combiner usage (based off of example)

[Ethan Mallove]

I think in my last post I veered too far off the test/combiner example :)

$ cat count2.xml
<?xml version="1.0" standalone="no"?>
<!DOCTYPE query SYSTEM "../../dtd/pb_query.dtd">

<query id="min_max">
  <experiment>test_run_correctness</experiment>
  <description>
    Combiner test
  </description>

  <parameter id="p.msgsize">
    <value>test_name</value>
  </parameter>

  <source id="src.latency">
    <result>test_name</result>
    <input>p.msgsize</input>
  </source>

  <operator id="op.min" type="count">
    <input>src.latency</input>
  </operator>

  <combiner id="comb.minmaxavg">
    <input>op.min</input>
  </combiner>

  <output>
    <input>comb.minmaxavg</input>
  </output>

</query>
$ cat count2.xml | perfbase query --desc=-
#* ERROR: can not process query specification
    "<value> 'test_name' is not a parameter in the specified experiment"
$ grep -C 2 test_name www/inp_test_run_correctness.xml

  <named_location mode="set" len="0">
    <name>test_name</name>
    <match>test_name: </match>
  </named_location>


I dont understand why perfbase throws the error, because in test/combiner I see
the following in the query*.xml:

  <parameter id="p.msgsize">
    <value>S_msg</value>
  </parameter>

And this in the inp*.xml:

    <tabular_value>
      <name>S_msg</name>
      <pos>0</pos>
    </tabular_value>

I thought it would make sense to swap "S_msg" with "test_name" would work.

-Ethan

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Ethan Mallove]

Lest I didn't provide enough context for my last question, I've attached the
experiment xml file.



Ethan Mallove wrote On 07/25/06 16:04,:

> I think in my last post I veered too far off the test/combiner example :)
>
> $ cat count2.xml
> <?xml version="1.0" standalone="no"?>
> <!DOCTYPE query SYSTEM "../../dtd/pb_query.dtd">
>
> <query id="min_max">
>   <experiment>test_run_correctness</experiment>
>   <description>
>     Combiner test
>   </description>
>
>   <parameter id="p.msgsize">
>     <value>test_name</value>
>   </parameter>
>
>   <source id="src.latency">
>     <result>test_name</result>
>     <input>p.msgsize</input>
>   </source>
>
>   <operator id="op.min" type="count">
>     <input>src.latency</input>
>   </operator>
>
>   <combiner id="comb.minmaxavg">
>     <input>op.min</input>
>   </combiner>
>
>   <output>
>     <input>comb.minmaxavg</input>
>   </output>
>
> </query>
> $ cat count2.xml | perfbase query --desc=-
> #* ERROR: can not process query specification
>     "<value> 'test_name' is not a parameter in the specified experiment"
> $ grep -C 2 test_name www/inp_test_run_correctness.xml
>
>   <named_location mode="set" len="0">
>     <name>test_name</name>
>     <match>test_name: </match>
>   </named_location>
>
>
> I dont understand why perfbase throws the error, because in test/combiner I see
> the following in the query*.xml:
>
>   <parameter id="p.msgsize">
>     <value>S_msg</value>
>   </parameter>
>
> And this in the inp*.xml:
>
>     <tabular_value>
>       <name>S_msg</name>
>       <pos>0</pos>
>     </tabular_value>
>
> I thought it would make sense to swap "S_msg" with "test_name" would work.
>
> -Ethan
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@...
> For additional commands, e-mail: users-help@...
>

<?xml version="1.0"?>
<!DOCTYPE experiment SYSTEM "../../dtd/pb_experiment.dtd">

[snip]

<experiment>
  <name>test_run_correctness</name>

[snip]

  <!-- SINGLE-OCCURENCE FIELDS -->

  <result occurence="once">
    <name>mtt_version_major</name>
    <synopsis>MTT major version</synopsis>
    <datatype>integer</datatype>
  </result>

  <result occurence="once">
    <name>mtt_version_minor</name>
    <synopsis>MTT minor version</synopsis>
    <datatype>integer</datatype>
  </result>

  <result occurence="once">
    <name>hostname</name>
    <synopsis>host name</synopsis>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>os_name</name>
    <synopsis>operating system name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>os_version</name>
    <synopsis>operating system version</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>platform_hardware</name>
    <synopsis>platform architecture</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>platform_type</name>
    <synopsis>platform os/architecture</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>platform_id</name>
    <synopsis>platform identifier</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>mpi_name</name>
    <synopsis>MPI name</synopsis>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>mpi_version</name>
    <synopsis>MPI version</synopsis>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>mpi_get_section_name</name>
    <synopsis>MPI get section name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>mpi_install_section_name</name>
    <synopsis>MPI install section name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>test_build_section_name</name>
    <synopsis>test build section name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>test_run_section_name</name>
    <synopsis>test run section name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="once">
    <name>merge_stdout_stderr</name>
    <synopsis>true if stdout and stderr are merged</synopsis>
    <default>0</default>
    <datatype>integer</datatype>
  </result>
 
  <result occurence="once">
    <name>environment</name>
    <synopsis>environment the run was performed in</synopsis>
    <default>N/A</default>
    <datatype>text</datatype>
  </result>
 
  <parameter occurence="once">
    <name>submit_timestamp</name>
    <synopsis>date and time the run results were submitted</synopsis>
    <datatype>timestamp</datatype>
  </parameter>

  <!-- MULTIPLE-OCCURENCE FIELDS -->
 
  <result occurence="multiple">
    <name>test_command</name>
    <synopsis>full command line for this test</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <result occurence="multiple">
    <name>test_name</name>
    <synopsis>name of test being run</synopsis>
    <datatype>string</datatype>
  </result>

  <result occurence="multiple">
    <name>test_np</name>
    <synopsis>number of processes used for test</synopsis>
    <datatype>integer</datatype>
  </result>

  <result occurence="multiple">
    <name>test_message</name>
    <synopsis>test result message</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </result>

  <parameter occurence="multiple">
    <name>test_pass</name>
    <synopsis>whether or not the test passed</synopsis>
    <datatype>integer</datatype>
  </parameter>

  <result occurence="multiple">
    <name>stderr</name>
    <synopsis>stderr output from run</synopsis>
    <default>N/A</default>
    <datatype>text</datatype>
  </result>

  <result occurence="multiple">
    <name>stdout</name>
    <synopsis>stdout output from run</synopsis>
    <default>N/A</default>
    <datatype>text</datatype>
  </result>
 
  <result occurence="multiple">
    <name>start_timestamp</name>
    <synopsis>date and time this test was started</synopsis>
    <datatype>timestamp</datatype>
  </result>

  <result occurence="multiple">
    <name>stop_timestamp</name>
    <synopsis>date and time this test was finished</synopsis>
    <datatype>timestamp</datatype>
  </result>

</experiment>



---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Joachim Worringen-4]

Ethan Mallove wrote:

> Lest I didn't provide enough context for my last question, I've attached the
> experiment xml file.
>
> Ethan Mallove wrote On 07/25/06 16:04,:
>> I think in my last post I veered too far off the test/combiner example :)
>>
>> $ cat count2.xml
>> <?xml version="1.0" standalone="no"?>
>> <!DOCTYPE query SYSTEM "../../dtd/pb_query.dtd">
>>
>> <query id="min_max">
>>   <experiment>test_run_correctness</experiment>
>>   <description>
>>     Combiner test
>>   </description>
>>
>>   <parameter id="p.msgsize">
>>     <value>test_name</value>
>>   </parameter>

To avoid confusion when queries become more complex, I recommend that
the given id's reflect the actual function of an object. I.e.,
"p.msgsize" has nothing to do with the value "test_name".

[...]

>> $ cat count2.xml | perfbase query --desc=-
>> #* ERROR: can not process query specification
>>     "<value> 'test_name' is not a parameter in the specified experiment"
 >>

>> $ grep -C 2 test_name www/inp_test_run_correctness.xml
>>
>>   <named_location mode="set" len="0">
>>     <name>test_name</name>
>>     <match>test_name: </match>
>>   </named_location>
>>
>>
>> I dont understand why perfbase throws the error, because in test/combiner I see
>> the following in the query*.xml:
>>
>>   <parameter id="p.msgsize">
>>     <value>S_msg</value>
>>   </parameter>
>>
>> And this in the inp*.xml:
>>
>>     <tabular_value>
>>       <name>S_msg</name>
>>       <pos>0</pos>
>>     </tabular_value>

You need to look what's *really* in the database for the experiment that
the query refers to - *that's* what perfbase refers to:
'perfbase info -e experiment_name --val='
will list all defined values in the experiment.

>> I thought it would make sense to swap "S_msg" with "test_name" would work.

In theory, yes, but you need to make sure that the database
representation of the experiment matches. Delete the old experiment and
create a new one.

>> <?xml version="1.0"?>
>> <!DOCTYPE experiment SYSTEM "../../dtd/pb_experiment.dtd">
>>
>> [snip]
>>
>> <experiment>
>>   <name>test_run_correctness</name>
>>
>> [snip]
>>
>>   <!-- SINGLE-OCCURENCE FIELDS -->
>>
>>   <result occurence="once">
>>     <name>mtt_version_major</name>
>>     <synopsis>MTT major version</synopsis>
>>     <datatype>integer</datatype>
>>   </result>

[...]

I'd say that things like version number etc. (many of the values you
declared as results) would actually better be declared as parameters.
They are input to the experiment run, not output, aren't they? And you
want to query for them, don't you?

  Joachim

--
Joachim Worringen, Software Architect, Dolphin Interconnect Solutions
phone ++49/(0)228/324 08 17 - http://www.dolphinics.com

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Ethan Mallove]

Joachim Worringen wrote On 07/26/06 08:54,:

> Ethan Mallove wrote:
>
>>Lest I didn't provide enough context for my last question, I've attached the
>>experiment xml file.
>>
>>Ethan Mallove wrote On 07/25/06 16:04,:
>>
>>>I think in my last post I veered too far off the test/combiner example :)
>>>
>>>$ cat count2.xml
>>><?xml version="1.0" standalone="no"?>
>>><!DOCTYPE query SYSTEM "../../dtd/pb_query.dtd">
>>>
>>><query id="min_max">
>>>  <experiment>test_run_correctness</experiment>
>>>  <description>
>>>    Combiner test
>>>  </description>
>>>
>>>  <parameter id="p.msgsize">
>>>    <value>test_name</value>
>>>  </parameter>
>
>
> To avoid confusion when queries become more complex, I recommend that
> the given id's reflect the actual function of an object. I.e.,
> "p.msgsize" has nothing to do with the value "test_name".
>
> [...]
>

I realize that the id's are non-sensisical, I wanted to veer as little as
possible from your example. Sorry it came out confusing :)

>
>>>$ cat count2.xml | perfbase query --desc=-
>>>#* ERROR: can not process query specification
>>>    "<value> 'test_name' is not a parameter in the specified experiment"
>
>  >>
>
>>>$ grep -C 2 test_name www/inp_test_run_correctness.xml
>>>
>>>  <named_location mode="set" len="0">
>>>    <name>test_name</name>
>>>    <match>test_name: </match>
>>>  </named_location>
>>>
>>>
>>>I dont understand why perfbase throws the error, because in test/combiner I see
>>>the following in the query*.xml:
>>>
>>>  <parameter id="p.msgsize">
>>>    <value>S_msg</value>
>>>  </parameter>
>>>
>>>And this in the inp*.xml:
>>>
>>>    <tabular_value>
>>>      <name>S_msg</name>
>>>      <pos>0</pos>
>>>    </tabular_value>
>
>
> You need to look what's *really* in the database for the experiment that
> the query refers to - *that's* what perfbase refers to:
> 'perfbase info -e experiment_name --val='
> will list all defined values in the experiment.
>

$ perfbase info -e test_run_correctness --val= | grep -w test_name
 test_name            :  string       []                RM


So the value is in the database, right?

>
>>>I thought it would make sense to swap "S_msg" with "test_name" would work.
>
>
> In theory, yes, but you need to make sure that the database
> representation of the experiment matches. Delete the old experiment and
> create a new one.
>
>
>>><?xml version="1.0"?>
>>><!DOCTYPE experiment SYSTEM "../../dtd/pb_experiment.dtd">
>>>
>>>[snip]
>>>
>>><experiment>
>>>  <name>test_run_correctness</name>
>>>
>>>[snip]
>>>
>>>  <!-- SINGLE-OCCURENCE FIELDS -->
>>>
>>>  <result occurence="once">
>>>    <name>mtt_version_major</name>
>>>    <synopsis>MTT major version</synopsis>
>>>    <datatype>integer</datatype>
>>>  </result>
>
> [...]
>
> I'd say that things like version number etc. (many of the values you
> declared as results) would actually better be declared as parameters.
> They are input to the experiment run, not output, aren't they? And you
> want to query for them, don't you?
>

We're not invalidating any versions, so we can use <result>, right?  But I'm
curious - will calling them <result> instead of <parameter> differ in how
perfbase inserts/queries this data?

-Ethan


>   Joachim
>

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Joachim Worringen-4]

Ethan Mallove wrote:
> $ perfbase info -e test_run_correctness --val= | grep -w test_name
>  test_name            :  string       []                RM
>
>
> So the value is in the database, right?

Yes - but it is a result ('R' means result, 'M' stands for multiple -.
or use '--val=test_name -v').

You can not filter for result values, but only for parameter values
because you want to see which results are delivered for certain
parameters. You can do filtering on the results lateron using <operators>.

The error message could be even more precise here, but perfbase does not
check if a value is a result if it is not a parameter.

> We're not invalidating any versions, so we can use <result>, right?  But I'm
> curious - will calling them <result> instead of <parameter> differ in how
> perfbase inserts/queries this data?

No, not for the inserting, but for the query. See above...

  Joachim

--
Joachim Worringen, Software Architect, Dolphin Interconnect Solutions
phone ++49/(0)228/324 08 17 - http://www.dolphinics.com

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Ethan Mallove]

Joachim Worringen wrote On 07/26/06 11:01,:

> Ethan Mallove wrote:
>
>>$ perfbase info -e test_run_correctness --val= | grep -w test_name
>> test_name            :  string       []                RM
>>
>>
>>So the value is in the database, right?
>
>
> Yes - but it is a result ('R' means result, 'M' stands for multiple -.
> or use '--val=test_name -v').
>
> You can not filter for result values, but only for parameter values
> because you want to see which results are delivered for certain
> parameters. You can do filtering on the results lateron using <operators>.
>
> The error message could be even more precise here, but perfbase does not
> check if a value is a result if it is not a parameter.
>

I changed all <result> to <parameter> across the board because I want to be able
to filter on anything, including results. E.g., test_result seems like it would
definitely be a <result>, but this prevents me from doing something like: "Show
me all runs of test 'foo' with test_result equal to 'fail' for all OS versions".


$ cat count2.xml | perfbase query --desc=- --sqltrace
#* SQL: SELECT * FROM exp_metadata WHERE name='test_run_correctness'
#* SQL: SELECT * FROM exp_values WHERE name = %s AND NOT is_result ('test_name',)
#* SQL: SELECT * FROM exp_values WHERE name = %s ('test_name',)
#* SQL: SELECT only_once FROM exp_values WHERE name='test_name'
#* ERROR: could not set up target table
   "duplicate entry of value 'test_name'"


What does this error mean?

In an attempt to get rid of the 'duplicate' error - I change the <value> to
_test_name:

#* ERROR: can not process query specification
    "<value> '_test_name' is not a parameter in the specified experiment"

I change <result> to _test_name and get:

#* ERROR: can not process query specification
    "<source> 'src': <result> _test_name not found in experiment"

??


$ cat count2.xml
<?xml version="1.0" standalone="no"?>
<!DOCTYPE query SYSTEM "../../dtd/pb_query.dtd">

<query id="min_max">
  <experiment>test_run_correctness</experiment>
  <description>
    Combiner test
  </description>

  <parameter id="name">
    <value>test_name</value>
  </parameter>

  <source id="src">
    <result>test_name</result>
    <input>name</input>
  </source>

  <operator id="_count" type="count">
    <input>src</input>
  </operator>

  <combiner id="name.count">
    <input>_count</input>
  </combiner>

  <output>
    <input>name.count</input>
  </output>

</query>
$ cat www/inp_test_run_correctness.xml
<?xml version="1.0"?>
<!DOCTYPE input SYSTEM "../../dtd/pb_input.dtd">
<!--
  Copyright (c) 2005-2006 The Trustees of Indiana University.
                          All rights reserved.
  Copyright (c) 2006      Cisco Systems, Inc.  All rights reserved.
  $COPYRIGHT$

  Additional copyrights may follow

  $HEADER$

-->

<input>
  <experiment>test_run_correctness</experiment>

  <set_separation>

<match>============================================================================</match>
  </set_separation>

  <named_location len="0">
    <name>mtt_version_major</name>
    <match>mtt_version_major: </match>
  </named_location>

  <named_location len="0">
    <name>mtt_version_minor</name>
    <match>mtt_version_minor: </match>
  </named_location>

  <named_location len="0">
    <name>hostname</name>
    <match>hostname: </match>
  </named_location>

  <named_location len="0">
    <name>os_name</name>
    <match>os_name: </match>
  </named_location>

  <named_location len="0">
    <name>os_version</name>
    <match>os_version: </match>
  </named_location>

  <named_location len="0">
    <name>platform_hardware</name>
    <match>platform_hardware: </match>
  </named_location>

  <named_location len="0">
    <name>platform_type</name>
    <match>platform_type: </match>
  </named_location>

  <named_location len="0">
    <name>platform_id</name>
    <match>platform_id: </match>
  </named_location>

  <named_location len="0">
    <name>environment</name>
    <match marker="begin" skip="1">ENVIRONMENT_BEGIN</match>
    <match marker="end">ENVIRONMENT_END</match>
  </named_location>

  <named_location len="0">
    <name>mpi_name</name>
    <match>mpi_name: </match>
  </named_location>

  <named_location len="0">
    <name>mpi_version</name>
    <match>mpi_version: </match>
  </named_location>

  <named_location len="0">
    <name>mpi_get_section_name</name>
    <match>mpi_get_section_name: </match>
  </named_location>

  <named_location len="0">
    <name>mpi_install_section_name</name>
    <match>mpi_install_section_name: </match>
  </named_location>

  <!-- MULTIPLE-occurrence FIELDS -->

  <named_location mode="set" len="0">
    <name>test_build_section_name</name>
    <match>test_build_section_name: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>test_run_section_name</name>
    <match>test_run_section_name: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>merge_stdout_stderr</name>
    <match>merge_stdout_stderr: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>test_command</name>
    <match>test_command: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>test_message</name>
    <match>test_message: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>test_name</name>
    <match>test_name: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>test_np</name>
    <match>test_np: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>test_pass</name>
    <match>test_pass: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>stderr</name>
    <match marker="begin" skip="1">STDERR_BEGIN</match>
    <match marker="end">STDERR_END</match>
  </named_location>

  <named_location mode="set" len="0">
    <name>stdout</name>
    <match marker="begin" skip="1">STDOUT_BEGIN</match>
    <match marker="end">STDOUT_END</match>
  </named_location>

  <named_location mode="set" len="0">
    <name>submit_timestamp</name>
    <match>submit_timestamp: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>start_timestamp</name>
    <match>start_timestamp: </match>
  </named_location>

  <named_location mode="set" len="0">
    <name>stop_timestamp</name>
    <match>stop_timestamp: </match>
  </named_location>

</input>
$ cat www/exp_test_run_correctness.xml
<?xml version="1.0"?>
<!DOCTYPE experiment SYSTEM "../../dtd/pb_experiment.dtd">
<!--
  Copyright (c) 2005-2006 The Trustees of Indiana University.
                          All rights reserved.
  Copyright (c) 2006      Cisco Systems, Inc.  All rights reserved.
  $COPYRIGHT$

  Additional copyrights may follow

  $HEADER$

-->

<!--
  One run in the test_run experiment corresponds to a run of an entire
  test suite.  A number of fields do not vary from test to test within
  a test suite, so occur only once.  Other fields have specific values
  for each test.  These fields occur multiple times.

  Generic experiment for correctness tests (i.e., "pass" / "fail"
  tests).
-->

<experiment>
  <name>test_run_correctness</name>

  <info>
    <performed_by>
      <name>MTT Engine</name>
      <organization>The Open MPI Project</organization>
    </performed_by>
    <project>MPI Testing Tool</project>
    <synopsis>Results of MTT MPI test suite runs: generic correctness
tests</synopsis>
    <description>Generic experiment for correctness tests (i.e., "pass" / "fail"
tests)</description>
  </info>


  <!-- SINGLE-occurrence FIELDS -->

  <parameter occurrence="once">
    <name>mtt_version_major</name>
    <synopsis>MTT major version</synopsis>
    <datatype>integer</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>mtt_version_minor</name>
    <synopsis>MTT minor version</synopsis>
    <datatype>integer</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>hostname</name>
    <synopsis>host name</synopsis>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>os_name</name>
    <synopsis>operating system name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>os_version</name>
    <synopsis>operating system version</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>platform_hardware</name>
    <synopsis>platform architecture</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>platform_type</name>
    <synopsis>platform os/architecture</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>platform_id</name>
    <synopsis>platform identifier</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>mpi_name</name>
    <synopsis>MPI name</synopsis>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>mpi_version</name>
    <synopsis>MPI version</synopsis>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>mpi_get_section_name</name>
    <synopsis>MPI get section name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>mpi_install_section_name</name>
    <synopsis>MPI install section name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>test_build_section_name</name>
    <synopsis>test build section name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>test_run_section_name</name>
    <synopsis>test run section name</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>merge_stdout_stderr</name>
    <synopsis>true if stdout and stderr are merged</synopsis>
    <default>0</default>
    <datatype>integer</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>environment</name>
    <synopsis>environment the run was performed in</synopsis>
    <default>N/A</default>
    <datatype>text</datatype>
  </parameter>

  <parameter occurrence="once">
    <name>submit_timestamp</name>
    <synopsis>date and time the run results were submitted</synopsis>
    <datatype>timestamp</datatype>
  </parameter>

  <!-- MULTIPLE-occurrence FIELDS -->

  <parameter>
    <name>test_command</name>
    <synopsis>full command line for this test</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter>
    <name>test_name</name>
    <synopsis>name of test being run</synopsis>
    <datatype>string</datatype>
  </parameter>

  <parameter>
    <name>test_np</name>
    <synopsis>number of processes used for test</synopsis>
    <datatype>integer</datatype>
  </parameter>

  <parameter>
    <name>test_message</name>
    <synopsis>test parameter message</synopsis>
    <default>N/A</default>
    <datatype>string</datatype>
  </parameter>

  <parameter>
    <name>test_pass</name>
    <synopsis>whether or not the test passed</synopsis>
    <datatype>integer</datatype>
  </parameter>

  <parameter>
    <name>stderr</name>
    <synopsis>stderr output from run</synopsis>
    <default>N/A</default>
    <datatype>text</datatype>
  </parameter>

  <parameter>
    <name>stdout</name>
    <synopsis>stdout output from run</synopsis>
    <default>N/A</default>
    <datatype>text</datatype>
  </parameter>

  <parameter>
    <name>start_timestamp</name>
    <synopsis>date and time this test was started</synopsis>
    <datatype>timestamp</datatype>
  </parameter>

  <parameter>
    <name>stop_timestamp</name>
    <synopsis>date and time this test was finished</synopsis>
    <datatype>timestamp</datatype>
  </parameter>

</experiment>

>
>>We're not invalidating any versions, so we can use <result>, right?  But I'm
>>curious - will calling them <result> instead of <parameter> differ in how
>>perfbase inserts/queries this data?
>
>
> No, not for the inserting, but for the query. See above...
>
>   Joachim
>

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Ethan Mallove]

I know I am not supposed to think of SQL when writing perfbase XML query files,
but could you tell me what the XML query would look like for this SQL query?

$ echo "SELECT test_name FROM rundata GROUP BY test_name;" | psql -d
pb_test_run_correctness

-Ethan

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Joachim Worringen-4]

Ethan Mallove wrote:
> I know I am not supposed to think of SQL when writing perfbase XML query files,
> but could you tell me what the XML query would look like for this SQL query?
>
> $ echo "SELECT test_name FROM rundata GROUP BY test_name;" | psql -d
> pb_test_run_correctness

You mean you want to see all distinct content of test_name?

This is done via "perfbase ls -d -s test_name".

A query does not support this alone, only in the context of a real query
(sic!). Use a blind sweep for this:
<parameter id="bla">
   <value>test_name</value>
   <sweep>
   </sweep>
</parameter>

  Joachim

--
Joachim Worringen, Software Architect, Dolphin Interconnect Solutions
phone ++49/(0)228/324 08 17 - http://www.dolphinics.com

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Joachim Worringen-4]

Ethan Mallove wrote:

> $ cat count2.xml | perfbase query --desc=- --sqltrace
> #* SQL: SELECT * FROM exp_metadata WHERE name='test_run_correctness'
> #* SQL: SELECT * FROM exp_values WHERE name = %s AND NOT is_result ('test_name',)
> #* SQL: SELECT * FROM exp_values WHERE name = %s ('test_name',)
> #* SQL: SELECT only_once FROM exp_values WHERE name='test_name'
> #* ERROR: could not set up target table
>    "duplicate entry of value 'test_name'"
>
>
> What does this error mean?

See below.

> In an attempt to get rid of the 'duplicate' error - I change the <value> to
> _test_name:
>
> #* ERROR: can not process query specification
>     "<value> '_test_name' is not a parameter in the specified experiment"
>
> I change <result> to _test_name and get:
>
> #* ERROR: can not process query specification
>     "<source> 'src': <result> _test_name not found in experiment"
>
> ??

Well, obviously, _test_name is not a value in the experiement?!

> $ cat count2.xml
> <?xml version="1.0" standalone="no"?>
> <!DOCTYPE query SYSTEM "../../dtd/pb_query.dtd">
>
> <query id="min_max">
>   <experiment>test_run_correctness</experiment>
>   <description>
>     Combiner test
>   </description>
>
>   <parameter id="name">
>     <value>test_name</value>
>   </parameter>
>
>   <source id="src">
>     <result>test_name</result>
>     <input>name</input>
>   </source>

You can not have 'test_name' being a parameter and result at the same time.

  Joachim

--
Joachim Worringen, Software Architect, Dolphin Interconnect Solutions
phone ++49/(0)228/324 08 17 - http://www.dolphinics.com

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...