Build Your Own ASIC Miner


#450

I begin the Masters of Science in Cybersecurity and Information Security next month.

FWM


#452

Hello je suis Français je mine en ce moment avec à peu pret 60 giant N et 5 giant b . J’aimerais vraiment discuter avec toi pour le projet de création d asic voici mon skype steve.guigs au plaisir d’ échanger avec toi


Guidance through the creation of an ASIC?
#453

Hey fnt ,

Could you drop me an email ebisu@tuta.io I’d like to hire you.

Thank you.


#454

Salut Steve33,

Laisse moi ton mail afin qu’on puisse communiquer.

Salutations.


#455

Hello Boy !!! I have a quest with the arrival of minor CryptoNight there was a lot of money xmr and many other fork that went in CryptoNight V7. My question is is it possible to change this algo on the miner type giant N? to continue to undermine the xmr?

Sorry for my english


#457

Hey mate, I have got similar approaches to this. Have you got a contact email/phone? Lets discuss


#458

Hey mate, you got pcb design experience? Hit me


#459

Hello everyone,
I am from south-Africa,very interesting couple of hours spent here.
Let me know if you need any help with the project.
Is there a progress update 13 days since last post?


#460

NodeHaven has been on the grind. You can check out their Slack and Telegram which can be found in this forum using the search function.

Other teams are practicing reverse engineer, printing their own boards and such.

How are you doing down in South Africa?

Thanks for your interest. This is our update :smiley: Happy Friday!


#461

I’m electronic.
I want participate seriously.
I’m new in the mining affaires but I know about digital systems.
There are two ways: design the ASIC (chip) embebed on FPGA OR buy it … AND buy other piece for build the Miner Machine (slang: ASIC).
Do you know how to apply the transform SHA-256? … including all data inputs…
Have you a Telegram account or channel?


#462

Well everybody just hold on. I just saw this thread and you all need a reality adjustment. Basically never in a million years it this project the way you want to take it ever going to come to market, be profitable, or even be fundable. You might just skip down to my suggestion on what WILL work if you can’t stomach some harsh criticism.

I’ve got the credentials, worked on everything from bit slice processors in the 80s, to Z80s, ASICs, FPGAs, up to and including NVMe and SSD firmware, even tweaked some compilers.

Plus, of course, I’ve got A3s, S9s, L3+, and even a few D3s. An open-source ASIC is absolutely impossible, and if you are even wondering why it is impossible, then you aren’t qualified to even make an assessment.

I’ll name just one absolute deal killer that nobody has even brought up – the necessary army of patent lawyers to deal with inevitable intellectual property concerns and royalty/licensing agreements that are going to have to be made with somebody)


-------------- WHAT IS VIABLE ---------

First, the key is not coming up with a magic ASIC. The way to make money is to come up with a coin using a new combination of OFF-THE-SHELF hardware and software that generates the work needed while minimizing clock cycles and maximizing parallelism.

We also need to come up with a coin that has a market, with low difficulty, that nobody else is attempting to optimize.

Now here is the logistic problem. Say coin XYZ has no hardware miners and the existing code is crap and can easily be optimized with some creative combination of hardware and software. This group identifies it, then WHAM. Bitcoin is off to the races, and their staff of 100+ engineers cranks it out while the rest of us are worrying about trying to fund building a prototype.

Why? BECAUSE THEY ARE READING THIS TOO, OR THEY WILL EVENTUALLY. IF not them, some other firm with the resources will do it.

History proves that first to market makes the money.

When I got my A3s, I had FIRST batch shipment. I was making $500+ PER DAY per miner. A few days later $300 per day, now it is $5.00 per day.

So there is absolutely no way that this can even be discussed in open forum. Think people, if it was possible to come up with a next generation SHA256 or Cryptonight or whatever miner that could possibly be profitable to warrant a redesign of what is already out there … then Bitmain would be shipping the S10 today (which does not exist).

Where do we go from here?? Frankly, IMO, this is unworkable. Best that can be done is improve mining EFFICIENCY on existing hardware by getting more work done in less time. That means optimizing code, pure and simple.

If an ASIC is the only way, you’re just going to have to farm out the entire project to a company that specializes in it … which means $$$$ up front, and you better have a design in mind and the software written so a simulator can test the design out.


#463

It isn’t practical. I’d be willing to bet Bitmain has some logic buried in the firmware that kills their ASICs at some time via some checkpointing firmware.

i.e, if the chip detects that it isn;'t running their code, the chip gets bricked. You’ll never find that failsafe, if it exists. I know if I was on the bitmain team, then I sure as heck would have done that. I’d want to protect my intellectual property, and insuring nobody is hacking my hardware is important.

Now, lets say the problem isn’t there, and you can reprogram all you want … It is a huge undertaking to figure out what code should be run and to make the necessary changes and benchmark it. You’d be better off starting from scratch almost.

Way too much work involved to just find out … and are you willing to donate a few miners that might get bricked to find out??


#464

You are not just going to find ASIC developers from these companies that has an employment agreement that lets them work on something else. They sure as heck won’t give up their six-figure salaries and retirement to work on something else.

But even if you did, then what would you use for a development platform, and who is going to pay for the software Now Xilinx has some decent online free-use software tools, but like all the “free” packages, they are married to the chip vendor who makes the device.

(Actually Altera has some low-end free stuff, but it isn’t going to work for this type of project)


#465

Cool, thanks for sharing! I love it!


#466

Hi guys,

I agree with dlethe info. Also don’t forget you have some economic issues to solve.

BTW: I “wrote” some software to modify the voltage of the A3 and D3.
Here is the link, it’s free: https://github.com/ForexNeuroTrader/A3_D3
Which this you can push the hashrate or lower the temp. of chips for less fan speed, etc…

Bye, FNT


#467

I think it’s about time I comment.

I’ve been following this thread since day 1.

I’m based in the UK and have a team already working on a new ASIC based miner.

We stripped down an S9 including the OS and what we found was very surprising. I’m not going to put anything here about what we found but it’s interesting.

One thing I will say is the hash rate is higher behind what you see.

I will post a few updates as the months progress regarding our research


#468

Hi,
I am squeezing out of S9´s about 20TH/s we would like to know if there is more potential.
Do you know the maximum possible mVolt a BM1387 ASIC is able to “consume”?
Would be really great if you share that info with me.


#469

I think the next generation of miners has appeared!
GMO B2 is a complete new design of a SHA-256 miner.
They understood, that the limiting edge of the S9 and all his copies is the power supply of each hash board.
Instead of feeding 63x ASIC with 3x 6Pin 12VDC they are using 4x 6Pin 12VDC to “feed” only 18x 7nm ASIC.
Each ASIC has a minimum hashing power of 450GH/s (compared to the 75GH/s of a S9)
Assuming, that we are able to overclock the 16nm BM1387 to roughly 110GH/s - it depends on the Volt settings and the individual chip quality - we can estimate that such a GMO B2 could easily run at 36-54TH/s.
(that would be up to 1000GH/s per ASIC - and GMO is expecting their ASIC will perform much more in future)


#470

For devices placed at home, I think the hash rate in Th/s level will be nice. Also one can assemble large power device by using more chips.


#471

I’m new here, can someone please give me a quick summary of this threads discussion? I don’t want to read through a half year of posts. uwu

@Duckegg Would you care to elaborate?