From 8272a44abd449661e78b9a9945f9ca9677fe98a8 Mon Sep 17 00:00:00 2001 From: JingLu5 Date: Tue, 15 Nov 2016 11:42:10 +0800 Subject: enhance test cases description and provide more info(in progress) JIRA: YARDSTICK-389 This work is about to improve test case.rst to provide more info Change-Id: If27fe462a43f6344a06e61944b72e912d2a516b7 Signed-off-by: JingLu5 --- docs/userguide/opnfv_yardstick_tc010.rst | 116 +++++++++++++++++++++++++------ 1 file changed, 94 insertions(+), 22 deletions(-) (limited to 'docs/userguide/opnfv_yardstick_tc010.rst') diff --git a/docs/userguide/opnfv_yardstick_tc010.rst b/docs/userguide/opnfv_yardstick_tc010.rst index ab793de76..202307de6 100644 --- a/docs/userguide/opnfv_yardstick_tc010.rst +++ b/docs/userguide/opnfv_yardstick_tc010.rst @@ -7,21 +7,71 @@ Yardstick Test Case Description TC010 ************************************* -.. _man-pages: http://manpages.ubuntu.com/manpages/trusty/lat_mem_rd.8.html +.. _lat_mem_rd: http://manpages.ubuntu.com/manpages/trusty/lat_mem_rd.8.html +-----------------------------------------------------------------------------+ |Memory Latency | | | +--------------+--------------------------------------------------------------+ -|test case id | OPNFV_YARDSTICK_TC010_Memory Latency | +|test case id | OPNFV_YARDSTICK_TC010_MEMORY LATENCY | | | | +--------------+--------------------------------------------------------------+ -|metric | Latency in nanoseconds | +|metric | Memory read latency (nanoseconds) | | | | +--------------+--------------------------------------------------------------+ -|test purpose | Measure the memory read latency for varying memory sizes and | -| | strides. Whole memory hierarchy is measured including all | -| | levels of cache. | +|test purpose | The purpose of TC010 is to evaluate the IaaS compute | +| | performance with regards to memory read latency. | +| | It measures the memory read latency for varying memory sizes | +| | and strides. Whole memory hierarchy is measured. | +| | | +| | The purpose is also to be able to spot the trends. | +| | Test results, graphs and similar shall be stored for | +| | comparison reasons and product evolution understanding | +| | between different OPNFV versions and/or configurations. | +| | | ++--------------+--------------------------------------------------------------+ +|test tool | Lmbench | +| | | +| | Lmbench is a suite of operating system microbenchmarks. This | +| | test uses lat_mem_rd tool from that suite including: | +| | * Context switching | +| | * Networking: connection establishment, pipe, TCP, UDP, and | +| | RPC hot potato | +| | * File system creates and deletes | +| | * Process creation | +| | * Signal handling | +| | * System call overhead | +| | * Memory read latency | +| | | +| | (LMbench is not always part of a Linux distribution, hence | +| | it needs to be installed. As an example see the | +| | /yardstick/tools/ directory for how to generate a Linux | +| | image with LMbench included.) | +| | | ++--------------+--------------------------------------------------------------+ +|test | LMbench lat_mem_rd benchmark measures memory read latency | +|description | for varying memory sizes and strides. | +| | | +| | The benchmark runs as two nested loops. The outer loop is | +| | the stride size. The inner loop is the array size. For each | +| | array size, the benchmark creates a ring of pointers that | +| | point backward one stride.Traversing the array is done by: | +| | | +| | p = (char **)*p; | +| | | +| | in a for loop (the over head of the for loop is not | +| | significant; the loop is an unrolled loop 100 loads long). | +| | The size of the array varies from 512 bytes to (typically) | +| | eight megabytes. For the small sizes, the cache will have an | +| | effect, and the loads will be much faster. This becomes much | +| | more apparent when the data is plotted. | +| | | +| | Only data accesses are measured; the instruction cache is | +| | not measured. | +| | | +| | The results are reported in nanoseconds per load and have | +| | been verified accurate to within a few nanoseconds on an SGI | +| | Indy. | | | | +--------------+--------------------------------------------------------------+ |configuration | File: opnfv_yardstick_tc010.yaml | @@ -33,20 +83,13 @@ Yardstick Test Case Description TC010 | | * Interval: 1 - there is 1 second delay between each | | | iteration. | | | | -+--------------+--------------------------------------------------------------+ -|test tool | Lmbench | -| | | -| | Lmbench is a suite of operating system microbenchmarks. This | -| | test uses lat_mem_rd tool from that suite. | -| | Lmbench is not always part of a Linux distribution, hence it | -| | needs to be installed in the test image | -| | | -+--------------+--------------------------------------------------------------+ -|references | man-pages_ | -| | | -| | McVoy, Larry W.,and Carl Staelin. "lmbench: Portable Tools | -| | for Performance Analysis." USENIX annual technical | -| | conference 1996. | +| | SLA is optional. The SLA in this test case serves as an | +| | example. Considerably lower read latency is expected. | +| | However, to cover most configurations, both baremetal and | +| | fully virtualized ones, this value should be possible to | +| | achieve and acceptable for black box testing. | +| | Many heavy IO applications start to suffer badly if the | +| | read latency is higher than this. | | | | +--------------+--------------------------------------------------------------+ |applicability | Test can be configured with different: | @@ -55,12 +98,21 @@ Yardstick Test Case Description TC010 | | * stop_size; | | | * iterations and intervals. | | | | -| | There are default values for each above-mentioned option. | +| | Default values exist. | | | | | | SLA (optional) : max_latency: The maximum memory latency | | | that is accepted. | | | | +--------------+--------------------------------------------------------------+ +|usability | This test case is one of Yardstick's generic test. Thus it | +| | is runnable on most of the scenarios. | +| | | ++--------------+--------------------------------------------------------------+ +|references | LMbench lat_mem_rd_ | +| | | +| | ETSI-NFV-TST001 | +| | | ++--------------+--------------------------------------------------------------+ |pre-test | The test case image needs to be installed into Glance | |conditions | with Lmbench included in the image. | | | | @@ -70,12 +122,32 @@ Yardstick Test Case Description TC010 |test sequence | description and expected result | | | | +--------------+--------------------------------------------------------------+ -|step 1 | The host is installed as client. Lmbench's lat_mem_rd tool | +|step 1 | The host is installed as client. LMbench's lat_mem_rd tool | | | is invoked and logs are produced and stored. | | | | | | Result: logs are stored. | | | | +--------------+--------------------------------------------------------------+ +|step 1 | A host VM with LMbench installed is booted. | +| | | ++--------------+--------------------------------------------------------------+ +|step 2 | Yardstick is connected with the host VM by using ssh. | +| | 'lmbench_latency_benchmark' bash script is copyied from Jump | +| | Host to the host VM via the ssh tunnel. | +| | | ++--------------+--------------------------------------------------------------+ +|step 3 | 'lmbench_latency_benchmark' script is invoked. LMbench's | +| | lat_mem_rd benchmark starts to measures memory read latency | +| | for varying memory sizes and strides. Memory read latency | +| | are recorded and checked against the SLA. Logs are produced | +| | and stored. | +| | | +| | Result: Logs are stored. | +| | | ++--------------+--------------------------------------------------------------+ +|step 4 | The host VM is deleted. | +| | | ++--------------+--------------------------------------------------------------+ |test verdict | Test fails if the measured memory latency is above the SLA | | | value or if there is a test case execution problem. | | | | -- cgit 1.2.3-korg