The S19 XP hash board is composed of 110 BM1366 chips, which are divided into 11 domains, and each domain is composed of 10 ASIC chips.
A S19 XP ends up being around $17.14/terrahash
Whereas a single Bitaxe is around $350.00/terrahash if you assume it's $175 to purchase + assemble, etc.
This means that the Bitaxe (while great for learning + fun) isn't efficient enough for real world use case.
What could we do to make it more efficient? I get that the purpose of this amazing project (nice job by the way) isn't to compete with ASICs/Bitmain products in terms of efficiency. It is for learning, documentation, etc.
I was just curious if there was some easy form of "copy + pasting" that could be done and then boom, you've got 4 ASICs per ESP32 and it isn't 20x less efficient than $2k ASIC anymore.
A S19 XP ends up being around $17.14/terrahash
Whereas a single Bitaxe is around $350.00/terrahash if you assume it's $175 to purchase + assemble, etc.
This means that the Bitaxe (while great for learning + fun) isn't efficient enough for real world use case.
What could we do to make it more efficient? I get that the purpose of this amazing project (nice job by the way) isn't to compete with ASICs/Bitmain products in terms of efficiency. It is for learning, documentation, etc.
I was just curious if there was some easy form of "copy + pasting" that could be done and then boom, you've got 4 ASICs per ESP32 and it isn't 20x less efficient than $2k ASIC anymore.