* work, and any subsequent call receives the same handle. With real
* MPI, communication is needed to coordinate the results. In all
* cases, any thread within a process may use the returned handle. */
-gmx_hw_info_t *gmx_detect_hardware(const gmx::MDLogger &mdlog,
- const PhysicalNodeCommunicator &physicalNodeComm);
+gmx_hw_info_t* gmx_detect_hardware(const gmx::MDLogger& mdlog,
+ const PhysicalNodeCommunicator& physicalNodeComm);
//! Return whether compatible GPUs were found.
-bool compatibleGpusFound(const gmx_gpu_info_t &gpu_info);
+bool compatibleGpusFound(const gmx_gpu_info_t& gpu_info);
} // namespace gmx