// according to this technique.
//
// Fair queuing for server requests is inspired by the fair queuing
// technique from the world of networking.  You can find a good paper
// on that at https://dl.acm.org/citation.cfm?doid=75247.75248 or
// http://people.csail.mit.edu/imcgraw/links/research/pubs/networks/WFQ.pdf
// and there is an implementation outline in the Wikipedia article at
// https://en.wikipedia.org/wiki/Fair_queuing .
//

 * conventional bounded queues, which either block or reject new elements when full.
 *
 * <p>This implementation is based on the <a
 * href="http://portal.acm.org/citation.cfm?id=6621">min-max heap</a> developed by Atkinson, et al.
 * Unlike many other double-ended priority queues, it stores elements in a single array, as compact
 * as the traditional heap data structure used in {@link PriorityQueue}.
 *

 * conventional bounded queues, which either block or reject new elements when full.
 *
 * <p>This implementation is based on the <a
 * href="http://portal.acm.org/citation.cfm?id=6621">min-max heap</a> developed by Atkinson, et al.
 * Unlike many other double-ended priority queues, it stores elements in a single array, as compact
 * as the traditional heap data structure used in {@link PriorityQueue}.
 *

  }];

  let description = [{
This op uses the algorithm by Marsaglia et al. to acquire samples via
transformation-rejection from pairs of uniform and normal random variables.
See http://dl.acm.org/citation.cfm?id=358414
  }];

  let arguments = (ins
    Arg<TF_I32OrI64Tensor, [{1-D integer tensor. Shape of independent samples to draw from each
distribution described by the shape parameters given in alpha.}]>:$shape,