Uploaded July 2026 | Updated September 2026, 2 weeks ago
Cross-architecture evaluation of ML and scientific-computing workloads using HPX (C++ standard library for parallelism and concurrency), benchmarking HPX's asynchronous tasking across RISC-V, ARM, and x86, including where each architecture excels, where gaps remain, and the implications for portable high-performance software.
Paper: arxiv.org/html/2604.00736v2
Cross-architecture evaluation of ML and scientific-computing workloads using HPX (C++ standard library for parallelism and concurrency), benchmarking HPX's asynchronous tasking across RISC-V, ARM, and x86, including where each architecture excels, where gaps remain, and the implications for portable high-performance software.
Paper: arxiv.org/html/2604.00736v2










