work_queue_stats Struct Reference

Statistics describing a work queue. More...

#include <work_queue.h>

Data Fields

int total_workers_connected
 Total number of workers currently connected to the master.
int workers_init
 Number of workers initializing.
int workers_idle
 Number of workers that are not running a task.
int workers_busy
 Number of workers that are running at least one task.
int total_workers_joined
 Total number of worker connections that were established to the master.
int total_workers_removed
 Total number of worker connections that were terminated by the master.
int tasks_waiting
 Number of tasks waiting to be run.
int tasks_running
 Number of tasks currently running.
int tasks_complete
 Number of tasks waiting to be returned to user.
int total_tasks_dispatched
 Total number of tasks dispatch to workers.
int total_tasks_complete
 Total number of tasks completed and returned to user.
int total_tasks_failed
 Total number of tasks completed and returned to user with result other than WQ_RESULT_SUCCESS.
int total_tasks_cancelled
 Total number of tasks cancelled.
timestamp_t start_time
 Absolute time at which the master started.
timestamp_t total_send_time
 Total time in microseconds spent in sending data to workers.
timestamp_t total_receive_time
 Total time in microseconds spent in receiving data from workers.
timestamp_t total_good_transfer_time
 Total time in microseconds spent in sending and receiving data to workers for tasks with result WQ_RESULT_SUCCESS.
timestamp_t total_execute_time
 Total time in microseconds workers spent executing completed tasks.
timestamp_t total_good_execute_time
 Total time in microseconds workers spent executing successful tasks.
int64_t total_bytes_sent
 Total number of file bytes (not including protocol control msg bytes) sent out to the workers by the master.
int64_t total_bytes_received
 Total number of file bytes (not including protocol control msg bytes) received from the workers by the master.
double efficiency
 Parallel efficiency of the system, sum(task execution times) / sum(worker lifetimes).
double idle_percentage
 The fraction of time that the master is idle waiting for workers to respond.
int capacity
 The estimated number of workers that this master can effectively support.
double bandwidth
 Average network bandwidth in MB/S observed by the master when transferring to workers.
int64_t total_cores
 Total number of cores aggregated across the connected workers.
int64_t total_memory
 Total memory in MB aggregated across the connected workers.
int64_t total_disk
 Total disk space in MB aggregated across the connected workers.
int64_t total_gpus
 Total number of GPUs aggregated across the connected workers.
int64_t committed_cores
 Committed number of cores aggregated across the connected workers.
int64_t committed_memory
 Committed memory in MB aggregated across the connected workers.
int64_t committed_disk
 Committed disk space in MB aggregated across the connected workers.
int64_t committed_gpus
 Committed number of GPUs aggregated across the connected workers.
int64_t min_cores
 The lowest number of cores observed among the connected workers.
int64_t max_cores
 The highest number of cores observed among the connected workers.
int64_t min_memory
 The smallest memory size in MB observed among the connected workers.
int64_t max_memory
 The largest memory size in MB observed among the connected workers.
int64_t min_disk
 The smallest disk space in MB observed among the connected workers.
int64_t max_disk
 The largest disk space in MB observed among the connected workers.
int64_t min_gpus
 The lowest number of GPUs observed among the connected workers.
int64_t max_gpus
 The highest number of GPUs observed among the connected workers.
int workers_ready
int workers_full
int total_worker_slots
int avg_capacity

Detailed Description

Statistics describing a work queue.


Field Documentation

Total number of workers currently connected to the master.

Number of workers initializing.

Number of workers that are not running a task.

Number of workers that are running at least one task.

Total number of worker connections that were established to the master.

Total number of worker connections that were terminated by the master.

Number of tasks waiting to be run.

Number of tasks currently running.

Number of tasks waiting to be returned to user.

Total number of tasks dispatch to workers.

Total number of tasks completed and returned to user.

Total number of tasks completed and returned to user with result other than WQ_RESULT_SUCCESS.

Total number of tasks cancelled.

Absolute time at which the master started.

Total time in microseconds spent in sending data to workers.

Total time in microseconds spent in receiving data from workers.

Total time in microseconds spent in sending and receiving data to workers for tasks with result WQ_RESULT_SUCCESS.

Total time in microseconds workers spent executing completed tasks.

Total time in microseconds workers spent executing successful tasks.

Total number of file bytes (not including protocol control msg bytes) sent out to the workers by the master.

Total number of file bytes (not including protocol control msg bytes) received from the workers by the master.

Parallel efficiency of the system, sum(task execution times) / sum(worker lifetimes).

The fraction of time that the master is idle waiting for workers to respond.

The estimated number of workers that this master can effectively support.

Average network bandwidth in MB/S observed by the master when transferring to workers.

Total number of cores aggregated across the connected workers.

Total memory in MB aggregated across the connected workers.

Total disk space in MB aggregated across the connected workers.

Total number of GPUs aggregated across the connected workers.

Committed number of cores aggregated across the connected workers.

Committed memory in MB aggregated across the connected workers.

Committed disk space in MB aggregated across the connected workers.

Committed number of GPUs aggregated across the connected workers.

The lowest number of cores observed among the connected workers.

The highest number of cores observed among the connected workers.

The smallest memory size in MB observed among the connected workers.

The largest memory size in MB observed among the connected workers.

The smallest disk space in MB observed among the connected workers.

The largest disk space in MB observed among the connected workers.

The lowest number of GPUs observed among the connected workers.

The highest number of GPUs observed among the connected workers.

Deprecated:
Use capacity instead.

The documentation for this struct was generated from the following file:

Generated on 29 Apr 2015 for cctools by  doxygen 1.6.1