Project Evil Genius – Custom SHA2-256 Circuits on a FPGA

[chunglam]

Even having a hundred dollars or two a month would be nice

How about this. Send bitstreams to early adopters with understanding that they do not share the bitstream with anyone (unless you ever decide to make it public), and they send you the earnings from the delta for 2 weeks.

Example I have 11 Ztex 1.15y (44 LX150 chips total) currently at ~9.5 GH/s. If you make it 3x = 28.5 GH/s , i would gladly pay you 19 GH/s worth of mining proceeds for 2 weeks. I would in fact point the equivalent number of boards to a pool of your choosing. Thats $30.84/day from me alone at todays difficulty, maybe by the time ur done its $10/day... maybe lesser...

I am sure im not alone in accepting this kind of offer... but this requires a leap of faith on both sides. (You trust that i donate the promissed hashrate to you, and I trust that you dont have something evil like timebombs in the bitstream)...

^ If you decide to opensource it(and provide bitstream that works with ztex firmware) , id donate a months worth of extra hash rate.

That would be an interesting option. I will have to think about. I am paranoid about this, so it might be hard to get past the trust factor. With 3x or better performance, it is a large jump, so I want to make sure I play my cards right.

But like I said, I have to get the circuit working, before anybody would be interested. People need to see proof in the performance, especially when I am talking about 3x or more in performance, which is a vast improvement. Also, so many people have tried before me, and got nowhere near the performance I am getting.

The good thing is I shouldn’t have too much more work to go through on the coding. With the first hahser done, I can reuse a lot of the code. Then I can move onto sims and timings. But timings are looking good at 150MHz, and I don’t believe I can get much more out of it. I even have done a few experiments, and I only have gotten about 5 MHz jump for a large area increase, so it is not worth it.

I have 100+ ztex quad boards. How about 50/50 split the profit of delta hash power until the boards no longer profitable? Currently, my operating expense is 1/3 of my farm's output. With difficulty skyrocketing, I believe my farm can only last for a few months.

[Isokivi]

When enterpoint rolled out their cairnsmores and was struggling with getting a well performing bitstream for them I made a bounty thread with input from the community that ended up being one of the largest on the forum to date. A bitstream was produced that met and exceeded the bounty terms and the bounty was paid out. I'd like to suggest someone does this again. This time there needs to be a larger weight on the difficulty at the time a bitstream comes out. Im afraid I do not have the time atm to champion the bounty, but I will gladly pledge a portion of what extended life (earnings) this could end up giving my Cairnsmores.

[Loredo]

Sir, I just have a couple comments for you.  Since they're free, I'm certain that they will be worth every penny you have to pay for them.

First, the following dichotomy is as interesting as hundreds or thousands of similar ones we can find  spread over the last several hundred years, in all forms of technology:

fpgaminer(at post 50) : "BitFury's open source design for the Spartan 6 LX150 is already at the zenith and will not be surpassed by any meaningful margin.  That design achieves ~300MH/s ..."

DoctorDoom: "I can fit a lot more Double Hashing circuits on a LX150 chip. It looks like I will break the 1GH/s on a LX150 chip ..."

Taken at its face value, that speaks to advances and economisations you have affected in your implementation of the double hash.  If these advances are original, then that has indirect value in terms of greatly enhancing your reputation, and thereby, your human capital.

Human capital can't be spent at the store to feed one's family, though.  The problem you face is that, as the bitcoin network hash rate heads relentlessly towards 500 TH and beyond, 1 GHps on an FPGA with a wholesale or used price in the $50-range won't feed one's family, either. 

On the other hand, if you could produce an incontrovertible video of a single Spartan 6 demonstrably hashing at or near 1 GHps, then someone might determine that such functionality had direct, scalable value to them in some way.  Then, through license, or though venture, you might be able to release some of the potential value locked in your labor.

Best of luck to you.

One other point, re: the recommendation to contact the current bitcoin-asic vendors.  You worked for Intel.  Your skills, at least as self-represented, appear quite impressive.  Your world view is probably not the same as theirs.

[chunglam]

Hi Doctor, any good news on the progress? In case you need access to ztex quad board, I can give you remote access to a linux VM with one or two ztex boards attached. PM me if you interested.

[JuniorJack]

I coded up a quick SL3 cracker about a year ago.  It either ran on my Spartan 6 devkit, or the X6500, I can't recall.  I could probably dump the code to github if people are interested.  I didn't optimize it particularly well, just got it working.

Hi,

Is it possible to run SHA-1 in single clock cycle, similar to the SHA256 in your implementation ? And what is the max clock of
say Spartan 6 ? 1000 Mhz ? If yes, this means single xtex quad board can match a single ATI 5970, which is not that bad.

BR

[Dexter770221]

I coded up a quick SL3 cracker about a year ago.  It either ran on my Spartan 6 devkit, or the X6500, I can't recall.  I could probably dump the code to github if people are interested.  I didn't optimize it particularly well, just got it working.

Hi,

Is it possible to run SHA-1 in single clock cycle, similar to the SHA256 in your implementation ? And what is the max clock of
say Spartan 6 ? 1000 Mhz ? If yes, this means single xtex quad board can match a single ATI 5970, which is not that bad.

BR

Max LUT speed at Spartan6 is 400MHz. Additional delays with routing and whole project runs at max 200MH.

[JuniorJack]

I coded up a quick SL3 cracker about a year ago.  It either ran on my Spartan 6 devkit, or the X6500, I can't recall.  I could probably dump the code to github if people are interested.  I didn't optimize it particularly well, just got it working.

Hi,

Is it possible to run SHA-1 in single clock cycle, similar to the SHA256 in your implementation ? And what is the max clock of
say Spartan 6 ? 1000 Mhz ? If yes, this means single xtex quad board can match a single ATI 5970, which is not that bad.

BR

Max LUT speed at Spartan6 is 400MHz. Additional delays with routing and whole project runs at max 200MH.

Hi,

Thanks! So the only chance to have performance boost is to fit 2 or more sha instances to run in parallel, but with unrolled SHA1, it might not be possible to fit it all.

BR

[fpgaminer]

Hi,

Thanks! So the only chance to have performance boost is to fit 2 or more sha instances to run in parallel, but with unrolled SHA1, it might not be possible to fit it all.

BR

https://github.com/fpgaminer/sha1_collider

[JuniorJack]

Hi,

Thanks! So the only chance to have performance boost is to fit 2 or more sha instances to run in parallel, but with unrolled SHA1, it might not be possible to fit it all.

BR

https://github.com/fpgaminer/sha1_collider

Thank you! Will check it out.

Regards

[poochone]

Hi,

Thanks! So the only chance to have performance boost is to fit 2 or more sha instances to run in parallel, but with unrolled SHA1, it might not be possible to fit it all.

BR

https://github.com/fpgaminer/sha1_collider

How much u wont for full source of sl3 unlocker for fpga ??