john, the real benchmarking tool

itt, a hupon egy masik szalon vetodott fel, hogy lehet-e egy ilyen geppel jelszot torni.

Nezzuk, az otthoni p4 3GHz masinat: 


Benchmarking: Traditional DES [64/64 BS MMX]... DONE
Many salts:     658086 c/s real, 663393 c/s virtual
Only one salt:  592832 c/s real, 597612 c/s virtual

Benchmarking: BSDI DES (x725) [64/64 BS MMX]... DONE
Many salts:     22067 c/s real, 22200 c/s virtual
Only one salt:  21729 c/s real, 21948 c/s virtual

Benchmarking: FreeBSD MD5 [32/32]... DONE
Raw:    9048 c/s real, 9102 c/s virtual

Benchmarking: OpenBSD Blowfish (x32) [32/32]... DONE
Raw:    448 c/s real, 451 c/s virtual

Benchmarking: Kerberos AFS DES [48/64 4K MMX]... DONE
Short:  197888 c/s real, 199483 c/s virtual
Long:   551321 c/s real, 555767 c/s virtual

Benchmarking: NT LM DES [64/64 BS MMX]... DONE
Raw:    6202K c/s real, 6252K c/s virtual

Benchmarking: NT MD4 [Generic 1x]... DONE
Raw:    10907K c/s real, 10995K c/s virtual

Benchmarking: M$ Cache Hash [Generic 1x]... DONE
Many salts:     19225K c/s real, 19341K c/s virtual
Only one salt:  6643K c/s real, 6683K c/s virtual

Benchmarking: LM C/R DES [netlm]... DONE
Many salts:     469186 c/s real, 472970 c/s virtual
Only one salt:  457504 c/s real, 461194 c/s virtual

Benchmarking: NTLMv1 C/R MD4 DES [netntlm]... DONE
Many salts:     601669 c/s real, 606521 c/s virtual
Only one salt:  563357 c/s real, 566757 c/s virtual

Es a HPC2009 fejgep (intel xeon E5520) egy magjat: 


Benchmarking: Traditional DES [64/64 BS]... DONE
Many salts:     1246K c/s real, 1246K c/s virtual
Only one salt:  1153K c/s real, 1153K c/s virtual

Benchmarking: BSDI DES (x725) [64/64 BS]... DONE
Many salts:     42339 c/s real, 42339 c/s virtual
Only one salt:  41799 c/s real, 41865 c/s virtual

Benchmarking: FreeBSD MD5 [32/64 X2]... DONE
Raw:    10604 c/s real, 10604 c/s virtual

Benchmarking: OpenBSD Blowfish (x32) [32/64 X2]... DONE
Raw:    624 c/s real, 624 c/s virtual

Benchmarking: Kerberos AFS DES [48/64 4K]... DONE
Short:  340958 c/s real, 340958 c/s virtual
Long:   1093K c/s real, 1092K c/s virtual

Benchmarking: LM DES [64/64 BS]... DONE
Raw:    12135K c/s real, 12135K c/s virtual

Benchmarking: NT MD4 [Generic 1x]... DONE
Raw:    10076K c/s real, 10076K c/s virtual

Benchmarking: Eggdrop [blowfish]... DONE
Raw:    19078 c/s real, 19078 c/s virtual

Benchmarking: M$ Cache Hash [Generic 1x]... DONE
Many salts:     13612K c/s real, 13626K c/s virtual
Only one salt:  5464K c/s real, 5455K c/s virtual

Benchmarking: LM C/R DES [netlm]... DONE
Many salts:     476625 c/s real, 476625 c/s virtual
Only one salt:  459901 c/s real, 459901 c/s virtual

Benchmarking: NTLMv1 C/R MD4 DES [netntlm]... DONE
Many salts:     683111 c/s real, 683111 c/s virtual
Only one salt:  651289 c/s real, 651289 c/s virtual

hm, hm, impozans, impozans. Hogyan is lehet teriteni 360 magra?

( mauzi v | 2010. 04. 26., h – 21:56 )

azért a blowfish úgymond "meggyőzően lassabb" a többinél :)

( mauzi v | 2010. 04. 26., h – 22:01 )

a "FreeBSD MD5" címszó alatt ugye a mindenféle oprendszerek "korszerű", MD5-alapú crpyt() kimenetét kell érteni?

( egeresz | 2010. 04. 29., cs – 14:12 )

hm, 


mpirun -hostfile /root/openmpi/hostfile /opt/j
Benchmarking: Traditional DES [128/128 BS SSE2-16]... DONE
Many salts:     883906K c/s real, 1007235K c/s virtual

Only one salt:  774239K c/s real, 854027K c/s virtual

Benchmarking: BSDI DES (x725) [128/128 BS SSE2-16]... DONE
Many salts:     29215K c/s real, 32217K c/s virtual

Only one salt:  28506K c/s real, 31441K c/s virtual

Benchmarking: FreeBSD MD5 [32/64 X2]... DONE
Raw:    3997K c/s real, 4403K c/s virtual


Benchmarking: OpenBSD Blowfish (x32) [32/64 X2]... DONE
Raw:    284083 c/s real, 314309 c/s virtual


Benchmarking: Kerberos AFS DES [48/64 4K]... DONE
Short:  156088K c/s real, 172179K c/s virtual

Long:   495935K c/s real, 544278K c/s virtual

Benchmarking: LM DES [128/128 BS SSE2-16]... DONE
Raw:    6264438K c/s real, 6852139K c/s virtual

( egeresz | 2010. 05. 04., k – 14:58 )

eddigi eredmenyek ;-/
john-mpi nem alkalmas tobb gepre. Egy gepen terit szepen.

most akkor megnezem a djohnt, hata ossze tudom barkacsolni a kettot

"bizonyos kisebb" cluser config problemakra derult feny es lett orvosolva, flock(), nfs3/nfs4, ntp, rsyslog, rsh, openmpi temakorben

meg egy kernelhiba workaround (2.6.32.20 -> 2.6.33.3)
ilyesmik

vegul a djohn csekely modositassal futtatja 704 szalon a john. Mondjuk ma 12:30 korul indult. egy napot hagyom futni.

egyebkent az ubuntu 10.4 amd64 architekturan a gyari john program 'generic' optimalicacioju (32 bites egeszeket hasznal, es mas semmit) mig van a john-nak 64 bit/sse2 optimalizacioja is, az egy cseppet (ketszer) gyorsabb.

( egeresz | 2010. 05. 06., cs – 16:49 )

dual xeon E5520 eseten
16 vcore (HT on) 1millio probalkozas 3 perc 55 sec
8 core (HT off) 1millio probalkozas 2 perc 13 sec

ami azt jelenti, hogy john futtatasakro a gep teljesitmenye 12%-13%
novekmenyt lehet elerni a HT (Hyper Threading) technologiaval

merthogy HT on eseteben 16 milliot vegez el 3 perc 55 sec alatt,
HT off eseteben 8 milliot vegez el 2 perc 13 sec alatt.