Jump to content

24 hour Trillian pmks... wpa crack default 8char[0a-9z]


Recommended Posts


Correct me if I'm wrong.

firstly, a default password written on the back of a router is 8 chars long? 0-9 and a-z?


Looks like 2.8 trillion password possibilities?

again correct me...

what would be the most cost effective way to build a machine capable of this crack

1000 video cards? Lol idk

Link to comment
Share on other sites

The Pyrit page says how many PMKS a card can test per second. Problem is that since they made that list substantial improvements have been made to the code and also newer, faster cards have emerged.

A quick google shows a GTX 970 does just over 76000 PMKS/s (can't find stats for 980) and an R9270x clocks in at 73000 PMKS/s. No idea if the latter was overclocked or if the AMD GPU cores have superior compute code written for them. Let's assume that a modern high-end card gives you 75000 PMKS/s

Now, the question boils down to the combination of time and money.

That 75000 PMKS/s card on its own needs just over 435 days to scan the entire keyspace. If you put 4 of those in a machine, it's 109 days. You're going to have to factor in electricity costs too.

One optimization you can try is to assume no char in the password repeats which for a good random password isn't much of a stretch. That way you reduce your keyspace to 36*35*34*33*32*31*30*29 = 1220096908800 a.k.a. half the original keyspace so you can take the previous numbers and halve them too. Just under 2 months to crack a password on a beastly machine is pretty decent I'd say.

Cost-effective can only be determined when you specify how long you're willing to wait for a result.

Edited by Cooper
Link to comment
Share on other sites

I did some experimenting with pyrit cluster configuration... took my usb-kali-nvidia-stick and cloned it using a dd method...

Started up my 2 sexy machines.

Labtop geforce gtx 560m + i7

Desktop geforce gtx 460 + AMD1090t

Labtop ->14000pmks


With pyrit configured to serve as a cluster I ran a benchmark and only got 24000pmks... could be my old linksys switch, ill test others another day...

I want to see 34000pmks!

So here is what I did... I took Super-WPA wordlist

wc -l Super-WPA => 980million lines

So I rounded up to the nearest billion and split the file into 2 parts based on performance

split Super-WPA --lines=800000000

Desktop gets 800,000,000

Labtopp gets 180,000,000

If my math is correct they should complete the task at close to the same time...

Edited by i8igmac
Link to comment
Share on other sites

Are you using OpenCL drivers and sure its going through GPU vs CPU? Also have to take into consideration, read capabilities from your HDD's from say, SATA HDD vs SATA SSD, vs SATA I, II and SATA III

Often its the HDD's that can be bottlenecks in performance, and not the GPU and CPU holding you back. Especially on laptops.


Anyone remember this video?

(BTW, I know its silly marketing video, but still fundamental in bottlenecks since most time, HDD' and Memory bus speeds are slower than GPU and CPU processing)

Edited by digip
Link to comment
Share on other sites

my desktop does have enough free space on its ssd for this wordlist... I spend more hours thinking about what I'm going to test before I actually perform them... interested to see if it shows a improvement...

I have not yet overclocked anything, this AMD1090t kicks ass @10000pmks :-p

At the moment I'm only testing attack_passthrough... I don't have much free harddrive space to build a database

Link to comment
Share on other sites


i did test my wordlist on my ssd with pyrit, i dont think my pmks have passed the speeds my standerd hd can serve... no bottleneck there

i would love to build a new computer, some small ssd's like the link i posted above could maybe produce insane speeds at afordable cost...

I have been looking at bang for the buck video cards and server boards (4 socket cpu/4 socket gpu)

Link to comment
Share on other sites

CPU's produce 1/10th the PMKs a potent vid card can achieve, so I think the cost-effective approach is to save the money you would otherwise spend on a big-ass server board which invariably require expensive Xeons, to consumer-grade gamer boards which allow 4-socket GPU. Put a CPU in there that's basically just a controller so go for the cheap here rather than bang for buck. Abundant and fast storage is useful, a good helping of memory probably even more so. If one's not fast enough, make multiple of the same. Trying to put more performance into the same box makes the price jump considerably faster than the performance.

I'd also take a long, hard look at 3-socket GPU boards. The 4-socket consumer boards tend to be enthousiast-focussed, with a price tag to match. Having just 3 might put your board in a price range where you can afford a second machine and 6 mid-range GPU's on 2 "mere mortals" boards might very well be cheaper than 4 high-end GPUs on some uber board, as well as faster.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Create New...