Bitcoin Forum
December 14, 2024, 03:34:05 AM *
News: Latest Bitcoin Core release: 28.0 [Torrent]
 
   Home   Help Search Login Register More  
Pages: [1]
  Print  
Author Topic: Hardware for My Next Project  (Read 2568 times)
DoctorDoom (OP)
Newbie
*
Offline Offline

Activity: 5
Merit: 0



View Profile
June 21, 2013, 08:07:21 PM
Last edit: September 16, 2016, 04:01:05 AM by DoctorDoom
 #1

e
tom_o
Sr. Member
****
Offline Offline

Activity: 308
Merit: 250


View Profile
June 21, 2013, 08:47:50 PM
 #2


I will start from scratch and use Verilog coding. I will make custom circuits for everything to get the max performance. Timing analysis will be key in this project. Meeting all set-up and hold times with guard band will take some time, and will be a pain. But the good thing, there are multiple phase that can be used.

I will start on the project after I get the first release out of my WinMiner/Dominatrix project. With the Dominatrix Engine, I already have a software platform that I can tie in the FPGA too. Depending on the performance I get out of the FPGA, will depend if I release the code to the public. If I get awesome performance, then I will keep the code. If I get OK performance, then I will release the code to the public for free.




Interesting, but what do you hope to gain by keeping it closed source? And you might need a bigger heatsink than than!
teknohog
Sr. Member
****
Offline Offline

Activity: 520
Merit: 253


555


View Profile WWW
June 27, 2013, 09:47:53 PM
 #3

Wow, your approach is seriously interesting. I wonder how much of the current ASIC mining designs are simply hardcoded implementations of the old FPGA stuff -- obviously, there must be some ASIC specific design, but to the outside the entire mining industry seems like an endless scaling up of old ideas.

Designing for a specific FPGA and using its timing quirks to your advantage goes against all the basics I hear -- that FPGAs are meant for quick prototyping. This is a good sign of the old hacker spirit.

I've also wondered how much FPGA circuit designs are limited by programmers who are used to sequential designs. For example, could you design a clockless network that spits out the result of SHA256. I dabble in analog electronics, where you can do fairly complicated things without any clocking, so I like to imagine how far you could extend these ideas in the digital realm. Then again, I understand it's the analog properties of a circuit that introduce delay, and one way to deal with that delay is clocking.

Anyway, good luck with the project Smiley I do hope you would opensource it at some point, so we can all learn something after you've earned enough coin.

world famous math art | masternodes are bad, mmmkay?
Every sha(sha(sha(sha()))), every ho-o-o-old, still shines
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!