Home Forums TrueRNG Hardware random number generator Extremely biased distribution from v2 device

This topic contains 6 replies, has 4 voices, and was last updated by  piersona 9 years, 4 months ago.

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #814

    zsaw
    Member

    I just received my v2 TrueRNG a couple days ago, and I’m running some basic tests, but the output is enormously uneven. I’m on Windows 8, installed the unsigned driver fine, no trouble communicating with it.

    I ran rngcapture a number of times, with similar results as this;

    Byte occurrences from rngcapture.exe 10000000 (10MB)
    Min: 3319
    Max: 157347
    Standard Deviation: 33101

    [0] : 7133
    [1] : 8893
    [2] : 39536
    [3] : 4715
    [4] : 20854
    [5] : 27888
    [6] : 46901
    [7] : 29522
    [8] : 18442
    [9] : 16722
    [10] : 36850
    [11] : 16239
    [12] : 4617
    [13] : 6761
    [14] : 37934
    [15] : 3608
    [16] : 52799
    [17] : 84518
    [18] : 10074
    [19] : 25794
    [20] : 16892
    [21] : 51053
    [22] : 65760
    [23] : 26489
    [24] : 39592
    [25] : 70190
    [26] : 26468
    [27] : 29782
    [28] : 39172
    [29] : 144984
    [30] : 8756
    [31] : 38772
    [32] : 10609
    [33] : 6890
    [34] : 150652
    [35] : 15350
    [36] : 11656
    [37] : 13505
    [38] : 136427
    [39] : 23197
    [40] : 33788
    [41] : 37923
    [42] : 38796
    [43] : 12403
    [44] : 4233
    [45] : 3319
    [46] : 143222
    [47] : 7011
    [48] : 8495
    [49] : 6054
    [50] : 32223
    [51] : 3355
    [52] : 6999
    [53] : 7792
    [54] : 103141
    [55] : 14772
    [56] : 30766
    [57] : 20747
    [58] : 41894
    [59] : 13883
    [60] : 8990
    [61] : 4504
    [62] : 41066
    [63] : 4241
    [64] : 13464
    [65] : 17302
    [66] : 96751
    [67] : 10992
    [68] : 22278
    [69] : 22627
    [70] : 92196
    [71] : 50798
    [72] : 22006
    [73] : 32354
    [74] : 38024
    [75] : 9053
    [76] : 13991
    [77] : 12382
    [78] : 29609
    [79] : 18232
    [80] : 39947
    [81] : 49713
    [82] : 57000
    [83] : 21934
    [84] : 23460
    [85] : 43437
    [86] : 110958
    [87] : 46503
    [88] : 85515
    [89] : 89591
    [90] : 72316
    [91] : 38456
    [92] : 39242
    [93] : 38972
    [94] : 25269
    [95] : 38727
    [96] : 16032
    [97] : 47499
    [98] : 103360
    [99] : 17539
    [100] : 49220
    [101] : 44934
    [102] : 143463
    [103] : 75689
    [104] : 65395
    [105] : 86261
    [106] : 90172
    [107] : 30797
    [108] : 20571
    [109] : 39440
    [110] : 66555
    [111] : 17456
    [112] : 17950
    [113] : 37006
    [114] : 81919
    [115] : 10625
    [116] : 24231
    [117] : 16561
    [118] : 84496
    [119] : 45025
    [120] : 42595
    [121] : 65395
    [122] : 87679
    [123] : 21308
    [124] : 31922
    [125] : 38158
    [126] : 38581
    [127] : 19895
    [128] : 18100
    [129] : 30735
    [130] : 29166
    [131] : 28828
    [132] : 32654
    [133] : 53014
    [134] : 31022
    [135] : 31100
    [136] : 45312
    [137] : 73078
    [138] : 48630
    [139] : 48326
    [140] : 14495
    [141] : 22833
    [142] : 29089
    [143] : 26953
    [144] : 54671
    [145] : 144801
    [146] : 54826
    [147] : 27312
    [148] : 38894
    [149] : 101038
    [150] : 59199
    [151] : 31058
    [152] : 84536
    [153] : 139087
    [154] : 113086
    [155] : 88491
    [156] : 106061
    [157] : 142699
    [158] : 69963
    [159] : 45617
    [160] : 22675
    [161] : 22177
    [162] : 65811
    [163] : 34794
    [164] : 72386
    [165] : 109307
    [166] : 94102
    [167] : 52949
    [168] : 15732
    [169] : 42617
    [170] : 41028
    [171] : 21560
    [172] : 8232
    [173] : 13992
    [174] : 51242
    [175] : 27493
    [176] : 11408
    [177] : 16593
    [178] : 15113
    [179] : 35257
    [180] : 17676
    [181] : 68972
    [182] : 36846
    [183] : 31330
    [184] : 25607
    [185] : 50591
    [186] : 24696
    [187] : 21990
    [188] : 7345
    [189] : 11535
    [190] : 17664
    [191] : 18940
    [192] : 15082
    [193] : 13636
    [194] : 23428
    [195] : 10181
    [196] : 37713
    [197] : 34775
    [198] : 29014
    [199] : 30568
    [200] : 10002
    [201] : 16679
    [202] : 23173
    [203] : 10171
    [204] : 3566
    [205] : 15068
    [206] : 28784
    [207] : 4710
    [208] : 29397
    [209] : 69283
    [210] : 4738
    [211] : 20470
    [212] : 14462
    [213] : 71048
    [214] : 52684
    [215] : 26492
    [216] : 79237
    [217] : 102223
    [218] : 53232
    [219] : 92444
    [220] : 30105
    [221] : 157347
    [222] : 9391
    [223] : 39768
    [224] : 12616
    [225] : 9322
    [226] : 66487
    [227] : 13186
    [228] : 17848
    [229] : 23250
    [230] : 132870
    [231] : 28865
    [232] : 34098
    [233] : 48500
    [234] : 46904
    [235] : 17278
    [236] : 6042
    [237] : 7375
    [238] : 82779
    [239] : 8534
    [240] : 18675
    [241] : 12435
    [242] : 29157
    [243] : 13644
    [244] : 11406
    [245] : 21476
    [246] : 54637
    [247] : 16872
    [248] : 19154
    [249] : 16984
    [250] : 44937
    [251] : 13071
    [252] : 9124
    [253] : 10966
    [254] : 56890
    [255] : 6972

    At first I thought it was my C#/.NET code that was at fault in reading from the virtual port, but then I found the same results from the rngcapture tool.

    The above breakdown was generated using this C# code:

    `var arr = File.ReadAllBytes(@”C:\10milrandom”);
    long[] map = new long[256];
    arr.ForEach(t => map[t]++);
    map.ForEach((d, i) => { File.AppendAllText(@”C:\rngdebug.txt”, “[" + i + "] : ” + d + “\r\n”); });`

    #832

    gee-one
    Member

    I just received a new V2 today, and ran it through some of its paces. I am not getting that much variation in results. I am on linux and pulled 64M from /dev/ttyAMC1 (I have a V1.0 in /dev/ttyAMC0). I used ent to analyze the byte count.

    $ ent -c true2-d.64M
    Value Char Occurrences Fraction
      0           262737   0.003915
      1           261696   0.003900
      2           261971   0.003904
      3           261731   0.003900
      4           262129   0.003906
      5           261891   0.003903
      6           261399   0.003895
      7           262226   0.003908
      8           262009   0.003904
      9           262641   0.003914
     10           262745   0.003915
     11           261910   0.003903
     12           262084   0.003905
     13           261890   0.003903
     14           262589   0.003913
     15           261598   0.003898
     16           261646   0.003899
     17           262201   0.003907
     18           262078   0.003905
     19           261423   0.003896
     20           261994   0.003904
     21           261609   0.003898
     22           261860   0.003902
     23           262776   0.003916
     24           261893   0.003903
     25           262981   0.003919
     26           262114   0.003906
     27           262193   0.003907
     28           262189   0.003907
     29           261808   0.003901
     30           260901   0.003888
     31           261621   0.003899
     32           261955   0.003903
     33   !       262377   0.003910
     34   "       262009   0.003904
     35   #       261406   0.003895
     36   $       262077   0.003905
     37   %       262730   0.003915
     38   &       262175   0.003907
     39   '       262527   0.003912
     40   (       261929   0.003903
     41   )       262344   0.003909
     42   *       260944   0.003888
     43   +       262077   0.003905
     44   ,       262244   0.003908
     45   -       261416   0.003895
     46   .       262570   0.003913
     47   /       261889   0.003903
     48   0       261547   0.003897
     49   1       262411   0.003910
     50   2       262164   0.003907
     51   3       262497   0.003912
     52   4       262601   0.003913
     53   5       261805   0.003901
     54   6       261278   0.003893
     55   7       262042   0.003905
     56   8       262377   0.003910
     57   9       261656   0.003899
     58   :       262322   0.003909
     59   ;       262651   0.003914
     60   <       262000   0.003904
     61   =       262212   0.003907
     62   >       261376   0.003895
     63   ?       262016   0.003904
     64   @       262925   0.003918
     65   A       262124   0.003906
     66   B       260942   0.003888
     67   C       263001   0.003919
     68   D       261776   0.003901
     69   E       261872   0.003902
     70   F       261472   0.003896
     71   G       262514   0.003912
     72   H       261916   0.003903
     73   I       261875   0.003902
     74   J       262615   0.003913
     75   K       262731   0.003915
     76   L       261961   0.003904
     77   M       263507   0.003927
     78   N       261971   0.003904
     79   O       262851   0.003917
     80   P       262297   0.003909
     81   Q       261781   0.003901
     82   R       262564   0.003913
     83   S       262058   0.003905
     84   T       262398   0.003910
     85   U       262133   0.003906
     86   V       261908   0.003903
     87   W       261417   0.003895
     88   X       261944   0.003903
     89   Y       262157   0.003907
     90   Z       262178   0.003907
     91   [       261272   0.003893
     92   \       262893   0.003917
     93   ]       263210   0.003922
     94   ^       262573   0.003913
     95   _       261595   0.003898
     96   **       262579   0.003913
     97   a       262351   0.003909
     98   b       261358   0.003895
     99   c       261700   0.003900
    100   d       262783   0.003916
    101   e       262844   0.003917
    102   f       262715   0.003915
    103   g       261563   0.003898
    104   h       262879   0.003917
    105   i       261373   0.003895
    106   j       261553   0.003898
    107   k       262110   0.003906
    108   l       261740   0.003900
    109   m       261474   0.003896
    110   n       263049   0.003920
    111   o       262302   0.003909
    112   p       262450   0.003911
    113   q       261837   0.003902
    114   r       262831   0.003917
    115   s       262886   0.003917
    116   t       262094   0.003906
    117   u       262207   0.003907
    118   v       262371   0.003910
    119   w       261831   0.003902
    120   x       262677   0.003914
    121   y       262283   0.003908
    122   z       262493   0.003912
    123   {       261532   0.003897
    124   |       262648   0.003914
    125   }       261743   0.003900
    126   ~       262630   0.003914
    127           262019   0.003904
    128           262279   0.003908
    129           263019   0.003919
    130           262821   0.003916
    131           262229   0.003908
    132           262176   0.003907
    133           262731   0.003915
    134           262294   0.003909
    135           262586   0.003913
    136           262403   0.003910
    137           261772   0.003901
    138           262268   0.003908
    139           262758   0.003915
    140           261138   0.003891
    141           262274   0.003908
    142           262259   0.003908
    143           263101   0.003921
    144           262499   0.003912
    145           262186   0.003907
    146           261640   0.003899
    147           261478   0.003896
    148           261500   0.003897
    149           263538   0.003927
    150           262180   0.003907
    151           262001   0.003904
    152           263051   0.003920
    153           262019   0.003904
    154           261841   0.003902
    155           262045   0.003905
    156           261095   0.003891
    157           262021   0.003904
    158           261967   0.003904
    159           262750   0.003915
    160           262739   0.003915
    161   �       262272   0.003908
    162   �       261582   0.003898
    163   �       261556   0.003898
    164   �       261527   0.003897
    165   �       262685   0.003914
    166   �       262730   0.003915
    167   �       261559   0.003898
    168   �       262949   0.003918
    169   �       262548   0.003912
    170   �       261889   0.003903
    171   �       262592   0.003913
    172   �       261857   0.003902
    173   �       262207   0.003907
    174   �       261008   0.003889
    175   �       261374   0.003895
    176   �       263191   0.003922
    177   �       262646   0.003914
    178   �       262025   0.003905
    179   �       261820   0.003901
    180   �       262377   0.003910
    181   �       262618   0.003913
    182   �       261443   0.003896
    183   �       262204   0.003907
    184   �       263069   0.003920
    185   �       261611   0.003898
    186   �       262668   0.003914
    187   �       262143   0.003906
    188   �       262294   0.003909
    189   �       262597   0.003913
    190   �       261958   0.003904
    191   �       261841   0.003902
    192   �       262354   0.003909
    193   �       262027   0.003905
    194   �       263238   0.003923
    195   �       261476   0.003896
    196   �       261458   0.003896
    197   �       261242   0.003893
    198   �       262049   0.003905
    199   �       262509   0.003912
    200   �       262196   0.003907
    201   �       261811   0.003901
    202   �       261771   0.003901
    203   �       262755   0.003915
    204   �       261878   0.003902
    205   �       261321   0.003894
    206   �       261766   0.003901
    207   �       262523   0.003912
    208   �       262727   0.003915
    209   �       261739   0.003900
    210   �       262929   0.003918
    211   �       263601   0.003928
    212   �       262411   0.003910
    213   �       262150   0.003906
    214   �       262585   0.003913
    215   �       262292   0.003909
    216   �       261865   0.003902
    217   �       261199   0.003892
    218   �       261752   0.003900
    219   �       262193   0.003907
    220   �       261539   0.003897
    221   �       262506   0.003912
    222   �       262314   0.003909
    223   �       262204   0.003907
    224   �       262207   0.003907
    225   �       262454   0.003911
    226   �       262118   0.003906
    227   �       263031   0.003920
    228   �       261540   0.003897
    229   �       262592   0.003913
    230   �       261572   0.003898
    231   �       261330   0.003894
    232   �       261458   0.003896
    233   �       261493   0.003897
    234   �       261328   0.003894
    235   �       262348   0.003909
    236   �       262572   0.003913
    237   �       261350   0.003894
    238   �       261854   0.003902
    239   �       262012   0.003904
    240   �       262563   0.003913
    241   �       262070   0.003905
    242   �       262355   0.003909
    243   �       261684   0.003899
    244   �       262401   0.003910
    245   �       262060   0.003905
    246   �       263629   0.003928
    247   �       261980   0.003904
    248   �       263142   0.003921
    249   �       261669   0.003899
    250   �       261664   0.003899
    251   �       262003   0.003904
    252   �       262115   0.003906
    253   �       261797   0.003901
    254   �       261693   0.003900
    255   �       262385   0.003910
    
    Total:      67107840   1.000000
    
    Entropy = 7.999997 bits per byte.
    
    Optimum compression would reduce the size
    of this 67107840 byte file by 0 percent.
    
    Chi square distribution for 67107840 samples is 271.85, and randomly
    would exceed this value 25.00 percent of the times.
    
    Arithmetic mean value of data bytes is 127.5058 (127.5 = random).
    Monte Carlo value for Pi is 3.142486124 (error 0.03 percent).
    Serial correlation coefficient is -0.000081 (totally uncorrelated = 0.0).
    
    • This reply was modified 9 years, 7 months ago by  gee-one.
    • This reply was modified 9 years, 7 months ago by  gee-one.
    #885

    Ubld.it Staff
    Moderator

    @zsaw We have been going back and forth amongst ourselves trying to theorize how this could happen. Those batch of v2′s were tested twice by 2 different people before shipping, and we ran them through ent/fips/histograms and they all checked out.

    I’m going to contact you via email and see if we can gather some other info and possibly do a RMA.

    For the record there is a 1mb test capture on the product page you can also run through your own code to rule out a capture problem (file is here) if you want to run that through your c# code.

    #886

    zsaw
    Member

    I already ran the 1mb capture through my same C# code, and the 1mb file was well distributed, so it doesn’t seem to be my code there.

    If there’s another environment to test in, without setting up a full VM with a unix distro, I’d be happy to run some other tests.
    I don’t know if using USB 3.0 ports or 2.0 port would make a difference, but I’ll check some other ports.

    #887

    Ubld.it Staff
    Moderator

    We have tested USB 3.0 ports and Win8 and haven’t noticed any issues. We’ve sent you an email (check your spam folder if you haven’t gotten it) about proceeding with RMA.

    #996

    piersona
    Member

    I just got mine and saw the same thing:

    0|14:35 ~/truerng$ dd if=/dev/ttyACM1 of=data.2.bin bs=1024 count=$((2**20)) iflag=fullblock
    ^C97419+0 records in
    97419+0 records out
    99757056 bytes (100 MB) copied, 2228.41 s, 44.8 kB/s
    
    130|15:13 ~/truerng$ ent data.2.bin
    Entropy = 7.999998 bits per byte.
    
    Optimum compression would reduce the size
    of this 99757056 byte file by 0 percent.
    
    Chi square distribution for 99757056 samples is 270.00, and randomly
    would exceed this value 25.00 percent of the times.
    
    Arithmetic mean value of data bytes is 127.5062 (127.5 = random).
    Monte Carlo value for Pi is 3.141212507 (error 0.01 percent).
    Serial correlation coefficient is 0.000179 (totally uncorrelated = 0.0).

    I did not get the same problem when I repeated with a smaller experiment without fullblock:

    0|15:13 ~/truerng$ dd if=/dev/ttyACM1 of=data.2.bin bs=1024 count=$((2**15))
    dd: warning: partial read (469 bytes); suggest iflag=fullblock
    3+32765 records in
    3+32765 records out
    2068055 bytes (2.1 MB) copied, 45.1346 s, 45.8 kB/s
    0|15:15 ~/truerng$ ent data.2.bin
    Entropy = 7.999914 bits per byte.
    
    Optimum compression would reduce the size
    of this 2068055 byte file by 0 percent.
    
    Chi square distribution for 2068055 samples is 246.19, and randomly
    would exceed this value 50.00 percent of the times.
    
    Arithmetic mean value of data bytes is 127.5313 (127.5 = random).
    Monte Carlo value for Pi is 3.140837020 (error 0.02 percent).
    Serial correlation coefficient is -0.001316 (totally uncorrelated = 0.0).

    I’m running two devices simultaneously, as the previous poster was, as well. /dev/ttyACM1 is the misbehaving unit. I’m currently running the following test script and will update once it’s finished:

    #!/bin/bash
    
    if test "${1}" == ""
    then
        echo "You must specify the device"
        exit 1
    fi
    
    device=/dev/ttyACM${1}
    logfile=test.${1}.log
    
    for i in 1 2 3 4 5 6 7 8 9 0
    do
        for bs in 3 4 5 6 7 8 9 10
        do
            bbs=$((2**${bs}))
            bcount=$((2**(24-${bs})))
            echo '==========================='                                            >> ${logfile} 2>&1
            filename=test.${1}.bs$((2**${bs})).fb0.${i}.bin
            echo ${filename}                                                              >> ${logfile} 2>&1
            time dd if=${device} of=${filename} bs=${bbs} count=${bcount}                 >> ${logfile} 2>&1
            ent ${filename}                                                               >> ${logfile} 2>&1
    
            filename=test.${1}.bs$((2**${bs})).fb1.${i}.bin
            echo ${filename}                                                              >> ${logfile} 2>&1
            time dd if=${device} of=${filename} bs=${bbs} count=${bcount} iflag=fullblock >> ${logfile} 2>&1
            ent ${filename}                                                               >> ${logfile} 2>&1
        done
    done
    

    Additionally, I was playing with the rngd in the background, so potentially running that at the same time could have been causing problems. I’ll run a second test with that running once this finishes.

    –Andrew

    • This reply was modified 9 years, 4 months ago by  piersona.
    • This reply was modified 9 years, 4 months ago by  piersona.
    • This reply was modified 9 years, 4 months ago by  piersona.
    #1000

    piersona
    Member

    I’ve learned two things:
    1. A chi squared result that “would exceed this value N percent of the times.” is not a failure for N != 50.00. See http://www.fourmilab.ch/random/

    2. The values I’m seeing for N are never as arbitrary as that website would lead me to believe, e.g. “40.98″. Instead, I’m seeing values of 25.00, 50.00, 75.00, 90.00, and 95.00.

    Even for the 95.00% case, the data examined with “ent -c” looks well distributed. I think they are performing well; it’s just my misunderstanding.

Viewing 7 posts - 1 through 7 (of 7 total)

You must be logged in to reply to this topic.