Function vm_engine_iter_outer_reduce_task

numexpr3/interpreter.cpp:304–336

}

static int
vm_engine_iter_outer_reduce_task(NpyIter *iter, 
                const NumExprObject *params, int *pc_error, char **errorMessage)
{
    NpyIter_IterNextFunc *iterNext;
    npy_intp task_size, *sizePtr;
    char **iterDataPtr;
    npy_intp *iterStrides;

    iterNext = NpyIter_GetIterNext(iter, errorMessage);
    if (iterNext == NULL) {
        return -1;
    }

    sizePtr = NpyIter_GetInnerLoopSizePtr(iter);
    iterDataPtr = NpyIter_GetDataPtrArray(iter);
    iterStrides = NpyIter_GetInnerStrideArray(iter);

    task_size = *sizePtr;
    // First do all the blocks with a compile-time fixed size.
    // This makes a big difference (30-50% on some tests).
    // RAM: Not-so-much with vectorized loops
    
    while( task_size > 0 ) {
#define NO_OUTPUT_BUFFERING
#include "interp_body_GENERATED.cpp"
#undef NO_OUTPUT_BUFFERING
        iterNext(iter);
        task_size = *sizePtr;   
    }
    
    return 0;
}

// Parallel iterator version of VM engine 
// This function fills out the global state and then unlocks the mutexes