Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 5 Next »

  1. Interface #1

    1. Visibility: Public

    2. Change Control: Stable

    3. Synopsis: PBS hook power control module

    4. A new class “pbs.Power” will be made available that will provide power functionality.  A hook will be able to access it via python import.

    5. Reference to more detail on the interface. The following define the PMI operations available:

    6. pmi_activate_profile

      1. Activate a given power profile on a set of hosts on behalf of a given job.  The parameter “profile_name” is a string containing the name of a profile.  The parameter “hosts” is a list containing strings that specify hostnames.  The parameter “job” is a PBS job object.  If the hosts parameter is not specified, the hosts will be calculated from the job object.  If the job parameter is not specified, the pbs.event().job object will be used.

      2. The return type is bool where True indicates success and False indicates the request was made without an indication from the PMI if it was successful or not.

      3. If an error occurs where it is appropriate for some or all of the job vnodes to be marked offline, this may be done before an exception is raised.

      4. If an error occurs where it is appropriate for the supported profile names for some or all of the job vnodes to be refreshed, this may be done before an exception is raised.

    7. pmi_get_usage

      1. Retrieve power usage for a job.  The parameter “job” is a PBS job object.

      2. The return will be a float which gives the cumulative energy usage for the job at the time of the call in kilowatt-hours (kWh).  If no power usage information is available, None is returned.

    8. pmi_deactivate_profile

      1. Inform the PMI that a job is no longer active.  This would be used when a job is suspended or terminated.  The parameter “job” is a PBS job object.  If it is not specified, the pbs.event().job object will be used.

      2. The return type is bool where True indicates success and False indicates the request was made without an indication from the PMI if it was successful or not.

    9. pmi_query

      1. Return information that matches a request type.  The parameter “query_type” is used to specify what should be returned.  The only value for  query_type is PMI_QUERY_PROFILE, and the return will be a list of strings giving profile names supported by the PMI.

    10. pmi_connect

      1. Connect to the PMI.  The parameter “endpoint” defaults to None and is a string which will be meaningful to the PMI.  The parameter “port” defaults to None and is an integer.  A typical usage would be “endpoint” specifying a hostname and “port” giving a network port for a network service connection.

      2. Currently the connection/disconnection will be done per hook instead of creating a long lasting session.

      3. Nothing is returned, the connection information is maintained in an instantiation of the Power class.

      4. If the endpoint or port parameters are not specified, the underlying code specific to the PMI will determine the connection details.

    11. pmi_disconnect

      1. Disconnect from the PMI.  There are no parameters needed since each instance of the Power class is associated to a backend power management interface.

    12. Exceptions

      1. InternalError - returned in cases where the underlying cause of a failure cannot be determined.

      2. BackendError - the backend PMI call was unsuccessful.

    13. Power module initialization

      1. A string can optionally be passed to specify the name of the PMI to be used (see I.1.11).  By default, the type of PMI to be used will be determined automatically based on the type of hardware used.

    14. Comments on the interface

      1. Standing of the interface: new interface

      2. Interface type: Other

    15. Examples

      1. Activate a profile from a job specific event.

        p = pbs.Power()

        p.pmi_connect(“power_master”)

        p.pmi_activate_profile(“LOW”)

        p.pmi_disconnect()

      2. Get profile name list.

        import pbs

        p = pbs.Power()

        p.pmi_connect(port=3564)

        pnames = p.pmi_query(p.PMI_QUERY_PROFILE)

        p.pmi_disconnect()

      3. Deactivate profile on a specific job.

        import pbs

        p = pbs.Power()

        badjob = pbs.server().job(“10”)

        p.pmi_connect()

        p.pmi_deactivate_profile(job=badjob)

        p.pmi_disconnect()

  2. Interface #2
    1. Visibility: Public
    2. Change Control: Stable
    3. Synopsis: Generically applicable server power_provisioning  flag
    4. Reference to more detail on the interface.
      1. The  power_provisioning boolean server attribute will have a default of unset, be visible to all and changeable by an administrator.  When it is set True, PMI operations may take place if allowed by power_enable flag (see I.1.10).  If it is unset or set False, no PMI operations will take place on any vnode.

      2. Use qmgr to set the power_provisioning flag true or false.  For example:   qmgr -c “set server power_provisioning = true”

    5. Comments on the interface
      1. Standing of the interface: new interface
      2. Interface type: Other
  3. Interface #3
    1. Visibility: Public
    2. Change Control: Stable
    3. Synopsis: Generically applicable energy usage for a job
      1. Add a new attribute for a job: resources_used.energy
    4. Reference to more detail on the interface.
      1. The type will be float.
      2. The units will be kWh.  For example:  resources_used.energy=64.2
      3. The resources_used.energy value will only be updated when PMI operations are allowed on the vnodes used by the job.
    5. Comments on the interface
      1. Standing of the interface: new interface
      2. Interface type: Other
  4. A.1.4      Interface #5
  5. A.1.4.1          Visibility: Public
  6. A.1.4.2          Change Control: Stable
  7. A.1.4.3          Synopsis: Generically applicable resource “eoe”
  8. A.1.4.3.1               A new resource similar to “aoe” is added to both jobs and vnodes to specify the energy operational environment.
  9. A.1.4.4          Reference to more detail on the interface.
  10. A.1.4.4.1               Is added to default resource list of scheduler in sched_config file.
  11. A.1.4.4.2               It is a non-consumable resource.
  12. A.1.4.4.3               It is of type resource, added to attribute resources_available. e.g. resources_available.eoe=”low,med,high”. It is a string array.
  13. A.1.4.4.4               Contains list of all power profile names that are available on a vnode. By default, resources_available.eoe is unset.
  14. A.1.4.4.5               The list is visible to all but settable only by manager.
  15. A.1.4.4.6               Job Resource_List.eoe per chunk in –l select as –l select=1:ncpus:eoe=low.
  16. A.1.4.4.7               Only one eoe value can be active on a vnode at a time.
  17. A.1.4.4.8               A job Resource_List.eoe may be requested in a select statement but no more than one distinct value for the requested eoe is currently supported. i.e. -lselect=1:ncpus=1:eoe=med+1:ncpus=2:eoe=med
  18. A.1.4.4.9               If a Job request is made with more than one value for eoe (I.e. –l select=1:eoe=low+1:eoe=high), it will be rejected by qsub with the error “qsub: only one value of eoe is allowed”.
  19. A.1.4.4.10            A value for resources_available.eoe will not be automatically set on the system(s) where the PBS server and scheduler are running.
  20. A.1.4.4.11            If both an aoe and eoe are set for a job, the aoe setting will be processed first by the scheduler.
  21. A.1.4.4.12            DELETED
  22. A.1.4.4.13            The scheduler will not prempt a job with eoe set using suspend or checkpoint.
  23. A.1.4.5          Comments on the interface
  24. A.1.4.5.1               Standing of the interface: new interface
  25. A.1.4.5.2               Interface type: Other
  26. A.1.5      Interface #6
  27. A.1.5.1          Visibility: Public
  28. A.1.5.2          Change Control: Stable
  29. A.1.5.3          Synopsis: Generically applicable vnode attribute: current_eoe
  30. A.1.5.4          Reference to more detail on the interface.
  31. A.1.5.4.1               Identifies the eoe active on a vnode. It is of type String. By default, it is unset. It is settable only by manager and visible to all.
  32. A.1.5.4.2               A job J1 running with a eoe setting X will cause the value of current_eoe to be set  to X on the vnodes assigned to J1 that allow PMI operations.
  33. A.1.5.4.3               Manually changing current_eoe is unsupported.
  34. A.1.5.4.4               The scheduler can run a job requesting an eoe on vnodes with a current_eoe value that matches the job eoe.
  35. A.1.5.4.5               The scheduler can only run a job on a vnode where the current_eoe does not match the job eoe if no jobs are running on the vnode and PMI operations are allowed on the vnode.
  36. A.1.5.4.6               When a job ends the pmi_deactivate operation will take place if all the vnodes used by the job have no other jobs running and allow PMI operations.  At this point, current_eoe will be unset on all the vnodes used by the job.
  37. A.1.5.5          Comments on the interface
  38. A.1.5.5.1               Standing of the interface: new interface
  39. A.1.5.5.2               Interface type: Other
  40. A.1.6      Interface #7
  41. A.1.6.1          Visibility: Public
  42. A.1.6.2          Change Control: Experimental
  43. A.1.6.3          Synopsis: Cray specific job attribute: pstate
  44. A.1.6.4          Reference to more detail on the interface.
  45. A.1.6.4.1               Cray ALPS reservation setting for p-state.  See Basil 1.4 documentation.
  46. A.1.6.4.2               It is of type String. By default, it is unset. It is settable and visible to all PBS users.
  47. A.1.6.4.3               DELETED
  48. A.1.6.5          Comments on the interface
  49. A.1.6.5.1               Standing of the interface: new interface
  50. A.1.6.5.2               Interface type: Other
  51. A.1.7      Interface #8
  52. A.1.7.1          Visibility: Public
  53. A.1.7.2          Change Control: Experimental
  54. A.1.7.3          Synopsis: Cray specific job attribute: pgov
  55. A.1.7.4          Reference to more detail on the interface.
  56. A.1.7.4.1               Cray ALPS reservation setting for p-governor.  See Basil 1.4 documentation.
  57. A.1.7.4.2               It is of type String. By default, it is unset. It is settable and visible to all PBS users.
  58. A.1.7.4.3               DELETED
  59. A.1.7.5          Comments on the interface
  60. A.1.7.5.1               Standing of the interface: new interface
  61. A.1.7.5.2               Interface type: Other
  62. A.1.8      Interface #9
  63. A.1.8.1          Visibility: Public
  64. A.1.8.2          Change Control: Experimental
  65. A.1.8.3          Synopsis: Cray specific job attribute: pcap_node
  66. A.1.8.4          Reference to more detail on the interface.
  67. A.1.8.4.1               Cray capmc set_power_cap --node setting.  See capmc documentation.
  68. A.1.8.4.2               It is of type Int. By default, it is unset. It is settable and visible to all PBS users.
  69. A.1.8.4.3               A  negative value will result in an PBSE_BADATVAL error.
  70. A.1.8.4.4               DELETED
  71. A.1.8.5          Comments on the interface
  72. A.1.8.5.1               Standing of the interface: new interface
  73. A.1.8.5.2               Interface type: Other
  74. A.1.9      Interface #10
  75. A.1.9.1          Visibility: Public
  76. A.1.9.2          Change Control: Experimental
  77. A.1.9.3          Synopsis: Cray specific job attribute: pcap_accelerator
  78. A.1.9.4          Reference to more detail on the interface.
  79. A.1.9.4.1               Cray capmc set_power_cap --accel setting.  See capmc documentation.
  80. A.1.9.4.2               It is of type Int. By default, it is unset. It is settable and visible to all PBS users.
  81. A.1.9.4.3               A negative value will result in an PBSE_BADATVAL error.
  82. A.1.9.4.4               DELETED
  83. A.1.9.5          Comments on the interface
  84. A.1.9.5.1               Standing of the interface: new interface
  85. A.1.9.5.2               Interface type: Other
  86. A.1.10   Interface #11
  87. A.1.10.1        Visibility: Public
  88. A.1.10.2        Change Control: Experimental
  89. A.1.10.3        Synopsis: Generically applicable vnode power enable flag
  90. A.1.10.4        Reference to more detail on the interface.
  91. A.1.10.4.1            The  power_enable boolean vnode attribute will have a default of unset, be visible to all and changeable by an administrator.
  92. A.1.10.4.2            Use qmgr to set the power_provisioning flag true or false.  For example:
  93.                                                 qmgr -c “set node bigbox power_enable = true”
  94. A.1.10.4.3            When it is set True, PMI operations may take place on the vnode.  If it is unset or set False, no PMI operations are allowed to take place on the vnode.
  95. A.1.10.5        Comments on the interface
  96. A.1.10.5.1            Standing of the interface: new interface
  97. A.1.10.5.2            Interface type: Other
  98. A.1.11   Interface #12
  99. A.1.11.1        Visibility: Private
  100. A.1.11.2        Change Control: Unstable
  101. A.1.11.3        Synopsis: Expose the hook PMI structure to allow additions to the supported PMI list.
  102. A.1.11.4        Reference to more detail on the interface.
  103. A.1.11.4.1            The PBS “power” hook can be modified to specify a PMI name in the pbs.Power() instantiation in the init_power function.  For example, the code below would cause the new file described in I.1.11.4.2 to be used by the hook:
  104.                                                 power = pbs.Power(“ipmitool”)
  105. A.1.11.4.2            Python code patterned after the file PBS_EXEC/lib/python/altair/pbs/v1/_pmi_none.py must be placed in a file where none is replaced by the PMI name being implemented.  For example:
  106.                                                 # cd $PBS_EXEC/lib/python/altair/pbs/v1
  107.                                                 # cp _pmi_none.py _pmi_ipmitool.py
  108.                                                 # vi _pmi_ipmitool.py
  109. A.1.11.4.3            The defined functions must all be present: __init__, _pmi_connect, _pmi_disconnect, _pmi_get_usage, _pmi_query, _pmi_activate_profile, _pmi_deactivate_profile.  These all have the same arguments as those in I.1.1 except the function name has an intial underbar ('_').
  110. A.1.11.5        Comments on the interface
  111. A.1.11.5.1            Standing of the interface: new interface
  112. A.1.11.5.2            Interface type: Other
  113. A.1.12   Interface #13
  114. A.1.12.1        Visibility: Public
  115. A.1.12.2        Change Control: Experimental
  116. A.1.12.3        Synopsis: Mom log using logjobmsg when a job ends and the value of current_eoe is unset.
  117. A.1.12.4        Reference to more detail on the interface.
  118. A.1.12.4.1            Example:
  119.                                     11/19/2014 14:44:15;0008;pbs_python;Job;165.bigcray;PMI: reset current_eoe
  120. A.1.12.5        Comments on the interface
  121. A.1.12.5.1            Standing of the interface: new interface
  122. A.1.12.5.2            Interface type:  Log message
  123. A.1.13   Interface #14
  124. A.1.13.1        Visibility: Public
  125. A.1.13.2        Change Control: Experimental
  126. A.1.13.3        Synopsis: When the energy for a job on an SGI system is obtained, it will be logged by MoM using  logjobmsg.
  127. A.1.13.4        Reference to more detail on the interface.
  128. A.1.13.4.1            Example:
  129. 11/06/2014 18:35:26;0008;pbs_python;Job;4856.iceberg;SGI: energy 1.456kWh
  130. A.1.13.5        Comments on the interface
  131. A.1.13.5.1            Standing of the interface: new interface
  132. A.1.13.5.2            Interface type:  Log message
  133. A.1.14   Interface #15
  134. A.1.14.1        Visibility: Public
  135. A.1.14.2        Change Control: Experimental
  136. A.1.14.3        Synopsis: The Cray capmc command invocations will be logged by MoM using LOG_DEBUG with the keyword “launch”.
  137. A.1.14.4        Reference to more detail on the interface.
  138. A.1.14.4.1            Example:
  139. 11/19/2014 15:20:58;0006;pbs_python;Hook;pbs_python;Cray: 167.bigcray launch: /opt/cray/capmc/default/bin/capmc get_node_energy_counter --nids 0
  140. A.1.14.5        Comments on the interface
  141. A.1.14.5.1            Standing of the interface: new interface
  142. A.1.14.5.2            Interface type:  Log message
  143. A.1.15   Interface #16
  144. A.1.15.1        Visibility: Public
  145. A.1.15.2        Change Control: Experimental
  146. A.1.15.3        Synopsis:  Following a successful Cray capmc invocation, a message will be logged by MoM using LOG_WARNING if the time used by capmc is greater than 30 seconds.
  147. A.1.15.4        Reference to more detail on the interface.
  148. A.1.15.4.1            Example:
  149. 11/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python; 21.bigcray;launch: finished  in 156 seconds
  150. A.1.15.5        Comments on the interface
  151. A.1.15.5.1            Standing of the interface: new interface
  152. A.1.15.5.2            Interface type:  Log message
  153. A.1.16   Interface #17
  154. A.1.16.1        Visibility: Private
  155. A.1.16.2        Change Control: Unstable
  156. A.1.16.3        Synopsis:  If Cray capmc writes anything to stderr, the first line will be logged by MoM using LOG_WARNING after the “launch” message.
  157. A.1.16.4        Reference to more detail on the interface.
  158. A.1.16.4.1            Cray has not documented the possible stderr output from capmc.
  159. A.1.16.4.2            Example:
  160. 11/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python; 21.bigcray;launch stderr: i fell and cannot get up
  161. A.1.16.5        Comments on the interface
  162. A.1.16.5.1            Standing of the interface: new interface
  163. A.1.16.5.2            Interface type:  Log message
  164. A.1.17   Interface #18
  165. A.1.17.1        Visibility: Private
  166. A.1.17.2        Change Control: Unstable
  167. A.1.17.3        Synopsis:  When Cray capmc is run with the argument “get_node_energy_counter”, the node count is checked and if it is wrong, a message will be logged by MoM using logjobmsg.
  168. A.1.17.4        Reference to more detail on the interface.
  169. A.1.17.4.1            The same command will be run one additional time if an error is seen.  No message will be logged for the first error.  If an error occurs after the second attempt, a message is logged.
  170. A.1.17.4.2            For example:
  171. 11/19/2014 15:20:05;0008;pbs_python;Job;166.centos1;error: node count 2, should be 1

 

  1. A.1.17.4.3            The output from capmc should include a node count.  If it does not, the messages will show “not set” instead of a number.
  2. A.1.17.4.4            Example:
  3. 11/19/2014 15:20:05;0008;pbs_python;Job;166.centos1;node count not set, should be 1

 

  1. A.1.17.5        Comments on the interface
  2. A.1.17.5.1            Standing of the interface: new interface
  3. A.1.17.5.2            Interface type:  Log message
  4. A.1.18   Interface #19
  5. A.1.18.1        Visibility: Public
  6. A.1.18.2        Change Control: Experimental
  7. A.1.18.3        Synopsis: If Cray RUR is configured (see I.2.1.5), log messages will be logged by MoM using logjobmsg when a job ends.
  8. A.1.18.4        Reference to more detail on the interface.
  9. A.1.18.4.1            A message will show the energy used by each aprun run by a job and a job tally in Joules.  For example:
  10. 11/19/2014 18:17:16;0008;pbs_python;Job;267.bigcray;Cray:RUR: {"apid":34876,"apid_energy":83876J,"job_energy":83876J}
  11. 11/19/2014 18:17:16;0008;pbs_python;Job;267.bigcray;Cray:RUR: {"apid":34972,"apid_energy":84272J,"job_energy":168148J}
  12. 11/19/2014 18:17:16;0008;pbs_python;Job;267.bigcray;Cray:RUR: {"apid":35234,"apid_energy":83194J,"job_energy":251342J}
  13. A.1.18.5        Comments on the interface
  14. A.1.18.5.1            Standing of the interface: new interface
  15. A.1.18.5.2            Interface type:  Log message
  16. A.1.19   Interface #20
  17. A.1.19.1        Visibility: Public
  18. A.1.19.2        Change Control: Experimental
  19. A.1.19.3        Synopsis: If Cray RUR is not configured, a log message will be logged by MoM using logjobmsg when a job ends.
  20. A.1.19.4        Reference to more detail on the interface.
  21. A.1.19.4.1            Example:
  22. 11/19/2014 18:17:16;0008;pbs_python;Job;267.bigcray;Cray: no RUR data
  23. A.1.19.5        Comments on the interface
  24. A.1.19.5.1            Standing of the interface: new interface
  25. A.1.19.5.2            Interface type:  Log message
  26. A.1.20   Interface #21
  27. A.1.20.1        Visibility: Public
  28. A.1.20.2        Change Control: Experimental
  29. A.1.20.3        Synopsis: At the end of a job on a Cray, the energy reported by capmc for the compute nodes used by the job will be logged  by MoM using logjobmsg.
  30. A.1.20.4        Reference to more detail on the interface.
  31. A.1.20.4.1            Example:
  32. 11/06/2014 18:35:26;0008;pbs_python;Job;156.bigcray;energy usage 554520J
  33. A.1.20.5        Comments on the interface
  34. A.1.20.5.1            Standing of the interface: new interface
  35. A.1.20.5.2            Interface type:  Log message
  36. A.1.21   Interface #22
  37. A.1.21.1        Visibility: Public
  38. A.1.21.2        Change Control: Experimental
  39. A.1.21.3        Synopsis: The energy reported by capmc for the compute nodes used by a job on a Cray will be logged by MoM using logjobmsg periodically every 5 minutes as the job runs.
  40. A.1.21.4        Reference to more detail on the interface.
  41. A.1.21.4.1            Example:
  42. 11/06/2014 18:35:26;0008;pbs_python;Job;156.bigcray;Cray: get_usage: energy 346342J
  43. A.1.21.5        Comments on the interface
  44. A.1.21.5.1            Standing of the interface: new interface
  45. A.1.21.5.2            Interface type:  Log message
  46. A.1.22   Interface #23
  47. A.1.22.1        Visibility: Private
  48. A.1.22.2        Change Control: Unstable
  49. A.1.22.3        Synopsis: When the PMI on a Cray is initialized, MoM will log messages at LOG_DEBUG.
  50. A.1.22.4        Reference to more detail on the interface.
  51. A.1.22.4.1            Example:
  52. 11/19/2014 15:20:58;0006;pbs_python;Hook;pbs_python;Cray: init
  53. 11/19/2014 15:20:58;0006;pbs_python;Hook;pbs_python;Cray: connect
  54. A.1.22.5        Comments on the interface
  55. A.1.22.5.1            Standing of the interface: new interface
  56. A.1.22.5.2            Interface type:  Log message
  57. A.1.23   Interface #24
  58. A.1.23.1        Visibility: Private
  59. A.1.23.2        Change Control: Unstable
  60. A.1.23.3        Synopsis: When pmi_get_usage is called for a job on a Cray, a message will be logged by MoM using logjobmsg.
  61. A.1.23.4        Reference to more detail on the interface.
  62. A.1.23.4.1            Example:
  63. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: get_usage
  64. A.1.23.5        Comments on the interface
  65. A.1.23.5.1            Standing of the interface: new interface
  66. A.1.23.5.2            Interface type:  Log message
  67. A.1.24   Interface #25
  68. A.1.24.1        Visibility: Private
  69. A.1.24.2        Change Control: Unstable
  70. A.1.24.3        Synopsis: When pmi_query is called on a Cray, a message will be logged by MoM using LOG_DEBUG.
  71. A.1.24.4        Reference to more detail on the interface.
  72. A.1.24.4.1            Example:
  73. 11/19/2014 15:20:58;0006;pbs_python;Hook;pbs_python;Cray: query
  74. A.1.24.5        Comments on the interface
  75. A.1.24.5.1            Standing of the interface: new interface
  76. A.1.24.5.2            Interface type:  Log message
  77. A.1.25   Interface #26
  78. A.1.25.1        Visibility: Private
  79. A.1.25.2        Change Control: Unstable
  80. A.1.25.3        Synopsis: When pmi_activate_profile is called on a Cray, a message will be logged by MoM using LOG_DEBUG.
  81. A.1.25.4        Reference to more detail on the interface.
  82. A.1.25.4.1            Example:
  83. 11/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python;Cray: 167.centos1 activate 'low'
  84. A.1.25.5        Comments on the interface
  85. A.1.25.5.1            Standing of the interface: new interface
  86. A.1.25.5.2            Interface type:  Log message
  87. A.1.26   Interface #27
  88. A.1.26.1        Visibility: Private
  89. A.1.26.2        Change Control: Unstable
  90. A.1.26.3        Synopsis: When pmi_activate_profile is called on a Cray but no compute nodes are allocated to the job, a message will be logged by MoM using logjobmsg.
  91. A.1.26.4        Reference to more detail on the interface.
  92. A.1.26.4.1            Example:
  93. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: no compute nodes for power setting
  94. A.1.26.5        Comments on the interface
  95. A.1.26.5.1            Standing of the interface: new interface
  96. A.1.26.5.2            Interface type:  Log message
  97. A.1.27   Interface #28
  98. A.1.27.1        Visibility: Private
  99. A.1.27.2        Change Control: Unstable
  100. A.1.27.3        Synopsis: When pmi_activate_profile is called on a Cray and the job has pcap_node set, a message will be logged by MoM using logjobmsg showing the pcap_node value.
  101. A.1.27.4        Reference to more detail on the interface.
  102. A.1.27.4.1            Example:
  103. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: pcap node 350
  104. A.1.27.5        Comments on the interface
  105. A.1.27.5.1            Standing of the interface: new interface
  106. A.1.27.5.2            Interface type:  Log message
  107. A.1.28   Interface #29
  108. A.1.28.1        Visibility: Private
  109. A.1.28.2        Change Control: Unstable
  110. A.1.28.3        Synopsis: When pmi_activate_profile is called on a Cray and the job has pcap_accelerator set, a message will be logged by MoM using logjobmsg showing the pcap_ accelerator value.
  111. A.1.28.4        Reference to more detail on the interface.
  112. A.1.28.4.1            Example:
  113. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: pcap accel 250
  114. A.1.28.5        Comments on the interface
  115. A.1.28.5.1            Standing of the interface: new interface
  116. A.1.28.5.2            Interface type:  Log message
  117. A.1.29   Interface #30
  118. A.1.29.1        Visibility: Private
  119. A.1.29.2        Change Control: Unstable
  120. A.1.29.3        Synopsis: When pmi_activate_profile is called on a Cray and the job has neither pcap_node or pcap_accelerator set, a message will be logged by MoM using logjobmsg.
  121. A.1.29.4        Reference to more detail on the interface.
  122. A.1.29.4.1            Example:
  123. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: no power cap to set
  124. A.1.29.5        Comments on the interface
  125. A.1.29.5.1            Standing of the interface: new interface
  126. A.1.29.5.2            Interface type:  Log message
  127. A.1.30   Interface #31
  128. A.1.30.1        Visibility: Private
  129. A.1.30.2        Change Control: Unstable
  130. A.1.30.3        Synopsis: When pmi_deactivate_profile is called on a Cray, a message will be logged by MoM using LOG_DEBUG.
  131. A.1.30.4        Reference to more detail on the interface.
  132. A.1.30.4.1            Example:
  133. 11/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python;Cray: deactivate 167.centos1
  134. A.1.30.5        Comments on the interface
  135. A.1.30.5.1            Standing of the interface: new interface
  136. A.1.30.5.2            Interface type:  Log message
  137. A.1.31   Interface #32
  138. A.1.31.1        Visibility: Private
  139. A.1.31.2        Change Control: Unstable
  140. A.1.31.3        Synopsis: When pmi_deactivate_profile is called on a Cray but no compute nodes are allocated to the job, a message will be logged by MoM using logjobmsg.
  141. A.1.31.4        Reference to more detail on the interface.
  142. A.1.31.4.1            Example:
  143. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: no compute nodes for power setting
  144. A.1.31.5        Comments on the interface
  145. A.1.31.5.1            Standing of the interface: new interface
  146. A.1.31.5.2            Interface type:  Log message
  147. A.1.32   Interface #33
  148. A.1.32.1        Visibility: Private
  149. A.1.32.2        Change Control: Unstable
  150. A.1.32.3        Synopsis: When pmi_deactivate_profile is called on a Cray and the job has pcap_node set, a message will be logged by MoM using logjobmsg showing the pcap_node value.
  151. A.1.32.4        Reference to more detail on the interface.
  152. A.1.32.4.1            Example:
  153. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: remove pcap node 350
  154. A.1.32.5        Comments on the interface
  155. A.1.32.5.1            Standing of the interface: new interface
  156. A.1.32.5.2            Interface type:  Log message
  157. A.1.33   Interface #34
  158. A.1.33.1        Visibility: Private
  159. A.1.33.2        Change Control: Unstable
  160. A.1.33.3        Synopsis: When pmi_deactivate_profile is called on a Cray and the job has pcap_accelerator set, a message will be logged by MoM using logjobmsg showing the pcap_ accelerator value.
  161. A.1.33.4        Reference to more detail on the interface.
  162. A.1.33.4.1            Example:
  163. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: remove pcap accel 250
  164. A.1.33.5        Comments on the interface
  165. A.1.33.5.1            Standing of the interface: new interface
  166. A.1.33.5.2            Interface type:  Log message
  167. A.1.34   Interface #35
  168. A.1.34.1        Visibility: Private
  169. A.1.34.2        Change Control: Unstable
  170. A.1.34.3        Synopsis: When pmi_deactivate_profile is called on a Cray and the job has neither pcap_node or pcap_accelerator set, a message will be logged by MoM using logjobmsg.
  171. A.1.34.4        Reference to more detail on the interface.
  172. A.1.34.4.1            Example:
  173. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: no power cap to remove
  174. A.1.34.5        Comments on the interface
  175. A.1.34.5.1            Standing of the interface: new interface
  176. A.1.34.5.2            Interface type:  Log message
  177. A.1.35   Interface #36
  178. A.1.35.1        Visibility: Private
  179. A.1.35.2        Change Control: Unstable
  180. A.1.35.3        Synopsis:  If Cray RUR is configured but the file created by the output plugin has a permission problem, a message will be logged by MoM using logjobmsg.
  181. A.1.35.4        Reference to more detail on the interface.
  182. A.1.35.4.1            The file owner must be 0 and it must not be writable by other.
  183. A.1.35.4.2            Example:
  184. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: RUR file:  /var/spool/PBS/spool/167.centos1.rur should only be writable by root
  185. A.1.35.5        Comments on the interface
  186. A.1.35.5.1            Standing of the interface: new interface
  187. A.1.35.5.2            Interface type:  Log message
  188. A.1.36   Interface #37
  189. A.1.36.1        Visibility: Private
  190. A.1.36.2        Change Control: Unstable
  191. A.1.36.3        Synopsis:  If Cray RUR is configured but the file created by the output plugin can be read, a message will be logged by MoM using logjobmsg.
  192. A.1.36.4        Reference to more detail on the interface.
  193. A.1.36.4.1            Example
  194. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray: reading RUR file:  /var/spool/PBS/spool/167.centos1.rur
  195. A.1.36.5        Comments on the interface
  196. A.1.36.5.1            Standing of the interface: new interface
  197. A.1.36.5.2            Interface type:  Log message
  198. A.1.37   Interface #38
  199. A.1.37.1        Visibility: Private
  200. A.1.37.2        Change Control: Unstable
  201. A.1.37.3        Synopsis: If the file created by the RUR output plugin can be read but the energy value cannot be parsed, a message will be logged by MoM using logjobmsg.
  202. A.1.37.4        Reference to more detail on the interface.
  203. A.1.37.4.1            A python exception error string will be output as part of the message.
  204. A.1.37.4.2            Example
  205. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray:RUR: energy_used not found: unexpected EOF while parsing
  206. A.1.37.5        Comments on the interface
  207. A.1.37.5.1            Standing of the interface: new interface
  208. A.1.37.5.2            Interface type:  Log message
  209. A.1.38   Interface #39
  210. A.1.38.1        Visibility: Private
  211. A.1.38.2        Change Control: Unstable
  212. A.1.38.3        Synopsis: If the file created by the RUR output plugin can be read but the Cray energy RUR plugin has not been enabled, a message will be logged by MoM using logjobmsg.
  213. A.1.38.4        Reference to more detail on the interface.
  214. A.1.38.4.1            Example
  215. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray:RUR: warning: energy plugin not enabled by RUR
  216. A.1.38.5        Comments on the interface
  217. A.1.38.5.1            Standing of the interface: new interface
  218. A.1.38.5.2            Interface type:  Log message
  219. A.1.39   Interface #40
  220. A.1.39.1        Visibility: Private
  221. A.1.39.2        Change Control: Unstable
  222. A.1.39.3        Synopsis: When the energy for a job is successfully obtained from RUR, MOM will log one of three possible messages using logjobmsg.
  223. A.1.39.4        Reference to more detail on the interface.
  224. A.1.39.4.1            If no energy value was obtained from capmc:
  225. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray:RUR: energy 4.234kWh
  226. A.1.39.4.2            If the energy value from capmc was smaller than what was obtained from RUR:
  227. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray:RUR: energy 4.234kWh replaces capmc energy 4.1432kWh
  228. A.1.39.4.3            If the energy value from capmc was greater than or equal to what was obtained from RUR:
  229. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;Cray:RUR: energy 4.234kWh last capmc usage 4.2432kWh
  230. A.1.39.5        Comments on the interface
  231. A.1.39.5.1            Standing of the interface: new interface
  232. A.1.39.5.2            Interface type:  Log message
  233. A.1.40   Interface #41
  234. A.1.40.1        Visibility: Private
  235. A.1.40.2        Change Control: Unstable
  236. A.1.40.3        Synopsis: When the PMI on an SGI is initialized, MoM will log messages at LOG_DEBUG.
  237. A.1.40.4        Reference to more detail on the interface.
  238. A.1.40.4.1            Example:
  239. 11/19/2014 15:20:58;0006;pbs_python;Hook;pbs_python;SGI: init
  240. 11/19/2014 15:20:58;0006;pbs_python;Hook;pbs_python;SGI: connect
  241. A.1.40.5        Comments on the interface
  242. A.1.40.5.1            Standing of the interface: new interface
  243. A.1.40.5.2            Interface type:  Log message
  244. A.1.41   Interface #42
  245. A.1.41.1        Visibility: Private
  246. A.1.41.2        Change Control: Unstable
  247. A.1.41.3        Synopsis: When pmi_get_usage is called for a job on an SGI, a message will be logged by MoM using logjobmsg.
  248. A.1.41.4        Reference to more detail on the interface.
  249. A.1.41.4.1            Example:
  250. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1;SGI: get_usage
  251. A.1.41.5        Comments on the interface
  252. A.1.41.5.1            Standing of the interface: new interface
  253. A.1.41.5.2            Interface type:  Log message
  254. A.1.42   Interface #43
  255. A.1.42.1        Visibility: Private
  256. A.1.42.2        Change Control: Unstable
  257. A.1.42.3        Synopsis: When pmi_query is called on an SGI, a message will be logged by MoM using LOG_DEBUG.
  258. A.1.42.4        Reference to more detail on the interface.
  259. A.1.42.4.1            Example:
  260. 11/19/2014 15:20:58;0006;pbs_python;Hook;pbs_python;SGI: query
  261. A.1.42.5        Comments on the interface
  262. A.1.42.5.1            Standing of the interface: new interface
  263. A.1.42.5.2            Interface type:  Log message
  264. A.1.43   Interface #44
  265. A.1.43.1        Visibility: Private
  266. A.1.43.2        Change Control: Unstable
  267. A.1.43.3        Synopsis: When pmi_activate_profile is called on an SGI, a message will be logged by MoM using LOG_DEBUG.
  268. A.1.43.4        Reference to more detail on the interface.
  269. A.1.43.4.1            Example:
  270. 11/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python;SGI: 167.centos1 activate '450W'
  271. A.1.43.5        Comments on the interface
  272. A.1.43.5.1            Standing of the interface: new interface
  273. A.1.43.5.2            Interface type:  Log message
  274. A.1.44   Interface #45
  275. A.1.44.1        Visibility: Private
  276. A.1.44.2        Change Control: Unstable
  277. A.1.44.3        Synopsis: When pmi_deactivate_profile is called on an SGI, a message will be logged by MoM using LOG_DEBUG.
  278. A.1.44.4        Reference to more detail on the interface.
  279. A.1.44.4.1            Example:
  280. 11/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python;SGI: deactivate
  281. A.1.44.5        Comments on the interface
  282. A.1.44.5.1            Standing of the interface: new interface
  283. A.1.44.5.2            Interface type:  Log message
  284. A.1.45   Interface #46
  285. A.1.45.1        Visibility: Private
  286. A.1.45.2        Change Control: Unstable
  287. A.1.45.3        Synopsis: If any PMI operation is attempted for a job with a vnode assigned that does not have power_enable=True, a message will be logged by MoM using logjobmsg.
  288. A.1.45.4        Reference to more detail on the interface.
  289. A.1.45.4.1            Example
  290. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1; power functionality is disabled on vnode v12
  291. A.1.45.5        Comments on the interface
  292. A.1.45.5.1            Standing of the interface: new interface
  293. A.1.45.5.2            Interface type:  Log message
  294. A.1.46   Interface #47
  295. A.1.46.1        Visibility: Private
  296. A.1.46.2        Change Control: Unstable
  297. A.1.46.3        Synopsis: If the PMI hook is run with an unexpected event,  MoM will log a message at LOG_WARNING.
  298. A.1.46.4        Reference to more detail on the interface.
  299. A.1.46.4.1            Example
  300. 11/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python; Event not serviceable for power provisioning.
  301. A.1.46.5        Comments on the interface
  302. A.1.46.5.1            Standing of the interface: new interface
  303. A.1.46.5.2            Interface type:  Log message
  304. A.1.47   Interface #48
  305. A.1.47.1        Visibility: Private
  306. A.1.47.2        Change Control: Unstable
  307. A.1.47.3        Synopsis: When the PMI hook handles the EXECHOST_STARTUP event and the MOM is running on the same host as the pbs_server or pbs_sched, MoM will log a message at LOG_DEBUG.
  308. A.1.47.4        Reference to more detail on the interface.
  309. A.1.47.4.1            Example
  310. 11/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python; Provisioning cannot be enabled on this host.
  311. A.1.47.5        Comments on the interface
  312. A.1.47.5.1            Standing of the interface: new interface
  313. A.1.47.5.2            Interface type:  Log message
  314. A.1.48   Interface #49
  315. A.1.48.1        Visibility: Private
  316. A.1.48.2        Change Control: Unstable
  317. A.1.48.3        Synopsis: If any PMI operation at job end throws a python exception, a message will be logged by MoM using logjobmsg showing the exception string.
  318. A.1.48.4        Reference to more detail on the interface.
  319. A.1.48.4.1            Example
  320. 11/19/2014 17:24:21;0008;pbs_python;Job;167.centos1; socket.error: [Errno 111] Connection refused
  321. A.1.48.5        Comments on the interface
  322. A.1.48.5.1            Standing of the interface: new interface
  323. A.1.48.5.2            Interface type:  Log message
  324. A.1.49   Interface #50
  325. A.1.49.1        Visibility: Private
  326. A.1.49.2        Change Control: Unstable
  327. A.1.49.3        Synopsis: If pmi_activate_profile throws either of the python exceptions defined in I.1.1.4.7, a message will be logged by MoM at LOG_WARNING.
  328. A.1.49.4        Reference to more detail on the interface.
  329. A.1.49.4.1            If the exception is BackendError, pmi_query is called to reset the eoe value for the natural vnode for the MoM.
  330. A.1.49.4.2            Example
  331. 1/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python; PMI:activate: set eoe: low,med,high
  332. A.1.49.4.3            If the exception is InternalError, the natural vnode for the MoM will be set offline.
  333. A.1.49.4.4            Example
  334. 1/19/2014 17:24:18;0006;pbs_python;Hook;pbs_python; PMI:activate: set myself offline
  335. A.1.49.5        Comments on the interface
  336. A.1.49.5.1            Standing of the interface: new interface
  337. A.1.49.5.2            Interface type:  Log message
  338. A.2  Administrator’s instructions
  339. A.2.1      Installation does not require any new or different steps.  Once the PBS installation is complete, additional steps will be needed to enable power functionality.
  340. A.2.1.1          Set power_provisioning on the server to true, and power_enable to true on the desired vnodes.  For example:
  341.                                     # qmgr -c “set server power_provisioning = True”
  342.                                     # qmgr -c “set node node1 power_enable = True”
  343.                                     # qmgr -c “set node node2 power_enable = True”
  344.                                     # qmgr -c “set node node3 power_enable = True”
  345. If all vnodes will have power_enable set, @default can be used instead of individual vnode names.  For example:
  346.             # qmgr -c “set node @default power_enable = True”
  347. A.2.1.2          If eoe values are not provided by the PMI, additional steps are needed.
  348. A.2.1.2.1               Use qmgr to set  resources_available.eoe for vnodes.
  349. A.2.1.2.2               Import a submit hook that will map the eoe values to job resources.  See I.2.4.5.
  350. A.2.1.3          To disable power provisioning, set  power_provisioning to false.
  351. A.2.1.3.1               If power_provisioning is set to false while jobs are running, the running jobs would not have their profile deactivated when they finished and the resources_used.energy value would not be set at the end of the job.
  352. A.2.1.3.2               DELETED
  353. A.2.1.4          To disable power provisioning on selected vnodes, set power_enable on the vnodes to False.
  354. A.2.1.5          An optional step on a Cray system can be preformed to use the RUR system to obtain the energy used by each aprun.  Please see the Cray document “Managing System Software for the Cray® Linux Environment” http://docs.cray.com/books/S-2393-5101/S-2393-5101.pdf.  Chapter 12 will provide information on RUR.
  355. A.2.1.5.1               The RUR config file has to be modified to use the PBS output plugin:
  356. /opt/pbs/default/lib/cray/pbs_output.py
  357. A.2.1.5.2                
  358. A.2.1.6          If eoe values are provided by the PMI, additional steps are needed to allow the server and MOMs to communicate the eoe values.
  359. A.2.1.6.1               DELETED
  360. A.2.1.6.2               After setting power_enable on the desired vnodes, restart or HUP the MOMs.
  361. A.2.1.6.3               Check that resources_available.eoe values are set on the vnodes.  It may take up to two minutes for the eoe values to be reported to the server.  If any vnodes do not report eoe values, restart or HUP the MOMs a second time for the vnodes missing eoe values.
  362. A.2.2      An upgrade may require some additional steps.
  363. A.2.2.1           If a job prologue script is defined as described in the PBS  Professional Administrator's Guide section 12.4.4, this must be converted into an execjob_prologue hook before power provisioning can be enabled.  A prologue script will no longer run after power_provisioning is enabled.
  364. A.2.3      If any host is running PBS with an alternate location for the pbs.conf file, PBS_CONF_FILE must be added to the pbs_environment file on that host.  On Linux systems, the default location for the pbs.conf file is /etc/pbs.conf.  The pbs.conf file is used by each MOM to check if the server or scheduler is running on the local host.  If so, the node will not be automatically configured for power provisioning.  For example, if /var/pbs.conf is the active pbs.conf file, the following line must be added to PBS_HOME/pbs_environment:
  365.                                     PBS_CONF_FILE=/var/pbs.conf
  366. A.2.4      New behavior
  367. A.2.4.1          When the  power_provisioning server attribute is set true, the PBS MOM will detect and use the PMI on the system where it is running.  The PMIs supported are:
  368. A.2.4.1.1               SGI Event Driven Framework part of SGI Management software
  369. A.2.4.1.2               CRAY capmc on XC30 hardware platforms with SMW software release 7.0.UP03 and later.
  370. A.2.4.2          The instructions to add or change power profile information are provided by the PMI provider if they are supported.  Here is a list of PMI vendors that support named power profiles.
  371. A.2.4.2.1               SGI
  372. A.2.4.3          If the PMI does not support named power profiles, the resources_available.eoe should be set manually for all the nodes to give a list power profiles.  The eoe values will be mapped to a set of options that are specific to the PMI (see I.2.4.5).  Here is a list of PMI providers that do not support named power profiles.  The options available for each provider follow their name.
  373. A.2.4.3.1               CRAY
  374. A.2.4.3.1.1   pstate: a value for the ALPS reservation p-state.
  375. A.2.4.3.1.2   pgov: a value for the ALPS reservation p-governor value.
  376. A.2.4.3.1.3   pcap_node: a power cap value for each job node in watts.
  377. A.2.4.3.1.4   pcap_accelerator: a power cap value for node accelerator.
  378. A.2.4.4          If the PMI power profile names are obtained from one of the vendors listed in I.2.4.2, then the resources_available.eoe values will be set automatically when power_provisioning is True.
  379. A.2.4.4.1               This will occur when MOM starts.
  380. A.2.4.4.2               DELETED
  381. A.2.4.4.3               A refresh can be forced to happen for a node by restarting or sending a HUP signal to the MOM.
  382. A.2.4.5          If the PMI power profile names are obtained from one of the vendors listed in I.2.4.3, then eoe values must be set manually on the vnodes and a submit hook needs to map the eoe values to the options listed for the PMI vendor.  The hook will set the desired job attributes for each possible eoe value.  For example:

 

  1.                                     # for n in node1 node2 node3 ;do
  2.                         >   qmgr -c “set node $n resources_available.eoe='low,med,high'”
  3.                         > done
  4.                         # cat map_eoe.py
  5.                         import pbs
  6.                         e = pbs.event()
  7.                         j = e.job
  8.                         profile = j.Resource_List['eoe']
  9. if profile is None:
  10.     res = j.Resource_List['select']
  11.     if res is not None:
  12.         for s in str(res).split('+')[0].split(':'):
  13.             if s[:4] == 'eoe=':
  14.                 profile = s.partition('=')[2]
  15.                 break
  16. pbs.logmsg(pbs.LOG_DEBUG, "got profile '%s'" % str(profile))
  17. if profile == "low":
  18.     j.Resource_List["pstate"] = "1900000"
  19.     j.Resource_List["pcap_node"] = 100
  20.     pbs.logmsg(pbs.LOG_DEBUG, "set low")
  21. elif profile == "med":
  22.     j.Resource_List["pstate"] = "220000"
  23.     j.Resource_List["pcap_node"] = 200
  24.     pbs.logmsg(pbs.LOG_DEBUG, "set med")
  25. elif profile == "high":
  26.     j.Resource_List["pstate"] = "240000"
  27.     pbs.logmsg(pbs.LOG_DEBUG, "set high")
  28. else:
  29.     pbs.logmsg(pbs.LOG_DEBUG, "unhandled profile '%s'" % str(profile))
  30.                         e.accept()
  31.                         # qmgr <<EOF
  32.                         create hook power_map event=queuejob
  33.                         import hook power_map application/x-python default map_eoe.py
  34.                         set hook power_map enabled=True
  35.                         EOF

 

  1. A.2.4.5.1               If settings for pstate, pgov, pcap_node, or pcap_accerator are made by the user, then the hook must be written to either overwrite or use the user values as desired.  For example, if the hook above were used as is and the user set a value for eoe to “high”and a value for pcap_node, then the pcap_node value would be in effect which would not normally happen when eoe was set to “high”.
  2. A.2.4.6          When a job is run without a eoe value and power_provisioning is True, no activation is done but the resources_used.energy value for jobs will still be calculated.
  3. A.2.4.7          If both aoe and eoe are set for a vnode, the eoe values must be the same for all the different application operating environments.
  4. A.2.4.8          DELETED
  5. A.2.5      No functionality was deprecated for this RFE.
  6. A.3  User’s instructions
  7. A.3.1      Submit a job which will request a specific power profile.
  8. A.3.1.1          Use the provisioning feature and set “eoe” to a power profile name.  For example:
  9.                                     qsub -leoe=low -lncpus=20 lackadaisical.sh
  10.                                     qsub -lselect=4:eoe=high:ncpus=8 zoomjob
  11. A.3.1.2          Submit a job without a value for “eoe”.  The behavior of the server and scheduler will not be changed for this case.  When this job runs, the power profile of the execution hosts may be changed depending on the implementation of the PMI.  For example, on a Cray (see I.2.4.3.1), a job can have job attributes (see I.1.6, I.1.7, I.1.8 and I.1.9) that affect the execution hosts.
  12. A.3.2      The “resources_used.energy” will be set with a value provided by the PMI.  As with existing behavior, all values for “resources_used” will be written in the accounting log.
  13. A.3.2.1          For example, energy could be included with resources_used for a job 'E' record:
  14.                                     04/14/2014 04:42:03;E;1.x44-mpi.pbspro.com;user=ashisha group=altair project=_pbs_project_default jobname=STDIN queue=workq ctime=1397475718 qtime=1397475718 etime=1397475718 start=1397475718 exec_host=x44-mpi/0 exec_vnode=(x44-mpi:ncpus=1) Resource_List.ncpus=1 Resource_List.nodect=1 Resource_List.place=pack Resource_List.select=1:ncpus=1 session=4746 end=1397475723 Exit_status=255 resources_used.cpupercent=0 resources_used.cput=00:00:01 resources_used.mem=0kb resources_used.ncpus=1 resources_used.vmem=0kb resources_used.walltime=00:00:05 resources_used.energy=1.67 run_count=1

 

  1. A.3.3      Monitor power usage of a job.
  2. A.3.3.1          Use qstat to see the resources_used.energy value as the job runs. 
  3. A.4  Integrator’s Instructions
  4. A.4.1      N/A
  5. A.5  Usage Notes
  6. A.5.1      N/A
  7. A.6  Changes to memory usage as a result of this new feature.
  8. A.6.1      TBD
  9. A.7  Changes to object file size as a result of this new feature.
  10. A.7.1      TBD
  11. A.8  Changes to the performance of any existing commands as a result of this feature.
  12. A.8.1      TBD
  13. A.9  Additional notes to break the administrator's instructions into one section for SGI and one for Cray.
  14. A.9.1      SGI
  15. A.9.1.1          Set power_provisioning on the server to true.  For example:
  16.                                     # qmgr -c “set server power_provisioning = True”
  17. A.9.1.2          DELETED
  18. A.9.1.3          Set power_enable to true on the desired vnodes.  For example:
  19.                                     # qmgr -c “set node node1 power_enable = True”
  20.                                     # qmgr -c “set node node2 power_enable = True”
  21.                                     # qmgr -c “set node node3 power_enable = True”
  22. A.9.1.4          Restart or HUP the MOMs.
  23. A.9.1.5          Check for eoe values as described in I.2.1.6.3.
  24. A.9.2      Cray
  25. A.9.2.1          Set power_provisioning on the server to true and power_enable to true on the desired vnodes.  For example:
  26.                                     # qmgr -c “set server power_provisioning = True”
  27.                                     # qmgr -c “set node node1 power_enable = True”
  28.                                     # qmgr -c “set node node2 power_enable = True”
  29.                                     # qmgr -c “set node node3 power_enable = True”
  30. A.9.2.2          Use qmgr to set  resources_available.eoe for vnodes.
  31. A.9.2.3          Import a submit hook that will map the eoe values to job resources.  See I.2.4.5.
  32. A.9.2.4          Setup RUR if desired.  See I.2.1.5.

 

  • No labels