Actionable insights tagged with ‘offload’ in Parallelware Analyzer:
PWD003: Missing array range in data copy to the GPU
PWD004: Out-of-memory-bounds array access
PWD005: Array range copied to or from the GPU does not cover the used range
PWD006: Missing deep copy of non-contiguous data to the GPU
PWD009: Incorrect privatization in OpenMP parallel region
PWD010: Incorrect sharing in OpenMP parallel region
PWD011: Missing OpenMP last private clause
PWR009: Use OpenMP teams to offload work to GPU
PWR013: Avoid copying unused variables to the GPU
PWR015: Avoid copying unnecessary array elements to or from the GPU
PWR026: Annotate function for OpenMP offload
PWR027: Annotate function for OpenACC offload
RMK002: Loop nesting that might benefit from hybrid parallelization using offloading and SIMD
