The spec sfs 2014 benchmark includes multiple workloads


For details on the rest of the parameters, go to the spec user guide [CITATION SPE17 \l 2058 ] page 39



Download 1.14 Mb.
Page4/9
Date09.09.2021
Size1.14 Mb.
#57296
1   2   3   4   5   6   7   8   9
SPECsfs
SPECsfs, SPECsfs, SPECsfs
For details on the rest of the parameters, go to the spec user guide [CITATION SPE17 \l 2058 ] page 39.



Installing the License Key

Create the netmist_license_keyfile in the current working directory (root of SPEC). This file should be a simple text file that contains:




LICENSE KEY ####

Where the ##### is the license number that you have received from SPEC. Create this file before running the benchmark. If not having a valid SPEC number, use LICENSE KEY 11.




2. Video Data Acquisition (VDA)


The workload generally simulates applications that store data acquired from a temporally volatile source (e.g. surveillance cameras). A stream refers to an instance of the application storing data from a single source (e.g. one video feed). The storage admin is concerned primarily about maintaining a minimum fixed bit rate per stream and secondarily about maintaining the fidelity of the stream. The goal of the storage admin is to provide as many simultaneous streams as possible while meeting the bit rate and fidelity constraints. The business metric for the benchmark is STREAMS. The benchmark consists of two workload objects: VDA1 (data stream) and VDA2 (companion applications). Each stream corresponds to a roughly 36 Mb/s bit rate, which is in the upper range of high definition video.


Table 3. VDA1 File Operation Distribution.



Table 4. VDA2 File Operation Distribution.
Configuration File

The following file fragment shows the prime parameters used to define a VDA test (BENCHMARK). This test has an initial load of 125 streams, with increments of 125, and runs 5 times (125,250,375,500 & 625). The mount point used in this example is the file clients.regression. It also includes the execution path that should be the same in all clients (location of netmist).




BENCHMARK=VDA

LOAD=125

INCR_LOAD=125

NUM_RUNS=5

CLIENT_MOUNTPOINTS=clients.regression

EXEC_PATH=/mnt/SPECsfs2014_SP2/bin/netmist

USER=root

WARMUP_TIME=300


The content of the clients.regression file is as follows. It includes the name of the client and the directory where the file system is mounted.


nc1 /nfs1

nc2 /nfs1

nc1 /nfs2

nc2 /nfs2




Note: this is a linux example.



Download 1.14 Mb.

Share with your friends:
1   2   3   4   5   6   7   8   9




The database is protected by copyright ©ininet.org 2024
send message

    Main page