FPGA for MD5: how fast can it go?

[Roasted]

Hi everyone,

I have just installed CUDA drivers for my notebook video card and with BarsWF CUDA bruteforcer I am able to check about 120M passwords per second. Now, I am very curious in raising this value and I am considering about FPGA. I know there are several models around and some of them are quite out of my budget but there are models from xilinx or altera that don't include the prices a lot.

Actually my question is that how many passwords per second can such a "entry level" model check against an MD5 hash? Do you have any idea about this ????

[Calvin K]

Hi guys,

I have some suggestion for you and you can say it is not the actual solution but it can be more helpful for you to solve your issues. If you want to go through my suggestion then I will suggest you something. Would you go with the cuda and you need not to go for the spending of money on fpga unless you actually required to. The fpga's cost a fortune. Are you sure you will regain your investment?


Thanks ..

[Roasted]

Hi guys,

IWould you go with the cuda and you need not to go for the spending of money on fpga unless you actually required to. The fpga's cost a fortune. Are you sure you will regain your investment?

No way !!! I am just telling that I have figured out that there are FPGA for every budget, such as CYCLONE III series from Altera. I am wondering how could they do.

[Lillebror]

I am with gromeo. I continued with the fpga path, and there are plenty of them out there. But in regard to what's already applied and used code, (unless you need to get comfortable with some other implement yourself) you primarily looking at going with something such as a pico e12 or something like that.

That being said that I am all for someone picking up one of the budget fpgas and working with utilities for them. I would be a glad dude, but since cuda hit and now I am on the cuda wagon.

[jesse]

FWIW (you may already be known about this) but work has been committed using the PS3 console against MD5, with some pretty decent results. Some of different projects are around for this, IIRC including a distributed platform. The code for PS3 MD5 is evolving and improving, apparently with more room for optimization still there. They have not been able to get the GPU yet, only the Cell SPU's. Even so, a figure of 80M-120M hashes/sec isn't bad.

Guggle "Rensselaer Polytechnic" and "Distributed Hash Cracker". you could work with the HTTP to send the hash back to the office and use the hardware there to perform the legwork. OFC precomp/rainbow is commonly an simpler method to go unless the hash is salted/stretched.