The proposed parallel state transformation procedures (PSTP) of fiber beam-column elements and multi-layered shell elements, combined with the parallel factorization of Jacobian (PF), are incorporated into a finite element program. In PSTP, elements are classified into different levels of workload prior to state determination in order to balance workload among different threads. In PF, the multi-threaded version of OpenBLAS is adopted to compute super-nodes. A case study on four super high-rise buildings, i.e. S1~S4, has demonstrated that the combination of PSTP and PF does not have any observable influence on computational accuracy. As number of elements and DOFs increases, the ratio of time consumed in the formation of the Jacobian to that consumed in the solution of algebraic equations tends to decrease. The introduction of parallel solver can yield a substantial reduction in computational cost. Combination of PSTP and PF can give rise to a further significant reduction. The acceleration ratios associated with PSTP do not exhibit a significant decrease as PGA level increases. Even PGA level is equal to 2.0g, PSTP still can result in acceleration ratios of 2.56 and 1.92 for S1 and S4, respectively. It is verified that it is more effective to accelerate analysis by reducing the time spent in solving algebraic equations rather than reducing that spent in the formation of the Jacobian for super high-rise buildings.