This paper proposes a fully decentralized distributed convex optimal power flow model for inverter-based distributed energy resources (DERs) integrated electric distribution networks based on Semi-Definite Programming (SDP) and alternating direction method of multipliers (ADMM) namely (SDP D-ADMM). The proposed approach is based on the SDP relaxed branch flow model of distribution networks within an auto-tuned accelerated decentralized ADMM architecture. The approach is based on dividing the power grid network into subproblems representing individual areas by interchanging minimum network information. In the proposed model the requirement of a central processor is also waived thus making the proposed approach more robust toward cyber-attacks. The effectiveness and scalability of the proposed method are validated by implementing modified IEEE 123 and IEEE 8500 bus systems with different levels of DER penetration. It has been observed that the proposed architecture outperforms other distributed optimization variants in terms of accuracy, global optimality, scalability, and computational time.