https://bitmine.ch/?page_id=863
You said on your web site that you've licensed an exclusive-patented sha-256 work engine. can you say more about which sha-256 engine youve licensed and what is its spec?
You also said that the process youve used is the Global Foundries 28nm HPP.... and that your voltage runs at 0.5, 0.65 and 0.75 volts, depending on low power, nominal and turbo modes? yet these voltages are much lower than Global Foundries claimed voltages for HPP. Is that intentionally running at much lower than the official voltage (0.85v) or a misprint ?
http://www.globalfoundries.com/technology/32-28nm.aspx
You said on your web site that you've licensed an exclusive-patented sha-256 work engine. can you say more about which sha-256 engine youve licensed and what is its spec?
You also said that the process youve used is the Global Foundries 28nm HPP.... and that your voltage runs at 0.5, 0.65 and 0.75 volts, depending on low power, nominal and turbo modes? yet these voltages are much lower than Global Foundries claimed voltages for HPP. Is that intentionally running at much lower than the official voltage (0.85v) or a misprint ?
http://www.globalfoundries.com/technology/32-28nm.aspx
As a chip designer, let me conjecture an answer. They said that they are using low power transistors. That might help a little, but not as much as you are showing here. For all current manufacturing technologies, the digital logic gates can run at a much lower voltage than the RAMs. The quoted number will be what makes the RAMs work correctly. SHA does not need any RAMs to be implemented, so if they skipped using RAMs in all the other logic (the control logic which submits the keys to the SHA engines and checks he results), then they can get by with a significantly lower vmin on their part. This difference is significant enough that most major manufacturers will be using separate voltage rails for RAM vs. logic in new mobile chips (Intel,Apple,Samsung,Qualcomm,nVidia,etc.).
Hope this helps.
Oatmo
