chipStar: Making HIP/CUDA applications cross-vendor portable by building on open standards

Paulius Velesko, Pekka Jääskeläinen, Henry Linjamäki, Michal Babej, Peng Tu, Sarbojit Sarkar, Ben Ashbaugh, Colleen Bertoni, Jenny Chen, Philip C. Roth, Wael Elwasif, Rahulkumar Gayatri, Jisheng Zhao, Karol Herbst, Kevin Harms, Brice Videau, PGLC Consulting, North Las Vegas, Tampere University, Intel Corporation, Santa Clara, Argonne National Laboratory, Purdue University, West Lafayette, Oak Ridge National Laboratory, Oak Ridge, National Energy Research Scientific Computing Center, Lawrence Berkeley National Laboratory, Georgia Institute of Technology, Red Hat

The International Journal of High Performance Computing Applications

Published online on February 14, 2026

Abstract

The International Journal of High Performance Computing Applications, Ahead of Print.
We describechipStar, an open source software stack that enables building unmodified CUDA and HIP programs into binaries that rely solely on open cross-vendor compute standards OpenCL and SPIR-V. The relevant technical aspects ofchipStarand the feature ...