G2A Many GEOs
SHARE
TWEET

BPi M2 Berry benchmarks

a guest Aug 5th, 2017 209 Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
  1. root@bpi-iot-ros-ai:~# 7za b
  2.  
  3. 7-Zip (A) 9.20  Copyright (c) 1999-2010 Igor Pavlov  2010-11-18
  4. p7zip Version 9.20 (locale=en_US.UTF-8,Utf16=on,HugeFiles=on,4 CPUs)
  5.  
  6. RAM size:     998 MB,  # CPU hardware threads:   4
  7. RAM usage:    850 MB,  # Benchmark threads:      4
  8.  
  9. Dict        Compressing          |        Decompressing
  10.       Speed Usage    R/U Rating  |    Speed Usage    R/U Rating
  11.        KB/s     %   MIPS   MIPS  |     KB/s     %   MIPS   MIPS
  12.  
  13. 22:    1336   299    434   1300  |    36258   397    824   3271
  14. 23:    1314   301    444   1339  |    35334   393    823   3233
  15. 24:    1337   313    458   1437  |    35029   395    821   3249
  16. 25:    1291   317    465   1475  |    33194   383    815   3121
  17. ----------------------------------------------------------------
  18. Avr:          308    451   1388               392    821   3219
  19. Tot:          350    636   2303
  20. root@bpi-iot-ros-ai:~# sysbench --test=cpu --cpu-max-prime=20000 run --num-threads=$(grep -c '^processor' /proc/cpuinfo)
  21. sysbench 0.4.12:  multi-threaded system evaluation benchmark
  22.  
  23. Running the test with following options:
  24. Number of threads: 4
  25.  
  26. Doing CPU performance benchmark
  27.  
  28. Threads started!
  29. Done.
  30.  
  31. Maximum prime number checked in CPU test: 20000
  32.  
  33.  
  34. Test execution summary:
  35.     total time:                          118.4670s
  36.     total number of events:              10000
  37.     total time taken by event execution: 473.8089
  38.     per-request statistics:
  39.          min:                                 46.72ms
  40.          avg:                                 47.38ms
  41.          max:                                226.98ms
  42.          approx.  95 percentile:              47.70ms
  43.  
  44. Threads fairness:
  45.     events (avg/stddev):           2500.0000/4.24
  46.     execution time (avg/stddev):   118.4522/0.01
  47.  
  48. root@bpi-iot-ros-ai:~# minerd --benchmark 2>&1 | grep Total
  49. [2017-08-05 15:01:57] Total: 2.06 khash/s
  50. [2017-08-05 15:02:02] Total: 2.06 khash/s
  51. [2017-08-05 15:02:07] Total: 2.06 khash/s
  52. [2017-08-05 15:02:12] Total: 2.05 khash/s
  53. [2017-08-05 15:02:17] Total: 2.06 khash/s
  54. [2017-08-05 15:02:22] Total: 2.05 khash/s
  55. [2017-08-05 15:02:27] Total: 2.05 khash/s
  56. [2017-08-05 15:02:32] Total: 2.05 khash/s
  57. [2017-08-05 15:02:37] Total: 2.06 khash/s
  58. [2017-08-05 15:02:42] Total: 2.05 khash/s
  59. [2017-08-05 15:02:47] Total: 2.06 khash/s
  60. [2017-08-05 15:02:52] Total: 2.05 khash/s
  61. [2017-08-05 15:02:57] Total: 2.06 khash/s
  62. [2017-08-05 15:03:02] Total: 2.05 khash/s
  63. [2017-08-05 15:03:07] Total: 2.06 khash/s
  64. ^C
  65.  
  66. root@bpi-iot-ros-ai:~# /usr/local/src/tinymembench/tinymembench
  67. tinymembench v0.4.9 (simple benchmark for memory throughput and latency)
  68.  
  69. ==========================================================================
  70. == Memory bandwidth tests                                               ==
  71. ==                                                                      ==
  72. == Note 1: 1MB = 1000000 bytes                                          ==
  73. == Note 2: Results for 'copy' tests show how many bytes can be          ==
  74. ==         copied per second (adding together read and writen           ==
  75. ==         bytes would have provided twice higher numbers)              ==
  76. == Note 3: 2-pass copy means that we are using a small temporary buffer ==
  77. ==         to first fetch data into it, and only then write it to the   ==
  78. ==         destination (source -> L1 cache, L1 cache -> destination)    ==
  79. == Note 4: If sample standard deviation exceeds 0.1%, it is shown in    ==
  80. ==         brackets                                                     ==
  81. ==========================================================================
  82.  
  83.  C copy backwards                                     :    268.7 MB/s (0.1%)
  84.  C copy backwards (32 byte blocks)                    :   1024.2 MB/s
  85.  C copy backwards (64 byte blocks)                    :   1029.1 MB/s (1.4%)
  86.  C copy                                               :   1029.3 MB/s (0.4%)
  87.  C copy prefetched (32 bytes step)                    :   1112.9 MB/s (0.4%)
  88.  C copy prefetched (64 bytes step)                    :   1106.1 MB/s (0.2%)
  89.  C 2-pass copy                                        :    787.3 MB/s (0.2%)
  90.  C 2-pass copy prefetched (32 bytes step)             :    880.5 MB/s
  91.  C 2-pass copy prefetched (64 bytes step)             :    902.0 MB/s
  92.  C fill                                               :   4126.0 MB/s (0.9%)
  93.  C fill (shuffle within 16 byte blocks)               :   4113.3 MB/s (0.9%)
  94.  C fill (shuffle within 32 byte blocks)               :    376.9 MB/s (3.6%)
  95.  C fill (shuffle within 64 byte blocks)               :    356.4 MB/s (1.7%)
  96.  ---
  97.  standard memcpy                                      :   1110.6 MB/s (0.3%)
  98.  standard memset                                      :   3571.6 MB/s (0.2%)
  99.  ---
  100.  NEON read                                            :   1300.5 MB/s
  101.  NEON read prefetched (32 bytes step)                 :   1474.6 MB/s
  102.  NEON read prefetched (64 bytes step)                 :   1476.7 MB/s
  103.  NEON read 2 data streams                             :    367.2 MB/s
  104.  NEON read 2 data streams prefetched (32 bytes step)  :    706.9 MB/s
  105.  NEON read 2 data streams prefetched (64 bytes step)  :    752.3 MB/s
  106.  NEON copy                                            :   1049.4 MB/s (0.3%)
  107.  NEON copy prefetched (32 bytes step)                 :   1034.0 MB/s (0.3%)
  108.  NEON copy prefetched (64 bytes step)                 :   1124.4 MB/s (0.8%)
  109.  NEON unrolled copy                                   :   1040.3 MB/s (0.4%)
  110.  NEON unrolled copy prefetched (32 bytes step)        :   1055.4 MB/s (0.3%)
  111.  NEON unrolled copy prefetched (64 bytes step)        :   1093.3 MB/s (0.3%)
  112.  NEON copy backwards                                  :   1041.2 MB/s (0.3%)
  113.  NEON copy backwards prefetched (32 bytes step)       :   1028.5 MB/s (0.5%)
  114.  NEON copy backwards prefetched (64 bytes step)       :   1088.7 MB/s (0.3%)
  115.  NEON 2-pass copy                                     :    884.0 MB/s (1.2%)
  116.  NEON 2-pass copy prefetched (32 bytes step)          :    952.6 MB/s
  117.  NEON 2-pass copy prefetched (64 bytes step)          :    990.1 MB/s (0.2%)
  118.  NEON unrolled 2-pass copy                            :    773.3 MB/s (0.2%)
  119.  NEON unrolled 2-pass copy prefetched (32 bytes step) :    751.4 MB/s (0.5%)
  120.  NEON unrolled 2-pass copy prefetched (64 bytes step) :    784.5 MB/s
  121.  NEON fill                                            :   3908.3 MB/s
  122.  NEON fill backwards                                  :   3918.2 MB/s (0.5%)
  123.  VFP copy                                             :   1043.8 MB/s
  124.  VFP 2-pass copy                                      :    777.2 MB/s
  125.  ARM fill (STRD)                                      :   3009.9 MB/s
  126.  ARM fill (STM with 8 registers)                      :   3927.3 MB/s (0.5%)
  127.  ARM fill (STM with 4 registers)                      :   3570.3 MB/s
  128.  ARM copy prefetched (incr pld)                       :   1073.5 MB/s (0.2%)
  129.  ARM copy prefetched (wrap pld)                       :   1081.0 MB/s (2.3%)
  130.  ARM 2-pass copy prefetched (incr pld)                :    876.7 MB/s (0.3%)
  131.  ARM 2-pass copy prefetched (wrap pld)                :    829.5 MB/s
  132.  
  133. ==========================================================================
  134. == Framebuffer read tests.                                              ==
  135. ==                                                                      ==
  136. == Many ARM devices use a part of the system memory as the framebuffer, ==
  137. == typically mapped as uncached but with write-combining enabled.       ==
  138. == Writes to such framebuffers are quite fast, but reads are much       ==
  139. == slower and very sensitive to the alignment and the selection of      ==
  140. == CPU instructions which are used for accessing memory.                ==
  141. ==                                                                      ==
  142. == Many x86 systems allocate the framebuffer in the GPU memory,         ==
  143. == accessible for the CPU via a relatively slow PCI-E bus. Moreover,    ==
  144. == PCI-E is asymmetric and handles reads a lot worse than writes.       ==
  145. ==                                                                      ==
  146. == If uncached framebuffer reads are reasonably fast (at least 100 MB/s ==
  147. == or preferably >300 MB/s), then using the shadow framebuffer layer    ==
  148. == is not necessary in Xorg DDX drivers, resulting in a nice overall    ==
  149. == performance improvement. For example, the xf86-video-fbturbo DDX     ==
  150. == uses this trick.                                                     ==
  151. ==========================================================================
  152.  
  153.  NEON read (from framebuffer)                         :     50.1 MB/s
  154.  NEON copy (from framebuffer)                         :     49.4 MB/s
  155.  NEON 2-pass copy (from framebuffer)                  :     49.3 MB/s
  156.  NEON unrolled copy (from framebuffer)                :     50.0 MB/s
  157.  NEON 2-pass unrolled copy (from framebuffer)         :     49.0 MB/s
  158.  VFP copy (from framebuffer)                          :    343.2 MB/s (0.2%)
  159.  VFP 2-pass copy (from framebuffer)                   :    304.6 MB/s
  160.  ARM copy (from framebuffer)                          :    187.5 MB/s (0.2%)
  161.  ARM 2-pass copy (from framebuffer)                   :    174.5 MB/s
  162.  
  163. ==========================================================================
  164. == Memory latency test                                                  ==
  165. ==                                                                      ==
  166. == Average time is measured for random memory accesses in the buffers   ==
  167. == of different sizes. The larger is the buffer, the more significant   ==
  168. == are relative contributions of TLB, L1/L2 cache misses and SDRAM      ==
  169. == accesses. For extremely large buffer sizes we are expecting to see   ==
  170. == page table walk with several requests to SDRAM for almost every      ==
  171. == memory access (though 64MiB is not nearly large enough to experience ==
  172. == this effect to its fullest).                                         ==
  173. ==                                                                      ==
  174. == Note 1: All the numbers are representing extra time, which needs to  ==
  175. ==         be added to L1 cache latency. The cycle timings for L1 cache ==
  176. ==         latency can be usually found in the processor documentation. ==
  177. == Note 2: Dual random read means that we are simultaneously performing ==
  178. ==         two independent memory accesses at a time. In the case if    ==
  179. ==         the memory subsystem can't handle multiple outstanding       ==
  180. ==         requests, dual random read has the same timings as two       ==
  181. ==         single reads performed one after another.                    ==
  182. ==========================================================================
  183.  
  184. block size : single random read / dual random read
  185.       1024 :    0.0 ns          /     0.0 ns
  186.       2048 :    0.0 ns          /     0.0 ns
  187.       4096 :    0.0 ns          /     0.0 ns
  188.       8192 :    0.0 ns          /     0.0 ns
  189.      16384 :    0.0 ns          /     0.0 ns
  190.      32768 :    0.0 ns          /     0.1 ns
  191.      65536 :    5.4 ns          /     9.7 ns
  192.     131072 :    9.4 ns          /    14.0 ns
  193.     262144 :   11.8 ns          /    17.5 ns
  194.     524288 :   21.1 ns          /    29.2 ns
  195.    1048576 :   98.0 ns          /   142.2 ns
  196.    2097152 :  135.5 ns          /   179.6 ns
  197.    4194304 :  156.8 ns          /   191.3 ns
  198.    8388608 :  168.2 ns          /   198.7 ns
  199.   16777216 :  177.1 ns          /   208.3 ns
  200.   33554432 :  186.7 ns          /   222.8 ns
  201.   67108864 :  201.1 ns          /   248.5 ns
  202.  
  203. root@bpi-iot-ros-ai:~# cat /sys/devices/1c62000.dramfreq/devfreq/dramfreq/cur_freq
  204. 576000
RAW Paste Data
Ledger Nano X - The secure hardware wallet
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy. OK, I Understand
Top