport (39.5%).
Disclosure: I've worked as an analyst at the IEA for 4 years.