ParDcuhre, our CUDA-enabled parallel solution for multivariate integration, is derived from DCUHRE [1, 2] which employs the global adaptive algorithm over a hyperrectangular region in estimating a given integrand. However, with DCUHRE, as the integral dimension increases, the number of region evaluation points increases exponentially. ParDcuhre improves on DCUHRE by implementing an evaluation-point-level parallelization strategy in comparison to the region-level parallelization in DCUHRE, and pushes the integral dimension threshold well above dimension 15 (the highest dimension used for testing by DCUHRE’s authors). Furthermore, ParDcuhre achieves quadruple-digit speedups with selected integrals, yielding estimates comparable to published results.