6502.org Forum  Projects  Code  Documents  Tools  Forum
It is currently Fri Oct 18, 2024 9:27 pm

All times are UTC




Post new topic Reply to topic  [ 245 posts ]  Go to page Previous  1 ... 7, 8, 9, 10, 11, 12, 13 ... 17  Next
Author Message
PostPosted: Fri Apr 14, 2023 6:27 pm 
Offline

Joined: Mon Mar 13, 2023 9:38 pm
Posts: 80
As I said in a previous message, LFSR are building blocks. So are LCG.
The miracle is that if we mix the output of an LFSR and the output of an LCG, it passes ALL statistical tests : even PractRand. And mathematically we know that it will give 2^64 random bits.

First a 8-bytes seed must be put at seed:
The seed can vary from 0 to $FFFFFFFFFFFFFFFF.

And a 8-bytes seed must be put at seedlfsr:
The seed can vary from 1 to $FFFFFFFFFFFFFFFF but never 0 otherwise the LFSR produces only bits to 0.

Each call to the begin: function returns a pseudorandom byte in rndbyte:

First init: must be call one time.
After begin: must be call for each generated byte (rndbyte:) that can be use directly.

The test vector is :

For a seedlfsr = $FFFFFFFFFFFFFFFF
adn seed = $FFFFFFFFFFFFFFFF

the first 50 bytes are :

Quote:
72 BB 94 D2 AD 1D 2D 21 97 4B DF 5A 06 A5 38 C1 67 75 5D F6 29 DC 95 15 77 C5 38 0E 3B B3 20 1C ED EA 6A 76 79 AB 73 97 18 FA F8 A5 AD 0F E2 26 EC 74


Code:
//           Pukall LFSR 64 + Pukall LCG 64

//----------------------------------------------------
//          Main Program
//----------------------------------------------------

init:
    jsr initlfsr
    jsr initlcg
rts

begin:
    jsr startlfsr
    jsr beginlcg
    lda rndbytelfsr
    clc
    adc rndbyte
    sta rndbyte
    rts

initlfsr:   
    ldx #0
   
looplfsr:   
    jsr startlfsr
   inx
   cpx #64
   bne looplfsr
   rts
   
startlfsr:
        lda #$00
        sta rndbytelfsr
        lda #$01
        sta mask
   
beginlfsr:
        asl seedlfsr
        rol seedlfsr+1
        rol seedlfsr+2
        rol seedlfsr+3
        rol seedlfsr+4
        rol seedlfsr+5
        rol seedlfsr+6
        rol seedlfsr+7
        bcc notlfsr
        lda seedlfsr
        eor #$01
        sta seedlfsr
        lda seedlfsr+7
        eor #$B0
        sta seedlfsr+7

   lda rndbytelfsr
        ora mask
        sta rndbytelfsr
notlfsr:
   asl mask
        bcc beginlfsr
        rts

initlcg:
    lda     #6
    sta     pos
    rts
beginlcg:
    ldx     pos
    cpx     #6
    bne     rnd
    jsr     start
    ldx     #0
    clc
    php
loop0:
    lda     seed,x
    plp
    adc     dda,x
    php
    sta     seed,x
    inx
    cpx     #8
    bne     loop0
    plp
rnd:
    ldx     pos
    lda     seed,x
    sta     rndbyte
    dex
    cpx     #2
    bne     end2
end:
    lda     #6
    sta     pos
    rts
end2:
    stx     pos
    rts
start:
    lda     #0
    ldx     #15
loop:
    sta     result,x
    dex
    cpx    #7
    bne     loop
    ldy     #64       
part1:
    sty     storey
    ldx     #7
    lsr     seed,x
    php
    dex
loop1:
    plp
    ror     seed,x
    php
    dex
    cpx     #0
    bne     loop1
    plp
    ror     seed
    bcc     part2
    ldx     #7
    ldy     #0
    clc
    php
loop2:
    inx
    lda     result,x
    plp
    adc     two,y   
    php
    sta     result,x
    iny
    cpy     #8
    bne     loop2
    plp
part2:
    ror     
    php
    ldx     #15         
    sta     result,x
loop3:
    dex
    plp
    ror     result,x
    php
    cpx     #0
    bne     loop3
    plp
    ldy     storey
    dey             
    bne     jump
    ldx     #0
loop4:
    lda     result,x
    sta     seed,x
    inx
    cpx    #8
    bne     loop4
    rts
jump:
    jmp part1

seed: .byte $FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF
rndbyte: .byte $FF
result: .byte $FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF
two: .byte $2D,$7F,$95,$4C,$2D,$F4,$51,$58
dda: .byte $01,$00,$00,$00,$00,$00,$00,$00
storey: .byte $FF
pos: .byte $FF
seedlfsr: .byte $FF,$FF,$FF,$FF,$FF,$FF,$FF,$FF
mask:  .byte $00
rndbytelfsr: .byte $00


PRACTRAND tests :

Code:
RNG_test using PractRand version 0.94
RNG = RNG_stdin8, seed = unknown
test set = core, folding = standard (8 bit)

rng=RNG_stdin8, seed=unknown
length= 8 megabytes (2^23 bytes), time= 3.7 seconds
  no anomalies in 76 test result(s)

rng=RNG_stdin8, seed=unknown
length= 16 megabytes (2^24 bytes), time= 8.4 seconds
  no anomalies in 84 test result(s)

rng=RNG_stdin8, seed=unknown
length= 32 megabytes (2^25 bytes), time= 17.0 seconds
  no anomalies in 93 test result(s)

rng=RNG_stdin8, seed=unknown
length= 64 megabytes (2^26 bytes), time= 32.5 seconds
  no anomalies in 99 test result(s)

rng=RNG_stdin8, seed=unknown
length= 128 megabytes (2^27 bytes), time= 62.7 seconds
  no anomalies in 108 test result(s)

rng=RNG_stdin8, seed=unknown
length= 256 megabytes (2^28 bytes), time= 121 seconds
  no anomalies in 118 test result(s)

rng=RNG_stdin8, seed=unknown
length= 512 megabytes (2^29 bytes), time= 237 seconds
  no anomalies in 126 test result(s)

rng=RNG_stdin8, seed=unknown
length= 1 gigabyte (2^30 bytes), time= 470 seconds
  no anomalies in 135 test result(s)

rng=RNG_stdin8, seed=unknown
length= 2 gigabytes (2^31 bytes), time= 936 seconds
  no anomalies in 144 test result(s)

rng=RNG_stdin8, seed=unknown
length= 4 gigabytes (2^32 bytes), time= 1846 seconds
  no anomalies in 151 test result(s)

rng=RNG_stdin8, seed=unknown
length= 8 gigabytes (2^33 bytes), time= 3662 seconds
  no anomalies in 159 test result(s)

rng=RNG_stdin8, seed=unknown
length= 16 gigabytes (2^34 bytes), time= 7278 seconds
  no anomalies in 167 test result(s)

rng=RNG_stdin8, seed=unknown
length= 32 gigabytes (2^35 bytes), time= 14372 seconds
  no anomalies in 173 test result(s)

rng=RNG_stdin8, seed=unknown
length= 64 gigabytes (2^36 bytes), time= 28416 seconds
  no anomalies in 181 test result(s)

rng=RNG_stdin8, seed=unknown
length= 128 gigabytes (2^37 bytes), time= 55896 seconds
  no anomalies in 189 test result(s)

rng=RNG_stdin8, seed=unknown
length= 256 gigabytes (2^38 bytes), time= 110189 seconds
  no anomalies in 195 test result(s)

rng=RNG_stdin8, seed=unknown
length= 512 gigabytes (2^39 bytes), time= 208856 seconds
  no anomalies in 202 test result(s)

rng=RNG_stdin8, seed=unknown
length= 1 terabyte (2^40 bytes), time= 369656 seconds
  no anomalies in 208 test result(s)

rng=RNG_stdin8, seed=unknown
length= 2 terabytes (2^41 bytes), time= 686317 seconds
  no anomalies in 213 test result(s)


Top
 Profile  
Reply with quote  
PostPosted: Fri Apr 14, 2023 6:48 pm 
Offline
User avatar

Joined: Tue Nov 16, 2010 8:00 am
Posts: 2353
Location: Gouda, The Netherlands
How many cycles/byte, though? The big challenge is to produce good randomness with minimal cycle count.


Top
 Profile  
Reply with quote  
PostPosted: Fri Apr 14, 2023 8:06 pm 
Offline

Joined: Mon Mar 13, 2023 9:38 pm
Posts: 80
Arlet wrote:
How many cycles/byte, though? The big challenge is to produce good randomness with minimal cycle count.


Yes Arlet, for the number of cycles you win.


Top
 Profile  
Reply with quote  
PostPosted: Fri Apr 14, 2023 8:36 pm 
Offline

Joined: Mon Mar 13, 2023 9:38 pm
Posts: 80
@Arlet, now that I've tested my LFSR+LCG mixed algorithm, I'd like to test both of yours. Would you have the C code of your two PRNGs ?

https://github.com/Arlet/pseudo-random- ... l-6502.asm
https://github.com/Arlet/pseudo-random- ... g-6502.asm


Top
 Profile  
Reply with quote  
PostPosted: Sat Apr 15, 2023 6:16 am 
Offline
User avatar

Joined: Tue Nov 16, 2010 8:00 am
Posts: 2353
Location: Gouda, The Netherlands
I lost the C code for the big one, but I just added the C code for the minimal version to the github repository.


Top
 Profile  
Reply with quote  
PostPosted: Sat Apr 15, 2023 2:26 pm 
Offline

Joined: Mon Mar 13, 2023 9:38 pm
Posts: 80
Arlet wrote:
I lost the C code for the big one, but I just added the C code for the minimal version to the github repository.

I notice that there is something strange about your minimal code. With some seeds, Practrand detects unusual data. While in my LFSR + LCG algorithm there is no unusual up to 2TB whatever seed one chooses. This shows that your algorithm is not so random, if depending on the seeds it reacts differently.
Code:
 uint8_t s0 = 255, s1 = 255, s2 = 255, s3 = 255, s4 = 255, s5 = 255;

rng=RNG_stdin8, seed=unknown
length= 512 megabytes (2^29 bytes), time= 158 seconds
  Test Name                         Raw       Processed     Evaluation
  [Low1/8]DC6-9x1Bytes-1            R=  +5.5  p =  7.5e-3   unusual         
  ...and 125 test result(s) without anomalies




Top
 Profile  
Reply with quote  
PostPosted: Sun Apr 16, 2023 1:26 am 
Offline

Joined: Mon Mar 13, 2023 9:38 pm
Posts: 80
Arlet, did you test your algorithms on Windows ?

You can download the PractRand Windows version here :

https://sourceforge.net/projects/pracrand/files/

In PractRand_094\Bin you have the 32-bit .exe version or 64-bit .exe version.

With the Windows version your minimal algorithm fails at 1Mb :

Code:
minimal.exe | RNG_test32.exe stdin8 -tlmin 1MB

rng=RNG_stdin8, seed=unknown
length= 1 megabyte (2^20 bytes), time= 0.1 seconds
  Test Name                         Raw       Processed     Evaluation
  DC6-9x1Bytes-1                    R=  +9.3  p =  4.3e-5   suspicious
  Gap-16:A                          R=  +7.3  p =  1.1e-5   suspicious
  Gap-16:B                          R= +32.1  p =  1.4e-24    FAIL !!
  FPF-14+6/16:(0,14-4)              R=+179.6  p =  1.1e-146   FAIL !!!!!
  FPF-14+6/16:all                   R=+114.2  p =  1.9e-96    FAIL !!!!!
  [Low1/8]BCFN(2+0,13-9,T)          R= +14.7  p =  1.9e-4   mildly suspicious
  [Low1/8]BCFN(2+1,13-9,T)          R= +27.8  p =  2.3e-7   very suspicious
  [Low1/8]BCFN(2+2,13-9,T)          R= +22.6  p =  3.4e-6   suspicious
  ...and 48 test result(s) without anomalies


or at 16 MB :

Code:
minimal.exe | RNG_test32.exe stdin8

RNG_test using PractRand version 0.94
RNG = RNG_stdin8, seed = unknown
test set = core, folding = standard (8 bit)

rng=RNG_stdin8, seed=unknown
length= 32 megabytes (2^25 bytes), time= 2.4 seconds
  Test Name                         Raw       Processed     Evaluation
  BCFN(2+0,13-4,T)                  R= +96.2  p =  5.7e-42    FAIL !!!
  BCFN(2+1,13-4,T)                  R= +93.6  p =  7.9e-41    FAIL !!!
  BCFN(2+2,13-5,T)                  R=+104.9  p =  2.9e-41    FAIL !!!
  BCFN(2+3,13-5,T)                  R=+105.1  p =  2.5e-41    FAIL !!!
  BCFN(2+4,13-5,T)                  R= +88.2  p =  1.0e-34    FAIL !!!
  BCFN(2+5,13-6,T)                  R=+103.5  p =  6.9e-36    FAIL !!!
  BCFN(2+6,13-6,T)                  R=+106.3  p =  7.8e-37    FAIL !!!
  BCFN(2+7,13-7,T)                  R=+114.8  p =  3.7e-35    FAIL !!!
  BCFN(2+8,13-8,T)                  R=+125.3  p =  9.9e-33    FAIL !!!
  BCFN(2+9,13-8,T)                  R=+126.4  p =  5.1e-33    FAIL !!!
  BCFN(2+10,13-9,T)                 R=+177.0  p =  6.9e-41    FAIL !!!
  BCFN(2+11,13-9,T)                 R=+172.2  p =  8.1e-40    FAIL !!!
  DC6-9x1Bytes-1                    R= +89.2  p =  3.1e-59    FAIL !!!!
  Gap-16:A                          R= +84.6  p =  6.0e-69    FAIL !!!!
  Gap-16:B                          R=+519.8  p =  1.6e-427   FAIL !!!!!!!
  FPF-14+6/16:(0,14-0)              R= +4220  p =  8e-3887    FAIL !!!!!!!!
  FPF-14+6/16:(1,14-1)              R=+190.4  p =  1.8e-168   FAIL !!!!!
  FPF-14+6/16:(2,14-2)              R=+133.6  p =  1.2e-116   FAIL !!!!!
  FPF-14+6/16:(3,14-2)              R=+174.7  p =  1.3e-152   FAIL !!!!!
  FPF-14+6/16:(4,14-3)              R=+122.6  p =  2.9e-107   FAIL !!!!!
  FPF-14+6/16:(5,14-4)              R= +87.3  p =  3.1e-71    FAIL !!!!
  FPF-14+6/16:(6,14-5)              R= +63.0  p =  4.4e-52    FAIL !!!!
  FPF-14+6/16:(7,14-5)              R= +38.4  p =  1.0e-31    FAIL !!!
  FPF-14+6/16:(8,14-6)              R= +17.2  p =  3.5e-13    FAIL
  FPF-14+6/16:(9,14-7)              R= +44.8  p =  1.7e-35    FAIL !!!
  FPF-14+6/16:all                   R= +3043  p =  2e-2855    FAIL !!!!!!!!
  FPF-14+6/16:cross                 R= +46.7  p =  3.0e-37    FAIL !!!
  mod3n(5):(0,9-2)                  R= +19.6  p =  8.3e-11   VERY SUSPICIOUS
  [Low1/8]BCFN(2+0,13-6,T)          R=+215.7  p =  2.5e-74    FAIL !!!!
  [Low1/8]BCFN(2+1,13-6,T)          R=+207.2  p =  2.2e-71    FAIL !!!!
  [Low1/8]BCFN(2+2,13-6,T)          R=+194.1  p =  6.8e-67    FAIL !!!!
  [Low1/8]BCFN(2+3,13-6,T)          R=+205.5  p =  7.9e-71    FAIL !!!!
  [Low1/8]BCFN(2+4,13-7,T)          R=+252.1  p =  1.8e-76    FAIL !!!!
  [Low1/8]BCFN(2+5,13-8,T)          R=+325.0  p =  2.0e-83    FAIL !!!!
  [Low1/8]BCFN(2+6,13-8,T)          R=+316.7  p =  2.5e-81    FAIL !!!!
  [Low1/8]BCFN(2+7,13-9,T)          R=+329.0  p =  4.7e-75    FAIL !!!!
  [Low1/8]BCFN(2+8,13-9,T)          R=+342.0  p =  5.8e-78    FAIL !!!!
  [Low1/8]DC6-9x1Bytes-1            R= +25.7  p =  2.9e-14    FAIL !
  [Low1/8]FPF-14+6/16:all           R=  +6.4  p =  2.2e-5   mildly suspicious
  [Low1/8]FPF-14+6/16:cross         R= +32.8  p =  2.1e-28    FAIL !!!
  ...and 53 test result(s) without anomalies


Top
 Profile  
Reply with quote  
PostPosted: Sun Apr 16, 2023 4:44 am 
Offline
User avatar

Joined: Tue Nov 16, 2010 8:00 am
Posts: 2353
Location: Gouda, The Netherlands
I saw your post on sourceforge. Did you see the reply?
Quote:
actually, are you sure your RNG output isn't getting mangled by CR/LF issues somehow? That's what that looks like to me.

this is indeed the problem. Every '0a' byte gets translated to '0d 0a':
Code:
% wine minimal.exe | hexdump -C | grep 0a | head -5
000000a0  a4 ab bc d7 2b d7 ce 55  6a d8 86 12 b6 44 46 55  |....+..Uj....DFU|
00000160  24 16 74 41 72 a5 1b 83  38 db 0d 0a 5e 0e 59 2f  |$.tAr...8...^.Y/|
00000210  96 0d 0a c5 32 2c 5a 2f  41 c5 3d c3 1b d7 5c 98  |....2,Z/A.=...\.|
00000250  37 6a 94 69 10 b7 b8 8d  07 0d 0a 67 dc bd 85 97  |7j.i.......g....|
000002f0  b6 7c 46 52 e3 d9 0d 0a  64 b9 4d c8 c9 18 09 5e  |.|FR....d.M....^|

I have added a windows fix to the source code in github.


Last edited by Arlet on Sun Apr 16, 2023 5:42 am, edited 3 times in total.

Top
 Profile  
Reply with quote  
PostPosted: Sun Apr 16, 2023 4:48 am 
Offline
User avatar

Joined: Tue Nov 16, 2010 8:00 am
Posts: 2353
Location: Gouda, The Netherlands
Alex1 wrote:
I notice that there is something strange about your minimal code. With some seeds, Practrand detects unusual data.


You should let it run longer. "unusual" is not a failure. It could simply be a case of some bad luck. You often see "unusual" pop up, and then it goes away with more data. If it's a real flaw, you should see it get worse with more data, and then fail.


Top
 Profile  
Reply with quote  
PostPosted: Sun Apr 16, 2023 1:40 pm 
Offline

Joined: Mon Mar 13, 2023 9:38 pm
Posts: 80
Arlet wrote:
Alex1 wrote:
I notice that there is something strange about your minimal code. With some seeds, Practrand detects unusual data.


You should let it run longer. "unusual" is not a failure. It could simply be a case of some bad luck. You often see "unusual" pop up, and then it goes away with more data. If it's a real flaw, you should see it get worse with more data, and then fail.



Thanks for finding the problem on Windows.
For Sourceforge why did he says "At a quick glance, your ADC implementation is bugged, though I don't quite see how that bug could lead to the results you see. "
For 'unuual' how do you explain that on my LFSR+LCG algorithm there is absolutely no 'unusual' even up to 2TB? This implies that its output is better, right?


Top
 Profile  
Reply with quote  
PostPosted: Sun Apr 16, 2023 2:05 pm 
Offline
User avatar

Joined: Tue Nov 16, 2010 8:00 am
Posts: 2353
Location: Gouda, The Netherlands
Alex1 wrote:
For 'unuual' how do you explain that on my LFSR+LCG algorithm there is absolutely no 'unusual' even up to 2TB? This implies that its output is better, right?


Just luck. For example, a p-value of 1e-3 means there's a 0.1% chance the outcome could have happened purely by chance. Keep in mind that each run of PractRand does hundreds of tests in parallel, so something with a probability of 0.1% can happen quite often.

This is from a test I'm running right now:
Code:
rng=RNG_stdin, seed=unknown
length= 4 gigabytes (2^32 bytes), time= 35.0 seconds
  Test Name                         Raw       Processed     Evaluation
  [Low4/32]BCFN(2+2,13-2,T)         R=  +9.1  p =  2.9e-4   unusual         
  ...and 282 test result(s) without anomalies

rng=RNG_stdin, seed=unknown
length= 8 gigabytes (2^33 bytes), time= 67.6 seconds
  no anomalies in 300 test result(s)

You can see it says "unusual" after 4GB, but then everything is normal after 8GB. Imagine tossing a coin 200 times, and after 100 times you have 55 heads and 45 tails, and you suspect something is wrong with the coin, but then you throw 100 more times, and you get 46 heads and 54 tails. The more data you have, the more reliable the verdict is. In case of doubt, just keep running the test with more data.

Also keep in mind that my "minimal" implementation is just that. It is primarily designed to be fast and small, and sufficiently good for applications where someone might use a 6502 for. It is highly unlikely that someone is going to use a 6502 to generate billions (or even millions) of random bytes with super strict randomness requirements. And if you make the code really high quality, it's going to take too long anyway. Imagine you have a subroutine that takes 1 millisecond to generate a perfect random byte. Generating 1 TB of output using that code would take 30 years. Ain't nobody got time for that.


Top
 Profile  
Reply with quote  
PostPosted: Sun Apr 16, 2023 4:04 pm 
Offline

Joined: Mon Mar 13, 2023 9:38 pm
Posts: 80
Arlet, yes, I understand. But my question is why doesn't my algorithm have this problem?
It is based on strong mathematics: LFSR and LCG and it never produces these problems while yours does.
Then, if you test your big algorithm, chances are it will produce the same problems because it's not based on strong maths.


Top
 Profile  
Reply with quote  
PostPosted: Sun Apr 16, 2023 4:09 pm 
Offline
User avatar

Joined: Tue Nov 16, 2010 8:00 am
Posts: 2353
Location: Gouda, The Netherlands
Quote:
But my question is why doesn't my algorithm have this problem?

Have you tried running it multiple times with different seeds? I can cause the same problem with chacha(8) when trying different seeds:

Code:
% RNG_test 'chacha(8)' -seed 3 -tlmax 512M
RNG_test using PractRand version 0.95
RNG = chacha(8), seed = 0x3
test set = core, folding = standard (32 bit)

rng=chacha(8), seed=0x3
length= 256 megabytes (2^28 bytes), time= 2.7 seconds
  no anomalies in 168 test result(s)

rng=chacha(8), seed=0x3
length= 512 megabytes (2^29 bytes), time= 6.4 seconds
  Test Name                         Raw       Processed     Evaluation
  [Low1/32]Gap-16:A                 R=  +5.2  p =  9.9e-4   unusual         
  ...and 179 test result(s) without anomalies


Top
 Profile  
Reply with quote  
PostPosted: Sun Apr 16, 2023 4:12 pm 
Offline
User avatar

Joined: Tue Nov 16, 2010 8:00 am
Posts: 2353
Location: Gouda, The Netherlands
And also with chacha20:
Code:
% RNG_test 'chacha(20)' -seed 11 -tlmax 512M
RNG_test using PractRand version 0.95
RNG = chacha(20), seed = 0x11
test set = core, folding = standard (32 bit)

rng=chacha(20), seed=0x11
length= 256 megabytes (2^28 bytes), time= 2.9 seconds
  no anomalies in 168 test result(s)

rng=chacha(20), seed=0x11
length= 512 megabytes (2^29 bytes), time= 6.4 seconds
  Test Name                         Raw       Processed     Evaluation
  [Low8/32]Gap-16:A                 R=  -4.3  p =1-4.6e-4   unusual         
  ...and 179 test result(s) without anomalies

Keep in mind that the chacha algorithm isn't based on math either. It's a bunch of random operations that were picked because they seem to be efficient at mixing bits, and then repeated enough times to make it completely untractable.


Top
 Profile  
Reply with quote  
PostPosted: Sun Apr 16, 2023 8:17 pm 
Offline

Joined: Mon Mar 13, 2023 9:38 pm
Posts: 80
Yes i tried my algorithm with different seeds and no 'unusual' detected. Now it is at 4 TB and still no unusual.

Yes i know for chacha and salsa and hc256 and trivium too. All algorithms that produces random mix gives "unusual".

Code:
RNG_test using PractRand version 0.94
RNG = chacha(20), seed = 0xc4cbc0f4
test set = core, folding = standard (32 bit)

rng=chacha(20), seed=0xc4cbc0f4
length= 16 megabytes (2^24 bytes), time= 2.7 seconds
  Test Name                         Raw       Processed     Evaluation
  [Low1/32]DC6-9x1Bytes-1           R=  -4.7  p =1-4.7e-3   unusual         
  ...and 117 test result(s) without anomalies

rng=salsa(20), seed=0x75364184
length= 64 megabytes (2^26 bytes), time= 28.2 seconds
  Test Name                         Raw       Processed     Evaluation
  [Low1/32]BCFN(2+0,13-6,T)         R= +11.6  p =  2.1e-4   unusual         
  ...and 139 test result(s) without anomalies

rng=trivium, seed=0xda1ca1c3
length= 64 megabytes (2^26 bytes), time= 14.2 seconds
  Test Name                         Raw       Processed     Evaluation
  [Low1/64]DC6-9x1Bytes-1           R=  -5.0  p =1-2.0e-3   unusual         
  ...and 176 test result(s) without anomalies

rng=hc256, seed=0x96d2e54
length= 128 megabytes (2^27 bytes), time= 25.4 seconds
  Test Name                         Raw       Processed     Evaluation
  [Low8/32]DC6-9x1Bytes-1           R=  -4.6  p =1-2.7e-3   unusual         
  ...and 153 test result(s) without anomalies


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 245 posts ]  Go to page Previous  1 ... 7, 8, 9, 10, 11, 12, 13 ... 17  Next

All times are UTC


Who is online

Users browsing this forum: No registered users and 26 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to: