【Raspberry Pi】ラズパイ4 ベンチマーク(UnixBench)

Raspberry Pi
この記事は約20分で読めます。
スポンサーリンク

Raspberry Pi 4…4GB!

レイアウトは3B+とそっくりですね!

よく見るとUSBとLANコネクタの位置が逆になっています。これはケースも新調ですね。

HDMI端子が小さい!2つ!充電端子もType-Cに変更されています。


というわけで、Raspberry Pi 4 4GBモデル!ついに!

invhnd
invhnd

写真を送ってもらいました!待ちきれませんでした。

イギリス在住のK君!ありがとう。顔の割に優しい!

ベンチマーク

UnixBench

Raspberry Pi 3 Model B+(比較用)

invhnd所持品。クリーンインストールしてホスト名整えたくらいなので素の性能に近いです。

sshサーバは起動しています。

sdカードは適当に転がっていたものですので、ディスクIOはご参考程度に。。

   BYTE UNIX Benchmarks (Version 5.1.3)

   System: invhnd-server-master: GNU/Linux
   OS: GNU/Linux -- 4.19.50-v7+ -- #896 SMP Thu Jun 20 16:11:44 BST 2019
   Machine: armv7l (unknown)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: ARMv7 Processor rev 4 (v7l) (0.0 bogomips)
          
   CPU 1: ARMv7 Processor rev 4 (v7l) (0.0 bogomips)
          
   CPU 2: ARMv7 Processor rev 4 (v7l) (0.0 bogomips)
          
   CPU 3: ARMv7 Processor rev 4 (v7l) (0.0 bogomips)
          
   08:19:06 up  7:36,  1 user,  load average: 0.29, 0.12, 0.04; runlevel 3

------------------------------------------------------------------------
Benchmark Run: Tue Jul 09 2019 08:19:06 - 08:47:15
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables        5070566.6 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     1400.7 MWIPS (9.8 s, 7 samples)
Execl Throughput                                958.0 lps   (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        144575.2 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           43250.0 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        370798.1 KBps  (30.0 s, 2 samples)
Pipe Throughput                              299203.4 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  53456.1 lps   (10.0 s, 7 samples)
Process Creation                               2387.3 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   1942.8 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    543.0 lpm   (60.1 s, 2 samples)
System Call Overhead                         690925.8 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0    5070566.6    434.5
Double-Precision Whetstone                       55.0       1400.7    254.7
Execl Throughput                                 43.0        958.0    222.8
File Copy 1024 bufsize 2000 maxblocks          3960.0     144575.2    365.1
File Copy 256 bufsize 500 maxblocks            1655.0      43250.0    261.3
File Copy 4096 bufsize 8000 maxblocks          5800.0     370798.1    639.3
Pipe Throughput                               12440.0     299203.4    240.5
Pipe-based Context Switching                   4000.0      53456.1    133.6
Process Creation                                126.0       2387.3    189.5
Shell Scripts (1 concurrent)                     42.4       1942.8    458.2
Shell Scripts (8 concurrent)                      6.0        543.0    905.1
System Call Overhead                          15000.0     690925.8    460.6
                                                                   ========
System Benchmarks Index Score                                         331.3

------------------------------------------------------------------------
Benchmark Run: Tue Jul 09 2019 08:47:15 - 09:15:33
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables       17396951.8 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     4796.6 MWIPS (10.0 s, 7 samples)
Execl Throughput                               2084.5 lps   (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        215442.8 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           59428.9 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        580531.0 KBps  (30.0 s, 2 samples)
Pipe Throughput                             1027388.4 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 184299.5 lps   (10.0 s, 7 samples)
Process Creation                               4370.4 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   4270.8 lpm   (60.1 s, 2 samples)
Shell Scripts (8 concurrent)                    579.1 lpm   (60.3 s, 2 samples)
System Call Overhead                        2293455.1 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   17396951.8   1490.7
Double-Precision Whetstone                       55.0       4796.6    872.1
Execl Throughput                                 43.0       2084.5    484.8
File Copy 1024 bufsize 2000 maxblocks          3960.0     215442.8    544.0
File Copy 256 bufsize 500 maxblocks            1655.0      59428.9    359.1
File Copy 4096 bufsize 8000 maxblocks          5800.0     580531.0   1000.9
Pipe Throughput                               12440.0    1027388.4    825.9
Pipe-based Context Switching                   4000.0     184299.5    460.7
Process Creation                                126.0       4370.4    346.9
Shell Scripts (1 concurrent)                     42.4       4270.8   1007.3
Shell Scripts (8 concurrent)                      6.0        579.1    965.2
System Call Overhead                          15000.0    2293455.1   1529.0
                                                                   ========
System Benchmarks Index Score                                         733.3

Raspberry Pi 4

UnixBenchやってもらいました!

最大CPUクロックが1.4GHz→1.5GHzに微増しているはず、どの程度変化があるのでしょうか。

sdカードはClass10だそうです。

   BYTE UNIX Benchmarks (Version 5.1.3)

   System: raspberrypi: GNU/Linux
   OS: GNU/Linux -- 4.19.50-v7l+ -- #895 SMP Thu Jun 20 16:03:42 BST 2019
   Machine: armv7l (unknown)
   Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
   CPU 0: ARMv7 Processor rev 3 (v7l) (0.0 bogomips)
          
   CPU 1: ARMv7 Processor rev 3 (v7l) (0.0 bogomips)
          
   CPU 2: ARMv7 Processor rev 3 (v7l) (0.0 bogomips)
          
   CPU 3: ARMv7 Processor rev 3 (v7l) (0.0 bogomips)
          
   00:16:58 up 14 min,  1 user,  load average: 0.06, 0.02, 0.00; runlevel 5

------------------------------------------------------------------------
Benchmark Run: Fri Jul 12 2019 00:16:58 - 00:44:52
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       10084878.1 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     2394.7 MWIPS (9.6 s, 7 samples)
Execl Throughput                                921.3 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        113720.5 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           31189.4 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        315540.4 KBps  (30.0 s, 2 samples)
Pipe Throughput                              161541.0 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  48524.5 lps   (10.0 s, 7 samples)
Process Creation                               1900.0 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   2661.5 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    767.2 lpm   (60.1 s, 2 samples)
System Call Overhead                         494731.5 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   10084878.1    864.2
Double-Precision Whetstone                       55.0       2394.7    435.4
Execl Throughput                                 43.0        921.3    214.3
File Copy 1024 bufsize 2000 maxblocks          3960.0     113720.5    287.2
File Copy 256 bufsize 500 maxblocks            1655.0      31189.4    188.5
File Copy 4096 bufsize 8000 maxblocks          5800.0     315540.4    544.0
Pipe Throughput                               12440.0     161541.0    129.9
Pipe-based Context Switching                   4000.0      48524.5    121.3
Process Creation                                126.0       1900.0    150.8
Shell Scripts (1 concurrent)                     42.4       2661.5    627.7
Shell Scripts (8 concurrent)                      6.0        767.2   1278.6
System Call Overhead                          15000.0     494731.5    329.8
                                                                   ========
System Benchmarks Index Score                                         327.0

------------------------------------------------------------------------
Benchmark Run: Fri Jul 12 2019 00:44:52 - 01:12:49
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables       40257467.8 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     9577.1 MWIPS (9.6 s, 7 samples)
Execl Throughput                               2771.9 lps   (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        213747.4 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           59678.8 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        623830.5 KBps  (30.0 s, 2 samples)
Pipe Throughput                              612293.6 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 191996.3 lps   (10.0 s, 7 samples)
Process Creation                               4759.0 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   6105.8 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    835.0 lpm   (60.2 s, 2 samples)
System Call Overhead                        1854552.2 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   40257467.8   3449.7
Double-Precision Whetstone                       55.0       9577.1   1741.3
Execl Throughput                                 43.0       2771.9    644.6
File Copy 1024 bufsize 2000 maxblocks          3960.0     213747.4    539.8
File Copy 256 bufsize 500 maxblocks            1655.0      59678.8    360.6
File Copy 4096 bufsize 8000 maxblocks          5800.0     623830.5   1075.6
Pipe Throughput                               12440.0     612293.6    492.2
Pipe-based Context Switching                   4000.0     191996.3    480.0
Process Creation                                126.0       4759.0    377.7
Shell Scripts (1 concurrent)                     42.4       6105.8   1440.0
Shell Scripts (8 concurrent)                      6.0        835.0   1391.7
System Call Overhead                          15000.0    1854552.2   1236.4
                                                                   ========
System Benchmarks Index Score                                         866.5

所感

スコア伸びましたね!

正直、ラズパイ4の変化はCPUよりメモリが大きいと思っていたので、驚きました。

  • 4パラ試験の最終スコアは約2割の向上!
  • 個別スコアではかなり伸びているものもありますね。
    「Dhrystone 2 using register variables」なんか1490.7→3449.7と、2.3倍まで向上!


いくつか下がっているところもありますね。

なるべく試験環境は揃えてもらったつもりなので、バックで何か負荷が、ではないと思いますが。

何に起因する値だろう。

  • Pipe Throughputは825.9→492.2
  • System Call Overheadは1529.0→1236.4

メモリ

pi@raspberrypi:~ $ free
              total        used        free      shared  buff/cache   available
Mem:        3999792       94740     3805136        8608       99916     3776636
Swap:        102396           0      102396

知ってたけど!4GBモデルだし。

数字で見ると説得力がすごいわ。余裕じゃないか・・

最後に

その他にも、思いついたらまた比較してみようと思います!

コメント

タイトルとURLをコピーしました