get index of the first dimension of i-th block(task)

Get specified hist, initialize hist by zeros if it wasn't used before

* @brief Some old prediction methods accept the ntree_limit parameter and they use 0 to * indicate no limit. */

! * \brief Get weight of each instances. * \param i Instance index. * \return The weight. */

* @brief Check whether we should configure `cache_host_ratio`. * * Defined by @ref AutoHostRatio . */

* @param Alignment for resource read stream and aligned write stream. */

! * \brief step 1: initialize the helper, with hint of number keys * and thread used in the construction * \param max_key numb

Create the internal model parameter from user inputs, this requires the user input to be initialized first. Don't apply the link function if the base

see the dummy implementation of `poll` in rabit for more info.

! * \brief decides whether we can replace current entry with the given statistics * * This function gives better priority to lower index wh

* @brief perform poll on the set defined, read, write, exception * * @param timeout specify timeout in seconds. Block if negative. */

Return the set of nodes to be expanded This set has no dependencies between entries so they may be expanded in parallel or asynchronously

push element to the queue, return false if the queue is full and need to be flushed

* @brief Get the underlying resource. */

* \brief Set meta info about categorical features. * * \param has_cat Do we have categorical feature in the data? * \param max_cat The maximu

utils to get/set position, with encoded format return decoded position encode the encoded position value for ridx

! * \brief calculate the sigmoid of the input. * \param x input parameter * \return the transformed value. */

* @brief Returns the view start and length for the i^th page. */

* @brief Get a @ref TensorView for this tensor. */

Note: the idea for separating this function from the one above is to be able to trigger all R allocations first before doing non-R allocations. */