Vbulletin Low Bruteforce Speed?
#1
I am bruteforcing a couple of VB hashes with 3 char salt and getting low speeds.

Infact, I am a bit puzzled about the speeds I get for VB hashes in different modes.

My GPUs:

ATI Radeon HD 5870 (for oclhashcat only)
ATI Radeon HD 6770M (for both)

oclhashcat and mask attack as follows:

Code:
oclhashcat.exe -m 5 -n 160 --remove -o foundVB.txt vbulletin.txt -1 ?l?d?s wordlist.txt ?1?1?1?1 --increment
The speed of cracking is: 137.9 Mhashes/s

There are 8 VB hashes in vbulletin.txt with a 3 char salt.

I tried on another GPU: ATI Radeon HD 6770 M and got a speed of 29 Mhashes/sec.

Now, with oclhashcat-plus

These stats are with ATI Radeon HD 6770M

Code:
oclhashcat-plus64.exe -m 5 -n 160 --remove -o foundVB.txt vbulletin.txt wordlist.txt

With this Dictionary Attack, I got: 41993.2k c/s or 41 Mhashes/sec

Now, after applying rules to the attack:

Code:
oclhashcat-plus64.exe -m 5 -n 160 --remove -o foundVB.txt vbulletin.txt wordlist.txt --rules=T0XlC.rule

228 Mhashes/sec

Now, the way I see, it's an extreme under utilization of the GPU in the case of oclhashcat with a mask attack.

I understand that by selecting more rules and applying them to a decent size wordlist, the amount of computations done by GPU increases, but is there a way to replicate the same effect in the case of a bruteforce or mask attack too?

The above benchmarks clearly depict that there's a huge margin of difference in the speeds obtained in different modes and with different programs. It would be great to simulate the same effect in oclhashcat to better utilize the computational power of GPU.
#2
No, it's actually simple: oclhc-plus is far more optimized than oclhc, that's why you get better speeds.
The speeds are ok for 8 VB hashes.
#3
Yes, it's evident from the stats that oclhashcat-plus is more optimized than oclhashcat. So, to reiterate my question:

What's the reason behind the better performance of oclhashcat-plus over oclhashcat? Just a few technical details (if possible?)
Is it possible to have the same optimization in oclhashcat for a mask attack too?
#4
oclhc is being ported to oclhc-plus, so in time oclhc will be deprecated.