summaryrefslogtreecommitdiffstats
path: root/rteval
Commit message (Collapse)AuthorAgeFilesLines
* Fix mismatch between rteval.py and specfileClark Williams2009-11-251-1/+1
| | | | Signed-off-by: Clark Williams <williams@redhat.com>
* fix incorrect manipulation of cyclictest thread prioritiesClark Williams2009-11-251-1/+1
| | | | | | | | | | | When running multiple measurement realtime threads (-t -p<N>), cyclictest normally starts one thread with the specified priority N, then decrements that priority for each subsequent thread. But, using the histogram option (which rteval does) changes this logic so that all measurement threads are equal priority. Change this in cyclictest.py. Signed-off-by: Clark Williams <williams@redhat.com>
* version bump to 1.11Clark Williams2009-11-061-1/+1
|
* capture base OS infoClark Williams2009-11-062-0/+20
| | | | | | | | | Added routine to capture base OS (RHEL vs. Fedora) info and put into the XML summary; also modified xsl report file to show the base OS info in report Signed-off-by: Clark Williams <williams@redhat.com> CC: David Sommerseth <davids@redhat.com>
* version bump to 1.10Clark Williams2009-11-041-1/+1
|
* add config file sections for cyclictestClark Williams2009-11-043-3/+16
| | | | | | | | | | Added a config file section for setting parameters used by cyclictest. Handle two parameters: buckets: <n> - the number 1us histogram buckets to allocate [2000] interval: <n> - the polling interval in microseconds [100] Signed-off-by: Clark Williams <williams@redhat.com> CC: David Sommerseth <davids@redhat.com>
* version bump to 1.9Clark Williams2009-11-031-1/+1
|
* New feature: rteval <-H | --raw-histogram>David Sommerseth2009-10-292-5/+63
| | | | | A variant of -Z|--summarize, but it uses another XSLT template to generate a raw text formated version of the histogram data
* Cleaned up the rteval_text.xslDavid Sommerseth2009-10-291-48/+148
| | | | | Using <xsl:text/> tags extensively to better control the text flow, which also increases the readablity of the stylesheet.
* Sort CPU cores as numbers, not as string (which is the default)David Sommerseth2009-10-291-1/+1
|
* Merge commit 'clark/master' into local clark work branchDavid Sommerseth2009-10-261-3/+11
|\ | | | | | | | | | | Conflicts: Makefile rteval.spec
| * fixed command line issue with sysreport optionClark Williams2009-10-201-2/+4
| | | | | | | | | | | | | | Fixed incorrect option being passed to sysreport option and added debugging messages around generation and moving of sysreport Signed-off-by: Clark Williams <williams@redhat.com>
| * fixed specfile issues and added better failure message on NFSClark Williams2009-10-201-1/+7
| | | | | | | | | | | | | | | | | | | | Fixed bogus %changelog date and removed leftover load tarball in rteval %files section. Caught permissions failure when running on NFS filesystem with root squash. Signed-off-by: Clark Williams <williams@redhat.com>
* | Merge branch 'master_ipv4' into clarkDavid Sommerseth2009-10-261-24/+46
|\ \ | |/ |/| | | | | | | Conflicts: Makefile rteval.spec
| * A few of fixesDavid Sommerseth2009-10-231-22/+44
| | | | | | | | | | | | | | | | | | | | | | | | | | - rteval do not collect system info before it's really needed This avoids system data collection when running rteval with --summarize. - Fixed get_kthreads() to ignore lines which do not have number in the first column. On some boxes, rtctl prints some warnings before it prints the thread list. - Made --summarize work more independently of the self.xmlreport variable.
| * Cleaned up the service XML tagsDavid Sommerseth2009-10-231-1/+1
| |
| * version bump to 1.7Clark Williams2009-10-231-1/+1
| | | | | | | | | | Added <kthread> section to XML summary and merged David's fixes to option processing and additions to XML.
| * added <kthread> section to XML reportClark Williams2009-10-231-7/+29
| | | | | | | | | | added code to grab kthread policy/priority via 'rtctl' and store that in the xml summary report of a run
| * version bump to 1.6Clark Williams2009-10-231-1/+1
| | | | | | | | | | changed stats calculation to loop less added methods for retrieving services status and kthread status
| * reworked stats calculation in cyclictest::reduce()Clark Williams2009-10-231-8/+2
| | | | | | | | avoid looping so much when calculating MAD and variance
| * added get_services() and get_kthread() methods to rteval/rteval.pyClark Williams2009-10-231-0/+26
| | | | | | | | Add methods to retrieve status of services and priorities of kthreads
| * Added hardware product name and manufacturer to the screen reportDavid Sommerseth2009-10-231-0/+1
| |
| * Fixed wrong data type for self.config.duration parsingDavid Sommerseth2009-10-231-1/+1
| |
| * rteval did not consider -f|--inifile argumentDavid Sommerseth2009-10-231-3/+3
| |
| * version bump to 1.5Clark Williams2009-10-231-1/+1
| | | | | | | | | | updated option processing, changed stats to use less memory, changed debug output to prefix with module name
| * modified Rundata::reduce in rteval/cyclictest.py to use less memoryClark Williams2009-10-231-5/+10
| | | | | | | | | | | | | | | | | | Rundata::reduce was using a range() statement, which generated a potentially HUGE list on long runs, many times exhausting memory. Avoid the range statment and just manually iterate Also update the cyclictest::debug() method to prefix all output with cyclictest:
| * modified rteval::debug() to prefix strings with rteval:Clark Williams2009-10-231-1/+1
| | | | | | | | | | updated the rteval::debug method to prefix output with reval: to differentiate output from other objects.
| * modified load::debug to print object nameClark Williams2009-10-233-16/+16
| | | | | | | | | | | | updated the load::debug method to print the object name as well as the input string, then updated call sites for this method in rteval/hackbench.py and rteval/kcompile.py
| * rearranged option processing logic in rteval/rteval.pyClark Williams2009-10-231-9/+18
| | | | | | | | | | | | rearranged order of command line and config file processing so that command line is processed, then config file and finally command line parameters override the config file values.
| * Changed from sample array to histogram (-h 1000)Clark Williams2009-10-231-22/+29
| | | | | | | | | | | | | | | | | | Calculate statistics directly from histogram (removed the temporary array which was eating *tons* of memory in Cyclictest::reduce()) Changed sampling interval to 100us (-i100 option to cyclictest) Added -a option to force affinity for measurement threads
| * updated version to 1.4Clark Williams2009-10-231-1/+2
| |
| * rteval/rteval_text.xsl changes:Clark Williams2009-10-231-1/+3
| | | | | | | | | | added units to mode parameter added mean absolute deviation
| * adde code to rteval/rtevalConfig.py to allow first call toClark Williams2009-10-231-1/+6
| | | | | | | | __find_config to fail (and allow command line to specify new value)
| * rteval.py changes:Clark Williams2009-10-231-1/+6
| | | | | | | | | | | | added code to change xslt_report path when -i specified added synchronization code with cyclictest to avoid trying to generate XML before cyclictest has finished stats calcuations
| * Massive changes:Clark Williams2009-10-231-98/+65
| | | | | | | | | | | | | | | | - keep data as histogram instead of sample array - change stats calculation to deal with histogram - add Mean Absolute Deviation calculation - print variance - add synchronization with rteval to prevent race
| * Add XML-RPC submissions into a submission queue instead of parsing it directlyDavid Sommerseth2009-09-301-1/+1
| | | | | | | | | | | | This is to avoid overloading the Apache process doing XML parsing and database inserts on bigger XML reports. A separate parser process will need to pick files from the submission queue and parse them in separate thread(s).
* | split kcompile and hackbench into sub-packagesClark Williams2009-10-141-1/+1
| | | | | | | | | | | | Modified specfile to generate kcompile and hackbench sub-packages, so that updates to rteval won't require re-downloading a kernel tarball
* | version bump to 1.7Clark Williams2009-10-131-1/+1
| | | | | | | | | | Added <kthread> section to XML summary and merged David's fixes to option processing and additions to XML.
* | added <kthread> section to XML reportClark Williams2009-10-131-7/+29
| | | | | | | | | | added code to grab kthread policy/priority via 'rtctl' and store that in the xml summary report of a run
* | Merge commit 'davids/clark' into clarkClark Williams2009-10-132-4/+5
|\ \
| * | Added hardware product name and manufacturer to the screen reportDavid Sommerseth2009-10-131-0/+1
| | |
| * | Fixed wrong data type for self.config.duration parsingDavid Sommerseth2009-10-131-1/+1
| | |
| * | rteval did not consider -f|--inifile argumentDavid Sommerseth2009-10-131-3/+3
| | |
* | | version bump to 1.6Clark Williams2009-10-131-1/+1
| | | | | | | | | | | | | | | changed stats calculation to loop less added methods for retrieving services status and kthread status
* | | reworked stats calculation in Rundata::reduce()Clark Williams2009-10-131-8/+2
| | | | | | | | | | | | avoid looping so much when calculating MAD and variance
* | | added get_services() and get_kthread() methods to rteval/rteval.pyClark Williams2009-10-131-0/+26
|/ / | | | | | | Add methods to retrieve status of services and priorities of kthreads
* | version bump to 1.5Clark Williams2009-10-121-1/+1
| | | | | | | | | | updated option processing, changed stats to use less memory, changed debug output to prefix with module name
* | modified Rundata::reduce in rteval/cyclictest.py to use less memoryClark Williams2009-10-121-5/+10
| | | | | | | | | | | | | | | | | | Rundata::reduce was using a range() statement, which generated a potentially HUGE list on long runs, many times exhausting memory. Avoid the range statment and just manually iterate Also update the cyclictest::debug() method to prefix all output with cyclictest:
* | modified rteval::debug() to prefix strings with rteval:Clark Williams2009-10-121-1/+1
| | | | | | | | | | updated the rteval::debug method to prefix output with reval: to differentiate output from other objects.
* | modified load::debug to print object nameClark Williams2009-10-123-16/+16
| | | | | | | | | | | | updated the load::debug method to print the object name as well as the input string, then updated call sites for this method in rteval/hackbench.py and rteval/kcompile.py