1 .. This work is licensed under a Creative Commons Attribution 4.0 International
3 .. http://creativecommons.org/licenses/by/4.0
4 .. (c) OPNFV, Huawei Technologies Co.,Ltd and others.
6 *************************************
7 Yardstick Test Case Description TC069
8 *************************************
10 .. _RAMspeed: http://alasir.com/software/ramspeed/
15 +-----------------------------------------------------------------------------+
18 +--------------+--------------------------------------------------------------+
19 |test case id | OPNFV_YARDSTICK_TC069_Memory Bandwidth |
21 +--------------+--------------------------------------------------------------+
22 |metric | Megabyte per second (MBps) |
24 +--------------+--------------------------------------------------------------+
25 |test purpose | To evaluate the IaaS compute performance with regards to |
26 | | memory bandwidth. |
27 | | Measure the maximum possible cache and memory performance |
28 | | while reading and writing certain blocks of data (starting |
29 | | from 1Kb and further in power of 2) continuously through ALU |
30 | | and FPU respectively. |
31 | | Measure different aspects of memory performance via |
32 | | synthetic simulations. Each simulation consists of four |
33 | | performances (Copy, Scale, Add, Triad). |
34 | | Test results, graphs and similar shall be stored for |
35 | | comparison reasons and product evolution understanding |
36 | | between different OPNFV versions and/or configurations. |
38 +--------------+--------------------------------------------------------------+
39 |configuration | File: opnfv_yardstick_tc069.yaml |
41 | | * SLA (optional): 7000 (MBps) min_bandwidth: The minimum |
42 | | amount of memory bandwidth that is accepted. |
43 | | * type_id: 1 - runs a specified benchmark |
44 | | (by an ID number): |
45 | | 1 -- INTmark [writing] 4 -- FLOATmark [writing] |
46 | | 2 -- INTmark [reading] 5 -- FLOATmark [reading] |
47 | | 3 -- INTmem 6 -- FLOATmem |
48 | | * block_size: 64 Megabytes - the maximum block |
50 | | * load: 32 Gigabytes - the amount of data load per pass. |
51 | | * iterations: 5 - test is run 5 times iteratively. |
52 | | * interval: 1 - there is 1 second delay between each |
55 +--------------+--------------------------------------------------------------+
56 |test tool | RAMspeed |
58 | | RAMspeed is a free open source command line utility to |
59 | | measure cache and memory performance of computer systems. |
60 | | RAMspeed is not always part of a Linux distribution, hence |
61 | | it needs to be installed in the test image. |
63 +--------------+--------------------------------------------------------------+
64 |references | RAMspeed_ |
68 +--------------+--------------------------------------------------------------+
69 |applicability | Test can be configured with different: |
71 | | * benchmark operations (such as INTmark [writing], |
72 | | INTmark [reading], FLOATmark [writing], |
73 | | FLOATmark [reading], INTmem, FLOATmem); |
74 | | * block size per array; |
75 | | * load per pass; |
76 | | * number of batch run iterations; |
77 | | * iterations and intervals. |
79 | | There are default values for each above-mentioned option. |
81 +--------------+--------------------------------------------------------------+
82 |pre-test | The test case image needs to be installed into Glance |
83 |conditions | with RAmspeed included in the image. |
85 | | No POD specific requirements have been identified. |
87 +--------------+--------------------------------------------------------------+
88 |test sequence | description and expected result |
90 +--------------+--------------------------------------------------------------+
91 |step 1 | The host is installed as client. RAMspeed is invoked and |
92 | | logs are produced and stored. |
94 | | Result: logs are stored. |
96 +--------------+--------------------------------------------------------------+
97 |test verdict | Test fails if the measured memory bandwidth is below the SLA |
98 | | value or if there is a test case execution problem. |
100 +--------------+--------------------------------------------------------------+