2.16.2.12. Yardstick Test Case Description TC015¶
Processing speed with impact on energy consumption and CPU load |
|
test case id |
OPNFV_YARDSTICK_TC015_PROCESSING SPEED |
metric |
score of single cpu running, score of parallel running, energy consumption cpu load |
test purpose |
The purpose of TC015 is to evaluate the IaaS compute performance with regards to CPU processing speed with its impact on the energy consumption It measures score of single cpu running and parallel running. Energy consumption and cpu load are monitored while the cpu test is running. 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, different server types. |
test tool |
UnixBench Unixbench is the most used CPU benchmarking software tool. It can measure the performance of bash scripts, CPUs in multithreading and single threading. It can also measure the performance for parallel tasks. Also, specific disk IO for small and large files are performed. You can use it to measure either linux dedicated servers and linux vps servers, running CentOS, Debian, Ubuntu, Fedora and other distros. (UnixBench 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 UnixBench included.) Redfish API This HTTPS interface is provided by BMC of every telco grade server. Is is a standard interface. |
test description |
The UnixBench runs system benchmarks on a compute, getting information on the CPUs in the system. If the system has more than one CPU, the tests will be run twice – once with a single copy of each test running at once, and once with N N copies, where N is the number of CPUs. UnixBench will process a set of results from a single test by averaging the individual pass results into a single final value. While the cpu test is running Energy scenario run in background to monitor the number of watt consumed by the compute server on the fly. The same is done using Cpuload scenario to monitor the overall percentage of CPU used on the fly. This enables to balance the CPU score with its impact on energy consumption. Synchronized measurements enables to look at any relation between CPU load and energy consumption. |
configuration |
file: opnfv_yardstick_tc015.yaml
Duration and Interval are set globally for Energy and Cpuload, aligned with duration of UnixBench test. SLA can be set for each scenario type. Default is NA. For SLA with single_score and parallel_score, both can be set by user, default is NA. |
applicability |
Test shall be applied to node context only It can be configured with different:
Default values exist. SLA (optional) : min_score: The minimun UnixBench score that is accepted. |
usability |
This test case is one of Yardstick’s generic test. Thus it is runnable on most of the scenarios. |
references |
ETSI-NFV-TST001 |
pre-test conditions |
The target shall have unixbench installed on it. |
test sequence |
description and expected result |
step 1 |
Yardstick is connected with the target node using ssh. |
step 2 |
Energy and Cpuload are launched silently in background one after the other. Then UnixBench is invoked. All the tests are executed using the “Run” script in the top-level of UnixBench directory. The “Run” script will run a standard “index” test, and save the report in the “results” directory. Then the report is processed by “unixbench_benchmark” and checked against the SLA. While unibench runs energy and cpu load are catched periodically according to interval value. Result: Logs are stored. |
test verdict |
Fails only if SLA is not passed, or if there is a test case execution problem. |