Benchmarks

[FrankM]

Hier möchte ich alle Benchmarks usw. sammeln. Bitte unbedingt das vorher lesen! Ich werde die Version des Images dabei schreiben. Einsetzen werde ich ausschließlich folgendes Image.

https://frank-mankel.org/topic/69/bionic-minimal-rockpro64-0-7-x-228-arm64-img-xz

Macht bis jetzt für mich den stabilsten Eindruck.

[FrankM]

USB2/3 (Version 0.7.3)

Ich benutze eine SAN Disk 240GB SSD an einem Inateck USB 3.0 2,5 Zoll Adapter.

Info zum USB-Adapter

lsusb
Bus 004 Device 002: ID 174c:55aa ASMedia Technology Inc. ASM1051E SATA 6Gb/s bridge, ASM1053E SATA 6Gb/s bridge, ASM1153 SATA 3Gb/s bridge

2,5 Zoll SSD am USB2-Port

sudo dd if=/dev/zero of=sd.img bs=1M count=4096 conv=fdatasync
4096+0 records in
4096+0 records out
4294967296 bytes (4.3 GB, 4.0 GiB) copied, 160.058 s, **26.8 MB/s**

2,5 Zoll SSD am USB3 Port

sudo dd if=/dev/zero of=sd.img bs=1M count=4096 conv=fdatasync
4096+0 records in
4096+0 records out
4294967296 bytes (4.3 GB, 4.0 GiB) copied, 36.2588 s, **118 MB/s**

Der @tkaiser erreicht deutlich höhere Geschwindigkeiten. Bis zu 400 MB/s. Hier nachzulesen.

Ich habe mich mit @tkaiser noch mal unterhalten. Scheint sehr deutlich ein Problem des Adapters zu sein. Da müsste eigentlich mehr gehen. Mal sehen....

[FrankM]

7-zip (Version 0.7.3)

Kleiner Stresstest für die CPU

Installation

sudo apt-get install p7zip p7zip-full p7zip-rar 

Test

7zr b

7-Zip (a) [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=de_DE.UTF-8,Utf16=on,HugeFiles=on,64 bits,6 CPUs LE)

LE
CPU Freq:   904  1276  1530  1721  1794  1793  1793  1794  1794

RAM size:    3876 MB,  # CPU hardware threads:   6
RAM usage:   1323 MB,  # Benchmark threads:      6

                       Compressing  |                  Decompressing
Dict     Speed Usage    R/U Rating  |      Speed Usage    R/U Rating
         KiB/s     %   MIPS   MIPS  |      KiB/s     %   MIPS   MIPS

22:       4400   502    852   4281  |      92530   522   1513   7891
23:       4227   513    840   4307  |      90668   523   1500   7845
24:       4180   534    842   4495  |      88868   525   1486   7800
25:       4207   564    852   4804  |      86102   526   1457   7663
----------------------------------  | ------------------------------
Avr:             528    847   4472  |              524   1489   7800
Tot:             526   1168   6136

[FrankM]

LAN (Version 0.7.3)

Geschwindigkeit der Schnittstelle

iperf3 -c 192.168.3.213
Connecting to host 192.168.3.213, port 5201
[  4] local 192.168.3.12 port 42350 connected to 192.168.3.213 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   116 MBytes   971 Mbits/sec    0    921 KBytes       
[  4]   1.00-2.00   sec   112 MBytes   941 Mbits/sec   11    460 KBytes       
[  4]   2.00-3.00   sec   112 MBytes   941 Mbits/sec   11    339 KBytes       
[  4]   3.00-4.00   sec   112 MBytes   941 Mbits/sec   10    355 KBytes       
[  4]   4.00-5.00   sec   112 MBytes   942 Mbits/sec   11    339 KBytes       
[  4]   5.00-6.00   sec   112 MBytes   941 Mbits/sec    0    382 KBytes       
[  4]   6.00-7.00   sec   112 MBytes   941 Mbits/sec   11    324 KBytes       
[  4]   7.00-8.00   sec   112 MBytes   942 Mbits/sec   11    243 KBytes       
[  4]   8.00-9.00   sec   112 MBytes   941 Mbits/sec   10    315 KBytes       
[  4]   9.00-10.00  sec   112 MBytes   942 Mbits/sec   11    308 KBytes       
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.10 GBytes   944 Mbits/sec   86             sender
[  4]   0.00-10.00  sec  1.10 GBytes   941 Mbits/sec                  receiver

iperf Done.
rock64@rockpro64:~$ iperf3 -s
-----------------------------------------------------------
Server listening on 5201
-----------------------------------------------------------
Accepted connection from 192.168.3.213, port 35834
[  5] local 192.168.3.12 port 5201 connected to 192.168.3.213 port 35836
[ ID] Interval           Transfer     Bandwidth
[  5]   0.00-1.00   sec   108 MBytes   908 Mbits/sec                  
[  5]   1.00-2.00   sec   112 MBytes   941 Mbits/sec                  
[  5]   2.00-3.00   sec   112 MBytes   941 Mbits/sec                  
[  5]   3.00-4.00   sec   112 MBytes   941 Mbits/sec                  
[  5]   4.00-5.00   sec   112 MBytes   941 Mbits/sec                  
[  5]   5.00-6.00   sec   112 MBytes   941 Mbits/sec                  
[  5]   6.00-7.00   sec   112 MBytes   941 Mbits/sec                  
[  5]   7.00-8.00   sec   112 MBytes   941 Mbits/sec                  
[  5]   8.00-9.00   sec   112 MBytes   941 Mbits/sec                  
[  5]   9.00-10.00  sec   112 MBytes   941 Mbits/sec                  
[  5]  10.00-10.02  sec  1.85 MBytes   930 Mbits/sec                  
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth
[  5]   0.00-10.02  sec  0.00 Bytes  0.00 bits/sec                  sender
[  5]   0.00-10.02  sec  1.09 GBytes   938 Mbits/sec                  receiver
-----------------------------------------------------------
Server listening on 5201
-----------------------------------------------------------
^Ciperf3: interrupt - the server has terminated

[FrankM]

Speichertest (Version 0.7.3)

Mit dem Tool tinymembench die Geschwindigkeit des Speichers testen.

./tinymembench 
tinymembench v0.4.9 (simple benchmark for memory throughput and latency)

==========================================================================
== Memory bandwidth tests                                               ==
==                                                                      ==
== Note 1: 1MB = 1000000 bytes                                          ==
== Note 2: Results for 'copy' tests show how many bytes can be          ==
==         copied per second (adding together read and writen           ==
==         bytes would have provided twice higher numbers)              ==
== Note 3: 2-pass copy means that we are using a small temporary buffer ==
==         to first fetch data into it, and only then write it to the   ==
==         destination (source -> L1 cache, L1 cache -> destination)    ==
== Note 4: If sample standard deviation exceeds 0.1%, it is shown in    ==
==         brackets                                                     ==
==========================================================================

 C copy backwards                                     :   2868.1 MB/s (0.3%)
 C copy backwards (32 byte blocks)                    :   2860.8 MB/s
 C copy backwards (64 byte blocks)                    :   2851.0 MB/s
 C copy                                               :   2724.3 MB/s (0.1%)
 C copy prefetched (32 bytes step)                    :   2775.6 MB/s
 C copy prefetched (64 bytes step)                    :   2778.9 MB/s
 C 2-pass copy                                        :   2546.9 MB/s
 C 2-pass copy prefetched (32 bytes step)             :   2577.6 MB/s
 C 2-pass copy prefetched (64 bytes step)             :   2577.3 MB/s
 C fill                                               :   4897.9 MB/s (0.4%)
 C fill (shuffle within 16 byte blocks)               :   4895.2 MB/s
 C fill (shuffle within 32 byte blocks)               :   4896.9 MB/s
 C fill (shuffle within 64 byte blocks)               :   4898.0 MB/s
 ---
 standard memcpy                                      :   2841.6 MB/s
 standard memset                                      :   4897.1 MB/s (0.4%)
 ---
 NEON LDP/STP copy                                    :   2842.3 MB/s
 NEON LDP/STP copy pldl2strm (32 bytes step)          :   2863.3 MB/s (0.3%)
 NEON LDP/STP copy pldl2strm (64 bytes step)          :   2863.2 MB/s
 NEON LDP/STP copy pldl1keep (32 bytes step)          :   2784.3 MB/s
 NEON LDP/STP copy pldl1keep (64 bytes step)          :   2777.8 MB/s
 NEON LD1/ST1 copy                                    :   2839.5 MB/s
 NEON STP fill                                        :   4896.0 MB/s (0.4%)
 NEON STNP fill                                       :   4862.2 MB/s
 ARM LDP/STP copy                                     :   2841.1 MB/s
 ARM STP fill                                         :   4896.6 MB/s (0.4%)
 ARM STNP fill                                        :   4861.4 MB/s

==========================================================================
== Framebuffer read tests.                                              ==
==                                                                      ==
== Many ARM devices use a part of the system memory as the framebuffer, ==
== typically mapped as uncached but with write-combining enabled.       ==
== Writes to such framebuffers are quite fast, but reads are much       ==
== slower and very sensitive to the alignment and the selection of      ==
== CPU instructions which are used for accessing memory.                ==
==                                                                      ==
== Many x86 systems allocate the framebuffer in the GPU memory,         ==
== accessible for the CPU via a relatively slow PCI-E bus. Moreover,    ==
== PCI-E is asymmetric and handles reads a lot worse than writes.       ==
==                                                                      ==
== If uncached framebuffer reads are reasonably fast (at least 100 MB/s ==
== or preferably >300 MB/s), then using the shadow framebuffer layer    ==
== is not necessary in Xorg DDX drivers, resulting in a nice overall    ==
== performance improvement. For example, the xf86-video-fbturbo DDX     ==
== uses this trick.                                                     ==
==========================================================================

 NEON LDP/STP copy (from framebuffer)                 :    606.2 MB/s
 NEON LDP/STP 2-pass copy (from framebuffer)          :    560.0 MB/s
 NEON LD1/ST1 copy (from framebuffer)                 :    672.9 MB/s
 NEON LD1/ST1 2-pass copy (from framebuffer)          :    614.2 MB/s
 ARM LDP/STP copy (from framebuffer)                  :    451.0 MB/s
 ARM LDP/STP 2-pass copy (from framebuffer)           :    433.7 MB/s

==========================================================================
== Memory latency test                                                  ==
==                                                                      ==
== Average time is measured for random memory accesses in the buffers   ==
== of different sizes. The larger is the buffer, the more significant   ==
== are relative contributions of TLB, L1/L2 cache misses and SDRAM      ==
== accesses. For extremely large buffer sizes we are expecting to see   ==
== page table walk with several requests to SDRAM for almost every      ==
== memory access (though 64MiB is not nearly large enough to experience ==
== this effect to its fullest).                                         ==
==                                                                      ==
== Note 1: All the numbers are representing extra time, which needs to  ==
==         be added to L1 cache latency. The cycle timings for L1 cache ==
==         latency can be usually found in the processor documentation. ==
== Note 2: Dual random read means that we are simultaneously performing ==
==         two independent memory accesses at a time. In the case if    ==
==         the memory subsystem can't handle multiple outstanding       ==
==         requests, dual random read has the same timings as two       ==
==         single reads performed one after another.                    ==
==========================================================================

block size : single random read / dual random read
      1024 :    0.0 ns          /     0.0 ns 
      2048 :    0.0 ns          /     0.0 ns 
      4096 :    0.0 ns          /     0.0 ns 
      8192 :    0.0 ns          /     0.0 ns 
     16384 :    0.0 ns          /     0.0 ns 
     32768 :    0.0 ns          /     0.0 ns 
     65536 :    4.5 ns          /     7.2 ns 
    131072 :    6.8 ns          /     9.7 ns 
    262144 :    9.8 ns          /    12.8 ns 
    524288 :   11.4 ns          /    14.7 ns 
   1048576 :   16.3 ns          /    22.8 ns 
   2097152 :  110.8 ns          /   169.8 ns 
   4194304 :  157.2 ns          /   213.9 ns 
   8388608 :  185.0 ns          /   234.5 ns 
  16777216 :  198.8 ns          /   244.2 ns 
  33554432 :  206.9 ns          /   249.3 ns 
  67108864 :  218.7 ns          /   261.9 ns 

Vergleichsergebnisse findet man hier.

Speichertest (Version 0.7.5)

Nachdem die Version 0.7.4 unstabil lief, hier die Ergebnisse von 0.7.5

rock64@rockpro64:~/tinymembench$ ./tinymembench 
tinymembench v0.4.9 (simple benchmark for memory throughput and latency)

==========================================================================
== Memory bandwidth tests                                               ==
==                                                                      ==
== Note 1: 1MB = 1000000 bytes                                          ==
== Note 2: Results for 'copy' tests show how many bytes can be          ==
==         copied per second (adding together read and writen           ==
==         bytes would have provided twice higher numbers)              ==
== Note 3: 2-pass copy means that we are using a small temporary buffer ==
==         to first fetch data into it, and only then write it to the   ==
==         destination (source -> L1 cache, L1 cache -> destination)    ==
== Note 4: If sample standard deviation exceeds 0.1%, it is shown in    ==
==         brackets                                                     ==
==========================================================================

 C copy backwards                                     :   2668.2 MB/s
 C copy backwards (32 byte blocks)                    :   2662.3 MB/s
 C copy backwards (64 byte blocks)                    :   2659.3 MB/s
 C copy                                               :   2673.1 MB/s
 C copy prefetched (32 bytes step)                    :   2648.6 MB/s
 C copy prefetched (64 bytes step)                    :   2653.3 MB/s
 C 2-pass copy                                        :   2404.3 MB/s
 C 2-pass copy prefetched (32 bytes step)             :   2441.8 MB/s
 C 2-pass copy prefetched (64 bytes step)             :   2442.8 MB/s (1.1%)
 C fill                                               :   4808.3 MB/s (0.4%)
 C fill (shuffle within 16 byte blocks)               :   4793.4 MB/s
 C fill (shuffle within 32 byte blocks)               :   4801.1 MB/s (0.4%)
 C fill (shuffle within 64 byte blocks)               :   4810.3 MB/s (0.2%)
 ---
 standard memcpy                                      :   2677.8 MB/s
 standard memset                                      :   4809.4 MB/s (0.4%)
 ---
 NEON LDP/STP copy                                    :   2673.6 MB/s
 NEON LDP/STP copy pldl2strm (32 bytes step)          :   2691.4 MB/s (0.9%)
 NEON LDP/STP copy pldl2strm (64 bytes step)          :   2690.8 MB/s
 NEON LDP/STP copy pldl1keep (32 bytes step)          :   2743.8 MB/s (1.1%)
 NEON LDP/STP copy pldl1keep (64 bytes step)          :   2741.6 MB/s
 NEON LD1/ST1 copy                                    :   2793.6 MB/s
 NEON STP fill                                        :   4897.8 MB/s (0.6%)
 NEON STNP fill                                       :   4864.0 MB/s (0.2%)
 ARM LDP/STP copy                                     :   2802.0 MB/s
 ARM STP fill                                         :   4898.0 MB/s (0.4%)
 ARM STNP fill                                        :   4863.8 MB/s (0.2%)

==========================================================================
== Framebuffer read tests.                                              ==
==                                                                      ==
== Many ARM devices use a part of the system memory as the framebuffer, ==
== typically mapped as uncached but with write-combining enabled.       ==
== Writes to such framebuffers are quite fast, but reads are much       ==
== slower and very sensitive to the alignment and the selection of      ==
== CPU instructions which are used for accessing memory.                ==
==                                                                      ==
== Many x86 systems allocate the framebuffer in the GPU memory,         ==
== accessible for the CPU via a relatively slow PCI-E bus. Moreover,    ==
== PCI-E is asymmetric and handles reads a lot worse than writes.       ==
==                                                                      ==
== If uncached framebuffer reads are reasonably fast (at least 100 MB/s ==
== or preferably >300 MB/s), then using the shadow framebuffer layer    ==
== is not necessary in Xorg DDX drivers, resulting in a nice overall    ==
== performance improvement. For example, the xf86-video-fbturbo DDX     ==
== uses this trick.                                                     ==
==========================================================================

 NEON LDP/STP copy (from framebuffer)                 :    539.1 MB/s (1.8%)
 NEON LDP/STP 2-pass copy (from framebuffer)          :    522.1 MB/s
 NEON LD1/ST1 copy (from framebuffer)                 :    583.0 MB/s
 NEON LD1/ST1 2-pass copy (from framebuffer)          :    564.2 MB/s
 ARM LDP/STP copy (from framebuffer)                  :    373.2 MB/s (0.1%)
 ARM LDP/STP 2-pass copy (from framebuffer)           :    418.2 MB/s

==========================================================================
== Memory latency test                                                  ==
==                                                                      ==
== Average time is measured for random memory accesses in the buffers   ==
== of different sizes. The larger is the buffer, the more significant   ==
== are relative contributions of TLB, L1/L2 cache misses and SDRAM      ==
== accesses. For extremely large buffer sizes we are expecting to see   ==
== page table walk with several requests to SDRAM for almost every      ==
== memory access (though 64MiB is not nearly large enough to experience ==
== this effect to its fullest).                                         ==
==                                                                      ==
== Note 1: All the numbers are representing extra time, which needs to  ==
==         be added to L1 cache latency. The cycle timings for L1 cache ==
==         latency can be usually found in the processor documentation. ==
== Note 2: Dual random read means that we are simultaneously performing ==
==         two independent memory accesses at a time. In the case if    ==
==         the memory subsystem can't handle multiple outstanding       ==
==         requests, dual random read has the same timings as two       ==
==         single reads performed one after another.                    ==
==========================================================================

block size : single random read / dual random read
      1024 :    0.0 ns          /     0.0 ns 
      2048 :    0.0 ns          /     0.0 ns 
      4096 :    0.0 ns          /     0.0 ns 
      8192 :    0.0 ns          /     0.0 ns 
     16384 :    0.0 ns          /     0.0 ns 
     32768 :    0.0 ns          /     0.0 ns 
     65536 :    4.1 ns          /     6.5 ns 
    131072 :    6.2 ns          /     8.7 ns 
    262144 :    8.9 ns          /    11.6 ns 
    524288 :   10.3 ns          /    13.3 ns 
   1048576 :   15.0 ns          /    21.3 ns 
   2097152 :  112.3 ns          /   173.0 ns 
   4194304 :  159.5 ns          /   217.2 ns 
   8388608 :  187.3 ns          /   237.9 ns 
  16777216 :  201.0 ns          /   246.2 ns 
  33554432 :  208.5 ns          /   250.8 ns 
  67108864 :  219.7 ns          /   264.1 ns

[FrankM]

Cpu Sysbench (Version 0.7.3)

sysbench --test=cpu --cpu-max-prime=20000 run
WARNING: the --test option is deprecated. You can pass a script name or path on the command line without any options.
sysbench 1.0.11 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 1
Initializing random number generator from current time


Prime numbers limit: 20000

Initializing worker threads...

Threads started!

CPU speed:
    events per second:   697.34

General statistics:
    total time:                          10.0006s
    total number of events:              6983

Latency (ms):
         min:                                  1.42
         avg:                                  1.43
         max:                                  6.58
         95th percentile:                      1.42
         sum:                               9993.83

Threads fairness:
    events (avg/stddev):           6983.0000/0.00
    execution time (avg/stddev):   9.9938/0.00

[FrankM]

Memtester (Version 0.7.5)

Installation

rock64@rockpro64:~$ sudo apt-get install memtester

Test

Im Beispiel testen wir 3072 MB und zwar einmal. Bei 1024 5 würde man 1024 MB fünfmal testen.

rock64@rockpro64:~$ sudo memtester 3072 1
memtester version 4.3.0 (64-bit)
Copyright (C) 2001-2012 Charles Cazabon.
Licensed under the GNU General Public License version 2 (only).

pagesize is 4096
pagesizemask is 0xfffffffffffff000
want 3072MB (3221225472 bytes)
got  3072MB (3221225472 bytes), trying mlock ...locked.
Loop 1/1:
  Stuck Address       : ok         
  Random Value        : ok
  Compare XOR         : ok
  Compare SUB         : ok
  Compare MUL         : ok
  Compare DIV         : ok
  Compare OR          : ok
  Compare AND         : ok
  Sequential Increment: ok
  Solid Bits          : ok         
  Block Sequential    : ok         
  Checkerboard        : ok         
  Bit Spread          : ok         
  Bit Flip            : ok         
  Walking Ones        : ok         
  Walking Zeroes      : ok         
  8-bit Writes        : ok
  16-bit Writes       : ok

Done.

[FrankM]

cryptsetup benchmark (v.0.6.44)

Dient dem Test des verbauten Speichers.

Installation

sudo apt-get install cryptsetup

Test

rock64@rockpro64:/usr/local/sbin$ cryptsetup benchmark
# Tests are approximate using memory only (no storage IO).
PBKDF2-sha1       793173 iterations per second for 256-bit key
PBKDF2-sha256    1483134 iterations per second for 256-bit key
PBKDF2-sha512     499321 iterations per second for 256-bit key
PBKDF2-ripemd160  381023 iterations per second for 256-bit key
PBKDF2-whirlpool  172463 iterations per second for 256-bit key
argon2i       4 iterations, 387040 memory, 4 parallel threads (CPUs) for 256-bit key (requested 2000 ms time)
argon2id      4 iterations, 374949 memory, 4 parallel threads (CPUs) for 256-bit key (requested 2000 ms time)
#     Algorithm | Key |  Encryption |  Decryption
        aes-cbc   128b   621.7 MiB/s   851.2 MiB/s
    serpent-cbc   128b           N/A           N/A
    twofish-cbc   128b    80.7 MiB/s    82.7 MiB/s
        aes-cbc   256b   536.2 MiB/s   759.3 MiB/s
    serpent-cbc   256b           N/A           N/A
    twofish-cbc   256b    81.0 MiB/s    82.7 MiB/s
        aes-xts   256b   686.9 MiB/s   691.4 MiB/s
    serpent-xts   256b           N/A           N/A
    twofish-xts   256b           N/A           N/A
        aes-xts   512b   637.8 MiB/s   638.4 MiB/s
    serpent-xts   512b           N/A           N/A
    twofish-xts   512b           N/A           N/A

Zum Vergleich die Ergebnisse meines Haupt-PC's

frank@frank-MS-7A34 ~ $ cryptsetup benchmark
# Die Tests sind nur annähernd genau, da sie nicht auf die Festplatte zugreifen.
PBKDF2-sha1      1106092 iterations per second
PBKDF2-sha256     740519 iterations per second
PBKDF2-sha512     555389 iterations per second
PBKDF2-ripemd160  668734 iterations per second
PBKDF2-whirlpool  262144 iterations per second
#  Algorithm | Key |  Encryption |  Decryption
     aes-cbc   128b  1022,9 MiB/s  3369,1 MiB/s
 serpent-cbc   128b    94,4 MiB/s   345,8 MiB/s
 twofish-cbc   128b   189,5 MiB/s   342,5 MiB/s
     aes-cbc   256b   779,6 MiB/s  2751,3 MiB/s
 serpent-cbc   256b    96,9 MiB/s   343,8 MiB/s
 twofish-cbc   256b   195,0 MiB/s   335,0 MiB/s
     aes-xts   256b  2653,5 MiB/s  2619,4 MiB/s
 serpent-xts   256b   339,4 MiB/s   339,3 MiB/s
 twofish-xts   256b   340,5 MiB/s   338,3 MiB/s
     aes-xts   512b  2294,2 MiB/s  2329,1 MiB/s
 serpent-xts   512b   327,4 MiB/s   337,8 MiB/s
 twofish-xts   512b   351,5 MiB/s   343,3 MiB/s

[FrankM]

Gestern mal was praxistaugliches aufgebaut. Auf dem ROCKPro64 einen NFS-Server installiert. Die Freigabe lag auf der NVMe SSD. Das ganze dann auf meinem Haupt-PC gemountet und mal einen Star Wars Film kopiert. 8,5 GB

Konstant 97 MB/s

Das sieht doch schon mal sehr erfreulich aus!

[FrankM]

iozone Test (0.6.52)

Hardware

Hardware ist eine Samsung EVO 960 m.2 mit 250GB

Eingabe

sudo iozone -e -I -a -s 100M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2 

Ausgabe

Run began: Thu Jun 14 12:04:01 2018

	Include fsync in write timing
	O_DIRECT feature enabled
	Auto Mode
	File size set to 102400 kB
	Record Size 4 kB
	Record Size 16 kB
	Record Size 512 kB
	Record Size 1024 kB
	Record Size 16384 kB
	Command line used: iozone -e -I -a -s 100M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2
	Output is in kBytes/sec
	Time Resolution = 0.000001 seconds.
	Processor cache size set to 1024 kBytes.
	Processor cache line size set to 32 bytes.
	File stride size set to 17 * record size.
                                                              random    random     bkwd    record    stride                                    
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
          102400       4    40859    79542   101334   101666    31721    60459                                                          
          102400      16   113215   202566   234307   233091   108334   154750                                                          
          102400     512   362864   412548   359279   362810   340235   412626                                                          
          102400    1024   400478   453205   381115   385746   372378   453548                                                          
          102400   16384   583762   598047   595752   596251   590950   604690

Zum direkten Vergleich hier heute mal mit 4.17.0-rc6-1019

rock64@rockpro64:/mnt$ uname -a
Linux rockpro64 4.17.0-rc6-1019-ayufan-gfafc3e1c913f #1 SMP PREEMPT Tue Jun 12 19:06:59 UTC 2018 aarch64 aarch64 aarch64 GNU/Linux

iozone Test

rock64@rockpro64:/mnt$ sudo iozone -e -I -a -s 100M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2 
	Iozone: Performance Test of File I/O
	        Version $Revision: 3.429 $
		Compiled for 64 bit mode.
		Build: linux 

	Contributors:William Norcott, Don Capps, Isom Crawford, Kirby Collins
	             Al Slater, Scott Rhine, Mike Wisner, Ken Goss
	             Steve Landherr, Brad Smith, Mark Kelly, Dr. Alain CYR,
	             Randy Dunlap, Mark Montague, Dan Million, Gavin Brebner,
	             Jean-Marc Zucconi, Jeff Blomberg, Benny Halevy, Dave Boone,
	             Erik Habbinga, Kris Strecker, Walter Wong, Joshua Root,
	             Fabrice Bacchella, Zhenghua Xue, Qin Li, Darren Sawyer,
	             Vangel Bojaxhi, Ben England, Vikentsi Lapa.

	Run began: Sat Jun 16 06:34:43 2018

	Include fsync in write timing
	O_DIRECT feature enabled
	Auto Mode
	File size set to 102400 kB
	Record Size 4 kB
	Record Size 16 kB
	Record Size 512 kB
	Record Size 1024 kB
	Record Size 16384 kB
	Command line used: iozone -e -I -a -s 100M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2
	Output is in kBytes/sec
	Time Resolution = 0.000001 seconds.
	Processor cache size set to 1024 kBytes.
	Processor cache line size set to 32 bytes.
	File stride size set to 17 * record size.
                                                              random    random     bkwd    record    stride                                    
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
          102400       4    48672   104754   115838   116803    47894   103606                                                          
          102400      16   168084   276437   292660   295458   162550   273703                                                          
          102400     512   566572   597648   580005   589209   534508   597007                                                          
          102400    1024   585621   624443   590545   599177   569452   630098                                                          
          102400   16384   504871   754710   765558   780592   777696   753426                                                          

iozone test complete.