oclHashcat v2.10 performance questions?

[halexan]

Version of oclHashcat is "v2.10 beta 192"

When cracking -m=5500 NetNTLMv1. The following results are extremely different! 

Running benchmark, speed is 1218.2 MH/s.

Code:

./oclHashcat64.bin -m 5500 -b --opencl-device-types 1

oclHashcat v2.01 (gaef8dc2) starting in benchmark-mode...

Device #1: GeForce GTX 970, skipped
Device #2: Intel(R) Xeon(R) CPU E5-2620 v3 @ 2.40GHz, 8016/32065 MB allocatable, 2400Mhz, 24MCU

Hashtype: NetNTLMv1-VANILLA / NetNTLMv1+ESS

Speed.Dev.#2.:  1218.2 MH/s (82.75ms)

Started: Wed Apr 27 16:17:57 2016
Stopped: Wed Apr 27 16:17:59 2016

Using wordlist, speed is 4558.7 kH/s!!! Its too slow! 

Code:

./oclHashcat64.bin -m 5500 /home/pptp.txt /home/dict_20160122.txt --opencl-device-types 1

oclHashcat v2.01 (gaef8dc2) starting...

Device #1: GeForce GTX 970, skipped                        
Device #2: Intel(R) Xeon(R) CPU E5-2620 v3 @ 2.40GHz, 8016/32065 MB allocatable, 2400Mhz, 24MCU

Hashes: 1 hashes; 1 unique digests, 1 unique salts
Bitmaps: 16 bits, 65536 entries, 0x0000ffff mask, 262144 bytes, 5/13 rotates
Rules: 1
Applicable Optimizers:
* Zero-Byte
* Precompute-Final-Permutation
* Not-Iterated
* Single-Hash
* Single-Salt
Watchdog: Temperature abort trigger set to 90c
Watchdog: Temperature retain trigger disabled

Device #2: Kernel /home/oclHashcat-2.10b192/kernels/m05500_a0.9b12b1e1.kernel (492680 bytes)

Cache-hit dictionary stats /home/dict_20160122.txt: 4473009478 bytes, 397531497 words, 397531497 keyspace

[s]tatus [p]ause [r]esume [b]ypass [c]heckpoint [q]uit => s

Session.Name...: oclHashcat
Status.........: Running
Input.Mode.....: File (/home/dict_20160122.txt)
Hash.Target....: XXXXXXX::::XXXXXXXXXXXXXXXXXXXXXXXXXX...
Hash.Type......: NetNTLMv1-VANILLA / NetNTLMv1+ESS
Time.Started...: Wed Apr 27 16:17:41 2016 (2 secs)
Time.Estimated.: Wed Apr 27 16:19:09 2016 (1 min, 25 secs)
Speed.Dev.#2...:  4558.7 kH/s (0.96ms)
Recovered......: 0/1 (0.00%) Digests, 0/1 (0.00%) Salts
Progress.......: 8848352/397531497 (2.23%)
Rejected.......: 992/8848352 (0.01%)
Restore.Point..: 8836064/397531497 (2.22%)
HWMon.GPU.#2...: N/A Util, N/A Temp, N/A Fan

I remember that hashcat v2.00's speeds of Straight and Brute-force mode do not have such difference, for example, benchmark speed is 190Mh/s and Straight mode is 80Mh/s.

Why the speeds are so different on oclHashcat v2.10?

[atom]

Your doing 2 things wrong!

Read this: https://hashcat.net/wiki/doku.php?id=fre...ck_so_slow Point 2 and 3

Basically 1) add a rule (important!) and 2) add -w 3

[halexan]

Your doing 2 things wrong!

Read this: https://hashcat.net/wiki/doku.php?id=fre...ck_so_slow Point 2 and 3

Basically 1) add a rule (important!) and 2) add -w 3

I tried "add a rule" and "add -w 3", and the speed rises to 105.6 MH/s.

However, I don't want to create more work for full speed, I just want to try every password in my wordlist faster. Is 4558.7 kH/s the highest speed that transmit passwords in wordlist from memory to the compute kernels (CPU)?

[atom]

Not sure what your problem is, 105MH/s is pretty good for a CPU.

[halexan]

Not sure what your problem is, 105MH/s is pretty good for a CPU.

I just want to know why "v2.10" using word list in straight mode leads to a big speed drop, as I mentioned above, 1200Mh/s downto 4500kh/s. While hashcat v2.00 does not have this serious problem (not too much drops).

Does the reading speed of wordlist affect the crack speed?

I mean, when cracking a password, you should take out a password in the wordlist, and caculate whether it is right, if it is not right, you should take out the next password in the wordlist. Does the speed of taking out passwords affect the crack speed? If read password faster, the crack speed will be faster?

[atom]

I think you're confusing hashcat and oclHashcat here, is that possible?

[undeath]

Looks like something is off to me. Check the outputs in the first post, atom.

[atom]

Don't see anything

[halexan]

I still don't understand. Why there is a big performance difference between brute-force and straight mode?
brute-force 1200Mh/s, using wordlist 4500kh/s

Is there any method to narrow the gap?

[undeath]

Is there any method to narrow the gap?

Basically 1) add a rule (important!) and 2) add -w 3

note: preferably a rule file with more than one rule