From 637eb126bb4ed30df61a293e03d36bd8ad5265e5 Mon Sep 17 00:00:00 2001 From: Maciek Konstantynowicz Date: Fri, 17 Aug 2018 15:31:08 +0100 Subject: report 18.07: moved performance change/compare sections out of rls notes to separate sections for vpp and dpdk. Change-Id: If620cd894ea906baf0391617e5440e34f7136c37 Signed-off-by: Maciek Konstantynowicz --- .../comparisons/3n-skx_vs_3n-hsw_testbeds.rst | 61 ++++++++ .../comparisons/current_vs_previous_release.rst | 76 ++++++++++ .../vpp_performance_tests/comparisons/index.rst | 13 ++ .../vpp_performance_tests/csit_release_notes.rst | 157 --------------------- docs/report/vpp_performance_tests/index.rst | 2 + .../vpp_performance_tests/throughput_trending.rst | 17 +++ 6 files changed, 169 insertions(+), 157 deletions(-) create mode 100644 docs/report/vpp_performance_tests/comparisons/3n-skx_vs_3n-hsw_testbeds.rst create mode 100644 docs/report/vpp_performance_tests/comparisons/current_vs_previous_release.rst create mode 100644 docs/report/vpp_performance_tests/comparisons/index.rst create mode 100644 docs/report/vpp_performance_tests/throughput_trending.rst (limited to 'docs/report/vpp_performance_tests') diff --git a/docs/report/vpp_performance_tests/comparisons/3n-skx_vs_3n-hsw_testbeds.rst b/docs/report/vpp_performance_tests/comparisons/3n-skx_vs_3n-hsw_testbeds.rst new file mode 100644 index 0000000000..c9e0bbbb99 --- /dev/null +++ b/docs/report/vpp_performance_tests/comparisons/3n-skx_vs_3n-hsw_testbeds.rst @@ -0,0 +1,61 @@ +3n-Skx vs. 3n-Hsw Testbeds +-------------------------- + +Relative comparison of |vpp-release| packet throughput (NDR, PDR and +MRR) is calculated for the same tests executed on 3-Node Skylake (3n- +skx) and 3-Node Haswell (3n-hsw) physical testbed types, in 1-core, +2-core and 4-core configurations. + +NDR Comparison +~~~~~~~~~~~~~~ + +NDR comparison between testbed types is available in CSV and pretty +ASCII formats: + + - `CSV NDR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-ndr.csv>`_, + - `ASCII NDR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-ndr.txt>`_. + +.. note:: + + Test results have been generated by + `FD.io test executor vpp performance job 3n-hsw`_ and + `FD.io test executor vpp performance job 3n-skx`_ + with RF result + files csit-vpp-perf-|srelease|-\*.zip + `archived here <../_static/archive/>`_. + +PDR Comparison +~~~~~~~~~~~~~~ + +PDR comparison between testbed types is available in CSV and pretty +ASCII formats: + + - `CSV PDR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-pdr.csv>`_, + - `ASCII PDR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-pdr.txt>`_. + +.. note:: + + Test results have been generated by + `FD.io test executor vpp performance job 3n-hsw`_ and + `FD.io test executor vpp performance job 3n-skx`_ + with RF result + files csit-vpp-perf-|srelease|-\*.zip + `archived here <../_static/archive/>`_. + +MRR Comparison +~~~~~~~~~~~~~~ + +MRR comparison between testbed types is available in CSV and pretty +ASCII formats: + + - `CSV MRR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-mrr.csv>`_, + - `ASCII MRR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-mrr.txt>`_. + +.. note:: + + Test results have been generated by + `FD.io test executor vpp performance job 3n-hsw`_ and + `FD.io test executor vpp performance job 3n-skx`_ + with RF result + files csit-vpp-perf-|srelease|-\*.zip + `archived here <../_static/archive/>`_. diff --git a/docs/report/vpp_performance_tests/comparisons/current_vs_previous_release.rst b/docs/report/vpp_performance_tests/comparisons/current_vs_previous_release.rst new file mode 100644 index 0000000000..c2a87dc2ab --- /dev/null +++ b/docs/report/vpp_performance_tests/comparisons/current_vs_previous_release.rst @@ -0,0 +1,76 @@ +Current vs. Previous Release +---------------------------- + +Relative comparison of VPP packet throughput (NDR, PDR and MRR) between +|vpp-release| and |vpp-release-1| (measured for |csit-release| and +|csit-release-1| respectively) is calculated from results of tests +running on 3-Node Intel Xeon Haswell testbeds (3n-hsw) in 1-core, 2-core +and 4-core (MRR only) configurations. + +Listed mean and standard deviation values are computed based on a series +of the same tests executed against respective VPP releases to verify +test results repeatability, with percentage change calculated for mean +values. Note that the standard deviation is quite high for a small +number of packet throughput tests, what indicates poor test results +repeatability and makes the relative change of mean throughput value not +fully representative for these tests. The root causes behind poor +results repeatability vary between the test cases. + +NDR Comparison +~~~~~~~~~~~~~~ + +NDR throughput comparison between releases is available in CSV and pretty +ASCII formats: + + - `CSV 1t1c NDR changes <../_static/vpp/performance-changes-1t1c-ndr.csv>`_, + - `CSV 2t2c NDR changes <../_static/vpp/performance-changes-2t2c-ndr.csv>`_, + - `ASCII 1t1c NDR changes <../_static/vpp/performance-changes-1t1c-ndr.txt>`_, + - `ASCII 2t2c NDR changes <../_static/vpp/performance-changes-2t2c-ndr.txt>`_. + +.. note:: + + Test results have been generated by + `FD.io test executor vpp performance job 3n-hsw`_, + with RF result + files csit-vpp-perf-|srelease|-\*.zip + `archived here <../_static/archive/>`_. + +PDR Comparison +~~~~~~~~~~~~~~ + +PDR throughput comparison between releases is available in CSV and pretty +ASCII formats: + + - `CSV 1t1c PDR changes <../_static/vpp/performance-changes-1t1c-pdr.csv>`_, + - `CSV 2t2c PDR changes <../_static/vpp/performance-changes-2t2c-pdr.csv>`_, + - `ASCII 1t1c PDR changes <../_static/vpp/performance-changes-1t1c-pdr.txt>`_, + - `ASCII 2t2c PDR changes <../_static/vpp/performance-changes-2t2c-pdr.txt>`_. + +.. note:: + + Test results have been generated by + `FD.io test executor vpp performance job 3n-hsw`_, + with RF result + files csit-vpp-perf-|srelease|-\*.zip + `archived here <../_static/archive/>`_. + +MRR Comparison +~~~~~~~~~~~~~~ + +MRR throughput comparison between releases is available in CSV and pretty +ASCII formats: + + - `CSV 1t1c MRR changes <../_static/vpp/performance-changes-1t1c-mrr.csv>`_, + - `CSV 2t2c MRR changes <../_static/vpp/performance-changes-2t2c-mrr.csv>`_, + - `CSV 4t4c MRR changes <../_static/vpp/performance-changes-4t4c-mrr.csv>`_, + - `ASCII 1t1c MRR changes <../_static/vpp/performance-changes-1t1c-mrr.txt>`_, + - `ASCII 2t2c MRR changes <../_static/vpp/performance-changes-2t2c-mrr.txt>`_, + - `ASCII 4t4c MRR changes <../_static/vpp/performance-changes-4t4c-mrr.txt>`_. + +.. note:: + + Test results have been generated by + `FD.io test executor vpp performance job 3n-hsw`_, + with RF result + files csit-vpp-perf-|srelease|-\*.zip + `archived here <../_static/archive/>`_. diff --git a/docs/report/vpp_performance_tests/comparisons/index.rst b/docs/report/vpp_performance_tests/comparisons/index.rst new file mode 100644 index 0000000000..a9179c6900 --- /dev/null +++ b/docs/report/vpp_performance_tests/comparisons/index.rst @@ -0,0 +1,13 @@ + +.. raw:: latex + + \clearpage + +Comparisons +=========== + +.. toctree:: + + current_vs_previous_release + 3n-skx_vs_3n-hsw_testbeds + diff --git a/docs/report/vpp_performance_tests/csit_release_notes.rst b/docs/report/vpp_performance_tests/csit_release_notes.rst index c9df5478bc..0318e7a01e 100644 --- a/docs/report/vpp_performance_tests/csit_release_notes.rst +++ b/docs/report/vpp_performance_tests/csit_release_notes.rst @@ -54,163 +54,6 @@ Changes in |csit-release| - **General Code Housekeeping**: Ongoing RF keywords optimizations, removal of redundant RF keywords. -Performance Changes -------------------- - -Relative performance changes in measured NDR, PDR and MRR packet -throughput in |csit-release| are calculated against the test results -from |csit-release-1| report, for tests running on 3-Node Intel Xeon -Haswell testbeds (3n-hsw) in 1-core, 2-core and 4-core (MRR only) -configurations. - -Listed mean and standard deviation values are computed based on a series -of the same tests executed against respective VPP releases to verify -test results repeatability, with percentage change calculated for mean -values. Note that the standard deviation is quite high for a small -number of packet throughput tests, what indicates poor test results -repeatability and makes the relative change of mean throughput value not -fully representative for these tests. The root causes behind poor -results repeatability vary between the test cases. - -NDR Changes -~~~~~~~~~~~ - -NDR throughput changes between releases are available in CSV and pretty -ASCII formats: - - - `CSV 1t1c NDR changes <../_static/vpp/performance-changes-1t1c-ndr.csv>`_, - - `CSV 2t2c NDR changes <../_static/vpp/performance-changes-2t2c-ndr.csv>`_, - - `ASCII 1t1c NDR changes <../_static/vpp/performance-changes-1t1c-ndr.txt>`_, - - `ASCII 2t2c NDR changes <../_static/vpp/performance-changes-2t2c-ndr.txt>`_. - -.. note:: - - Test results have been generated by - `FD.io test executor vpp performance job 3n-hsw`_, - with RF result - files csit-vpp-perf-|srelease|-\*.zip - `archived here <../_static/archive/>`_. - -PDR Changes -~~~~~~~~~~~ - -PDR throughput changes between releases are available in CSV and pretty -ASCII formats: - - - `CSV 1t1c PDR changes <../_static/vpp/performance-changes-1t1c-pdr.csv>`_, - - `CSV 2t2c PDR changes <../_static/vpp/performance-changes-2t2c-pdr.csv>`_, - - `ASCII 1t1c PDR changes <../_static/vpp/performance-changes-1t1c-pdr.txt>`_, - - `ASCII 2t2c PDR changes <../_static/vpp/performance-changes-2t2c-pdr.txt>`_. - -.. note:: - - Test results have been generated by - `FD.io test executor vpp performance job 3n-hsw`_, - with RF result - files csit-vpp-perf-|srelease|-\*.zip - `archived here <../_static/archive/>`_. - -MRR Changes -~~~~~~~~~~~ - -MRR throughput changes between releases are available in CSV and pretty -ASCII formats: - - - `CSV 1t1c MRR changes <../_static/vpp/performance-changes-1t1c-mrr.csv>`_, - - `CSV 2t2c MRR changes <../_static/vpp/performance-changes-2t2c-mrr.csv>`_, - - `CSV 4t4c MRR changes <../_static/vpp/performance-changes-4t4c-mrr.csv>`_, - - `ASCII 1t1c MRR changes <../_static/vpp/performance-changes-1t1c-mrr.txt>`_, - - `ASCII 2t2c MRR changes <../_static/vpp/performance-changes-2t2c-mrr.txt>`_, - - `ASCII 4t4c MRR changes <../_static/vpp/performance-changes-4t4c-mrr.txt>`_. - -.. note:: - - Test results have been generated by - `FD.io test executor vpp performance job 3n-hsw`_, - with RF result - files csit-vpp-perf-|srelease|-\*.zip - `archived here <../_static/archive/>`_. - -Skx vs. Hsw Comparison ----------------------- - -Relative performance comparison in measured NDR, PDR and MRR packet -throughput is calculated for tests executed on 3-Node Skylake (3n-skx) -and 3-Node Haswell (3n-hsw) physical testbed types in 1-core -configurations. - -NDR Comparison -~~~~~~~~~~~~~~ - -NDR comparison between testbed types is available in CSV and pretty -ASCII formats: - - - `CSV 1c NDR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-ndr.csv>`_, - - `ASCII 1c NDR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-ndr.txt>`_. - -.. note:: - - Test results have been generated by - `FD.io test executor vpp performance job 3n-hsw`_ and - `FD.io test executor vpp performance job 3n-skx`_ - with RF result - files csit-vpp-perf-|srelease|-\*.zip - `archived here <../_static/archive/>`_. - -PDR Comparison -~~~~~~~~~~~~~~ - -PDR comparison between testbed types is available in CSV and pretty -ASCII formats: - - - `CSV 1c PDR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-pdr.csv>`_, - - `ASCII 1c PDR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-pdr.txt>`_. - -.. note:: - - Test results have been generated by - `FD.io test executor vpp performance job 3n-hsw`_ and - `FD.io test executor vpp performance job 3n-skx`_ - with RF result - files csit-vpp-perf-|srelease|-\*.zip - `archived here <../_static/archive/>`_. - -MRR Comparison -~~~~~~~~~~~~~~ - -MRR comparison between testbed types is available in CSV and pretty -ASCII formats: - - - `CSV 1c MRR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-mrr.csv>`_, - - `ASCII 1c MRR comparison <../_static/vpp/performance-compare-testbeds-3n-hsw-3n-skx-mrr.txt>`_. - -.. note:: - - Test results have been generated by - `FD.io test executor vpp performance job 3n-hsw`_ and - `FD.io test executor vpp performance job 3n-skx`_ - with RF result - files csit-vpp-perf-|srelease|-\*.zip - `archived here <../_static/archive/>`_. - -Throughput Trending -------------------- - -In addition to reporting throughput changes between VPP releases, CSIT -provides continuous performance trending for VPP master branch: - -#. `VPP Performance Dashboard `_ - - per VPP test case throughput trend, trend compliance and summary of - detected anomalies. - -#. `Trending Methodology `_ - - throughput test metrics, trend calculations and anomaly - classification (progression, regression, outlier). - -#. `Trendline Graphs `_ - - per VPP build MRR throughput measurements against the trendline - with anomaly highlights, with associated CSIT test jobs. - Known Issues ------------ diff --git a/docs/report/vpp_performance_tests/index.rst b/docs/report/vpp_performance_tests/index.rst index abef2c8667..888be08ac2 100644 --- a/docs/report/vpp_performance_tests/index.rst +++ b/docs/report/vpp_performance_tests/index.rst @@ -9,5 +9,7 @@ VPP Performance throughput_speedup_multi_core/index packet_latency_graphs/index http_server_performance/index + comparisons/index + throughput_trending test_environment documentation/index diff --git a/docs/report/vpp_performance_tests/throughput_trending.rst b/docs/report/vpp_performance_tests/throughput_trending.rst new file mode 100644 index 0000000000..c2ca421d57 --- /dev/null +++ b/docs/report/vpp_performance_tests/throughput_trending.rst @@ -0,0 +1,17 @@ +Throughput Trending +------------------- + +In addition to reporting throughput comparison between VPP releases, +CSIT provides continuous performance trending for VPP master branch: + +#. `VPP Performance Dashboard `_ + - per VPP test case throughput trend, trend compliance and summary of + detected anomalies. + +#. `Trending Methodology `_ + - throughput test metrics, trend calculations and anomaly + classification (progression, regression). + +#. `Trendline Graphs `_ + - per VPP build MRR throughput measurements against the trendline + with anomaly highlights and associated CSIT test jobs. \ No newline at end of file -- cgit 1.2.3-korg