hashcat Forum
Why Speed difference so big between brute and dictionary for (wpa 22000) - Printable Version

+- hashcat Forum (https://hashcat.net/forum)
+-- Forum: Support (https://hashcat.net/forum/forum-3.html)
+--- Forum: hashcat (https://hashcat.net/forum/forum-45.html)
+--- Thread: Why Speed difference so big between brute and dictionary for (wpa 22000) (/thread-11635.html)

Why Speed difference so big between brute and dictionary for (wpa 22000) - jonas - 09-28-2023

Hi

I have three P104 graphics cards that can achieve a speed of 900kh/s using hashcat benchmark testing

I used them to crack the WPA22000 password

I have created many dictionaries myself and found that when using BF attacks and over 600000 dictionaries for Hybrid or Combinator attacks, the speed can reach 900kh/s

[Image: 1.png]

Code:
Conversation: Hashcat

Status: Running

Hash Mode...: 22000 (WPA-PBKDF2-PMKID+EAPOL)

Hash. Target...:.. TEM Tenda_ two hundred and nineteen thousand three hundred and sixty

Time. Start...: Saturday, September 23, 2023 20:23:45 (10 seconds)

Estimated time...: Saturday, September 23, 2023 20:29:31 (5 minutes and 36 seconds)

Kernel. Feature...: Pure kernel

Guess. Base...: File (.. Dic 60w_chinese name. dic), left side

Guess. Mod...: File (.. Dic birth_from_1970-2025. txt), right

Speed# 1......: 295.1 kHz/s (78.54ms) @ Acceleration: 64 Cycles: 1024 Thr: 512 Vec: 1

Speed# 2......: 298.1 kHz/s (76.73ms) @ Acceleration: 64 Cycles: 1024 Thr: 512 Vec: 1

Speed# 3...: 284.1 kHz/s (79.81ms) @ Acceleration: 64 Cycles: 1024 Thr: 512 Vec: 1

Speed. # *...: 877.8 kHz/s

Recovered...: 0/1 (0.00%) Summary (Total), 0/1 (0.00%) Summary (New)

Progress...: 7569870/302598180 (2.50%)

Rejected...: 0/7569870 (0.00%)

Recovery point....: 0/589860 (0.00%)

Restore. Sub. # 1...: Salt: 0 Amplifier: 26-27 Iteration: 2048-3072

Restore. Sub. # 2...: Salt: 0 Amplifier: 26-27 Iteration: 0-1

Restore. Sub. # 3...: Salt: 0 Amplifier: 25-26 Iteration: 0-1

Candidate. Engine.: Device Generator

Candidates# 1....: aan1155->fuzhuolin1155

Candidates# 2....: lanhuaimin1155->mobin1155

Candidates# 3....: fuzhuoming1144->lanhuaying1144

Hardware. Mon. # 1..: Temperature: 45c Fan: 37% Utilization: 100% Core: 1645MHz Memory: 5005MHz Bus: 4

Hardware. Mon. # 2.: Temperature: 46c Fan: 41% Utilization: 100% Core: 1670MHz Memory: 5005MHz Bus: 4

Hardware. Mon. # 3.: Temperature: 50c Fan: 31% Utilization: 100% Core: 1632MHz Memory: 5005MHz Bus: 4

However, when using dictionary attacks, the speed drops to 600kh/s

[Image: 2.png]

Code:
Conversation: Hashcat

Status: Running

Hash Mode...: 22000 (WPA-PBKDF2-PMKID+EAPOL)

Hash. Target...:.. TEM Tenda_ two hundred and nineteen thousand three hundred and sixty

Time. Start...: Saturday, September 23, 2023 20:29:59 (1 minute and 19 seconds)

Estimated time...: Saturday, September 23, 2023 20:51:14 (19:56 seconds)

Kernel. Feature...: Pure kernel

Guess. Base...: File (.. DIC Initials (10k)_ And birthday. dic)

Guess. Queue: 1/1 (100.00%)

Speed# 1......: 227.8 kHz/s (386.31ms) @ Acceleration: 64 Cycles: 1024 Thr: 512 Vec: 1

Speed# 2......: 231.9 kHz/s (380.40ms) @ Acceleration: 64 Cycles: 1024 Thr: 512 Vec: 1

Speed# 3...: 218.5 kHz/s (393.20ms) @ Acceleration: 64 Cycles: 1024 Thr: 512 Vec: 1

Speed. # *...: 678.6 kHz/s

Recovered...: 0/1 (0.00%) Summary (Total), 0/1 (0.00%) Summary (New)

Progress...: 51612106/863160889 (5.98%)

Reject...: 2506/51612106 (0.00%)

Recovery point...: 50628950/863160889 (5.87%)

Restore. Sub. # 1...: Salt: 0 Amplifier: 0-1 Iteration: 0-1

Restore. Sub. # 2...: Salt: 0 Amplifier: 0-1 Iteration: 2048-3072

Restore. Sub. # 3...: Salt: 0 Amplifier: 0-1 Iteration: 2048-3072

Candidate. Engine.: Device Generator

Candidates# 1....: byk730812->bzx19870112

Candidates# 2....: bzx19870113->blr980823

Candidates# 3...: blr980824->bcn20120124

Hardware. Mon. # 1..: Temperature: 52c Fan: 43% Utilization: 100% Core: 1645MHz Memory: 5005MHz Bus: 4

Hardware. Mon. # 2.: Temperature: 53c Fan: 45% Utilization: 100% Core: 1670MHz Memory: 5005MHz Bus: 4

Hardware. Mon. # 3.: Temperature: 56c Fan: 31% Utilization: 100% Core: 1632MHz Memory: 5005MHz Bus: 4

The performance of each graphics card has decreased by one-third.

I checked some information online

This guy asked me to create a password of the same length, but the speed still hasn't improved

I also saw this wiki, and I tried the aspects mentioned in it, but it didn't work

When running BF attacks and dictionary attacks, there is a significant difference in the utilization rate of the graphics card. During BF attacks, the occupancy rate of the three graphics cards remains in a straight line, but during dictionary attacks, it is up and down

[Image: Snipaste-2023-09-24-13-18-12.png]

I also have a workstation that runs a dictionary at a speed of only 180kh/s. The only difference between them is that it use a slower hard drive. This workstation uses HDD based NFS, while the workstation with three graphics cards uses SSD, which may be directly related to the hard drive?


To explore this issue, I found another new machine using E5-2696 v3 with 64G RAM and 1TB CT1000P3SSD8 nvme. This time, I used RTX 3090, which should have a normal speed of 1100kh/s. Using a dictionary to attack this computer, its speed also decreased to only 600kh/s. Therefore, it is not related to hardware. What is causing the speed to decrease by 1/3

I just want to know some technical details (if possible) about why this situation happened.



Thank you very much

RE: Why Speed difference so big between brute and dictionary for (wpa 22000) - Chick3nman - 09-28-2023

This is expected behavior, Dictionary attacks require moving data across the bus and that takes time, whereas bruteforce/mask attacks are computed entirely on the GPU which is much easier/faster. If your dictionary attack looks slower than expected, you may have poor workload: https://hashcat.net/wiki/doku.php?id=frequently_asked_questions#how_to_create_more_work_for_full_speed

RE: Why Speed difference so big between brute and dictionary for (wpa 22000) - jonas - 09-29-2023

In fact, I have been using - w 4 to improve performance, but sometimes the effect is not obvious. It seems that I need to create a new dictionary now