Jõudluse mõõtmine

Allikas: Imre kasutab arvutit
Mine navigeerimisribaleMine otsikasti

Sissejuhatus

Muu hulgas on olemas selline tarkvara

lmbench

TODO

# apt-get install lmbench

root@imre-ubuntu-2024-dpdk-01:/usr/lib/lmbench/scripts# ./config-run

root@imre-ubuntu-2024-dpdk-01:/usr/lib/lmbench/scripts# lmbench-run bw_pipe -m 10k -M 10M -P 2
/usr/bin/lmbench-run: 1: /var/lib/lmbench/config: Permission denied
Using config in CONFIG.imre-ubuntu-2024-dpdk-01
Sun Jul 28 12:48:06 UTC 2024
Latency measurements
Sun Jul 28 12:48:38 UTC 2024
Calculating file system latency
Sun Jul 28 12:48:39 UTC 2024
Local networking
Sun Jul 28 12:51:47 UTC 2024
Bandwidth measurements
Sun Jul 28 12:54:24 UTC 2024
Calculating context switch overhead
Sun Jul 28 12:54:28 UTC 2024
Calculating effective TLB size
Sun Jul 28 12:54:30 UTC 2024
Calculating memory load parallelism
Sun Jul 28 13:01:38 UTC 2024
McCalpin's STREAM benchmark
Sun Jul 28 13:01:46 UTC 2024
Calculating memory load latency
Sun Jul 28 13:43:20 UTC 2024
Benchmark run finished....
Remember you can find the results of the benchmark 
under /var/lib/lmbench/results

Tulemus salvestatakse faili, muu hulgas on seal registreeritud nt

root@imre-ubuntu-2024-dpdk-01:~# grep -A 8 "libc bcopy unaligned" /var/lib/lmbench/results/x86_64-linux-gnu/imre-ubuntu-2024-dpdk-01.0 
"libc bcopy unaligned
0.000512 112923.37
0.001024 109816.96
0.002048 114732.77
0.004096 99844.64
0.008192 109028.73
0.016384 109222.73
0.032768 59181.28
0.065536 58554.25
...

sysbench

TODO

root@imre-ubuntu-2024-dpdk-01:/usr/lib/lmbench/scripts# sysbench --time=60 --threads=6 --memory-block-size=1K --memory-scope=global --memory-total-size=1000G --memory-oper=read memory run
sysbench 1.0.20 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 6
Initializing random number generator from current time


Running memory speed test with the following options:
  block size: 1KiB
  total size: 1024000MiB
  operation: read
  scope: global

Initializing worker threads...

Threads started!

Total operations: 1048575996 (29916609.93 per second)

1024000.00 MiB transferred (29215.44 MiB/sec)


General statistics:
    total time:                          35.0493s
    total number of events:              1048575996

Latency (ms):
         min:                                    0.00
         avg:                                    0.00
         max:                                    0.20
         95th percentile:                        0.00
         sum:                                33916.86

Threads fairness:
    events (avg/stddev):           174762666.0000/0.00
    execution time (avg/stddev):   5.6528/0.02

kus

  • taustal on nt top väljundis paista, et tõesti tegevused toimuvad kuue cpu peal; ja kuus threadi on rakendatud

PTS

TODO

Kasulikud lisamaterjalid