Nanominer - Modular FPGA Mining Platform

[wondermine]

Currently I'm writing a SHA-2 IP core from scratch that can be rolled or unrolled as necessary, and use DSP circuitry as available for simple migration from chip to chip, generation to generation.  Currently the goal is to get the SHA core completed, verified, and running efficiently; once that's finished the rest will be easier in comparison.  From this follows that no, I do not have a mining core up and running.
The goal of a new SHA-2 core is to have something that can be configured as needed for high or low end FPGAs (versatility is key considering not everyone can buy a device with DSP cores), as well as something that is well documented and I can distribute freely as my own IP.
The SHA-2 core is going well, and the project will almost certainly speed up once another hardware engineer who's agreed to work with me is ready to start (next two weeks or so I believe).
Check the website for a large amount of information you're interested in, links are posted there regarding current prices and shipping availability etc: www.nonverba.org/nanominer
Also digikey.com is a great site to check out with anything regarding circuit purchases.
As some people have noted, buying chips themselves is generally more expensive than buying the development kit, likely (now that I think of it) because buying FPGAs that can go in whatever application you like is more what they want to make their money off of.  The development kits are there to "get you hooked" essentially, for large scale production.  For example, a great deal of IP is device-locked, or time-locked, or both. 
They will communicate over serial using USB, other interfaces can be added as they become necessary or convenient, thanks to the versatile nature of the technology.
If you'd like to buy multiple FPGAs and put them on one PCB, you can certainly do this; it's possible, but it means dealing with giving them clean power, clean clocking, and essentially requires you have some, if not a great deal of, experience in circuit design.  That and a place to have the board manufactured and assembled.
Hopefully this answers your questions, if not feel free to pm or email me.

[BinaryMage]

subbed

Ditto.

[wondermine]

Alright everyone, I have a request.
I need BTC or other currency enough to buy two DE0 Nano platforms at educational cost (since I'm a student).  This is to test inter-fpga interfacing and cluster compatibility.  The SHA-2 core is in a state where I can use it (beta) and I would like to start developing the framework for mining now.

This means I need $88 CAD (2x $59 Academic - contributions to-date, I'll pay s/h + tax).

Please, if you can spare even a few points of a BTC or a couple dollars, it's much appreciated and will go towards this technology seeing its full potential in a timely manner.

More information on the core's ETA will be available when my other hardware engineers are available to work.  I'm working with a number of both amateurs and experts in the field to make this the best mining technology available.

One note: I don't want to promise specific rates anymore.  I'm still confident I can get >50MH/s out of a DE0 but numbers cause controversy.  Here's what I will promise:

Nanominer, with your support, will be the most versatile, well documented, modular, and platform independent (we will support Xilinx, Altera, and others, playing on each of their strengths) FPGA mining system available.  The more donations we receive, the faster it will be released.

This is going to be the freely distributed, well supported, gold-standard for a technology that (in my opinion and with the right development) will supersede workstation mining.

We are also doing research into new and better ways of mining, seeking help from experts in the fields of cryptanalysis and computer engineering.  We mean business.  For more information, PM or email me, and as always visit http://www.nonverba.org/nanominer

Thank you in advance for your generosity.

[ZodiacDragon84]

list it on GLBSE, put the word out on what you are doing.

[wondermine]

list it on GLBSE, put the word out on what you are doing.

I actually started to... then realized you need 20BTC to register an asset.  If I do acquire that much BTC, I will certainly consider it though.  Until that time, this will have to suffice.  Thank you.

[AniceInovation]

subbed http://www.dinigroup.com/product/data/DN7020k10/images/7020k10_front.jpg

[Transisto]

sent 1.23...

At all cost have some sort of a business model, I'd donate much more If I could get an FPGA head-start.

-Fellow Canadian

[wondermine]

Sent a coin. You from UofM?

You bet!


And thanks for the donations people, this is a great kickstart!  I'll keep you all posted.

[Intention]

What kind of rate are you getting on the core now on any of your FPGA's?

[wondermine]

Running heavily modified fpgaminer I can still pull 3.6GH/s out of my Stratix IV without using DSP, however this does not yield accepted shares.  I tested the core and it hands out correct numbers (SHA-2 wise anyway), but the script to interface may not be working properly.  I've spoken with makomk and gotten another python based script but the JTAG library doesn't (apparently...) support the Stratix IV GX 230 toolchains (even when added from the Altera site, it's a little buggy, Azelphur and I spent some hours trying to debug it).  This means I'm stuck with the Tcl that doesn't seem to be working.

If anyone is interested in looking at this:
https://github.com/progranism/Open-Source-FPGA-Bitcoin-Miner

Especially the miner.tcl and related scripts, maybe you could tell me what is wrong with the way data is being sent.  I've said before and I'll say again, I'm not much good at scripting, I'm far better at realtime and hardware, so if anyone has any thoughts, I'd love to hear them.

Nanominer is not yielding packets yet because it's not supposed to yet; I'm still working out the kinks in the SHA-2 core and it'll be a little while before I add on the circuitry to manage bitcoin mining in particular, much less develop the microcontroller interface.

[wondermine]

Alright, everyone, we've got accepted packets with the modified fpgaminer firmware, now I'm just optimizing it.  Made a couple mistakes
of my own and had to fix some things, but I'll have some nice big MH/s figures on that modified code later this afternoon!

[ummas]

Give us somme screen to prove that and get more donnation BTC...
simple ?

[Transisto]

Am I missing something or this is better than BLF ?

[ummas]

BFL has better Mhas/$ but worse Mhash/W. BFL should be idiot proof, and nanominer is in the development phase.

[cypherdoc]

BFL has better Mhas/$ but worse Mhash/W. BFL should be idiot proof, and nanominer is in the development phase.

But is bfl real?

[ummas]

look at the photos in offtopic thread... could it be not real
the old lady convinced me it is

[Wandering Albatross]

To future-proof why not use CycloneV?  Too expensive? Can you add something to your site saying exactly where you are regarding $$ and/or BTC? Essentially you want a collaborative project. In that case you need to be open about all the funding details. A funding thermometer seems to be a simple approach to letting people know what's been raised so far.

[Transisto]

...
I have a modified core running on a Stratix IV at 3.6GH/s, and am redesigning the core from the ground up to allow for similar usage on the lower end FPGAs....

How much does that Stratix IV cost,  What exact model is it ?

Is it much more cost effective to go nano ?

[wondermine]

...
I have a modified core running on a Stratix IV at 3.6GH/s, and am redesigning the core from the ground up to allow for similar usage on the lower end FPGAs....

How much does that Stratix IV cost,  What exact model is it ?

Is it much more cost effective to go nano ?

Nano isn't necessarily cost effective, it's just easy to get into at $89.  The Stratix IV I'm using (DE4-230 board) is $2995 retail.

To future-proof why not use CycloneV?  Too expensive? Can you add something to your site saying exactly where you are regarding $$ and/or BTC? Essentially you want a collaborative project. In that case you need to be open about all the funding details. A funding thermometer seems to be a simple approach to letting people know what's been raised so far.

This core is future-proof in that it's not set up for one core, it can be configured with generics to be implemented in any core.  The only reason we're using the Cyclone IV is that Terasic has their cheap DE0 Nano available and it has a Cyclone IV.

Give us somme screen to prove that and get more donnation BTC...
simple ?

Assuming you all want screenshots that show me getting more than a measly 200MH/s out of a Stratix IV, I'll post them once I have them.  That said, as I type this I'm staring at Quartus, waiting for my compile to run.

As far as a funding thermometer, that's a good idea.  I'll get on that once I have a reasonable MH/s figure for you all to look at, I get the point, you guys need proof, and that's not a problem for me.  What kind of rate would help you all see that we're moving towards something reasonable?  And please note: I don't be posting Nano MH/s rates yet, I have a DE4 and I will be posting DE4 rates until I have a Nano. 

I'm in the process of setting up multiple cores, optimizing speed, using the DE4's DSP, and other improvements, so I don't want to prematurely post screenshots.  If there's a threshold (say, 400MH/s or something? please tell me) then I'll post screens once I pass that threshold.

[Transisto]

Of course I did not meant to ask you to distribute educational hardware,  I'll look into what they require for educational pricing.

I'd see a lot of people going for a few DE4 if it can do 3.6ghs,  the 400 price difference would make up for lower consumption ("Singles" would uses 345w for same hashrate)

Also there is much more fun to be had on a dev board than that BFL Single, I could even see some resale value.