Re: [ActiveMining] The Official Active Mining Discussion Thread [Self-Moderated]
...
Mabs, can you at least try to understand what's going on before posting? The Intellihash update is software (RTL/HDL) level code changing1. Software level changes are a far cry from a complete chip redesign. So your estimated time frame is not realistic at all.
...
Mabs, can you at least try to understand what's going on before posting? The Intellihash update is software (RTL/HDL) level code changing1. Software level changes are a far cry from a complete chip redesign. So your estimated time frame is not realistic at all.
...
Ken stated the chips were late. He attributed this to needing to redesign the chips, making them "up to 20% more efficient" (cue 20% Cooler FIM FTW!). Delayed chip production can not be attributed to software, learn to SHA256 ASIC. If the mysterious Intellihash is indeed software capable of being run on top of any chips, Ken's explanation of delayed chips is clearly a lie. Delayed chips is hardware, not software. Learn to difference.
There's simply too much fail in this thread.
Quote
Intellihash is our new trademark for our new Bitcoin mining software which gives up to a 20% increase in hashing speed and has the possibility to increase the speed of our mining machines as the difficulty goes up. We have had to modify the software in our chips to make it work with our new software. The chips are going to be late; however, our new Intellihash software could be a game changer for the company.
Changing the RTL code will result in a delay, albeit much less then a full redesign.
It would behoove you to do some research into how KnC got their speed boost.
(Hint: it wasn't a chip redesign!)tl;dr Software.
Quote
We have had to modify the software in our chips to make it work with our new software.
The above quote is nonsensical to anyone familiar with SHA256 ASIC design. Such ASICS have no more software than your CPU, which is none.
Any changes to the RTL code, no matter how minor, require a new mask set & a new foundry run. Fact.
TL;DR: Stop repeating nonsense.
I take it you didn't look into KnC?
You are right, there is delays in the process, but it is less than a full redesign - which is all that I stated. If you are just going to ignore test cases like KnC, who DID receive a significant speed boost via software optimization, (some at the hardware level) then we have nothing more to discuss.
KNC's software optimisation did not require silicon redesign. Ken claims that his Intellihash, while making his chips 20% Cooler(tm) Hasbro, required silicon redesign & was responsible for the delays. This is nonsense -- either he botched the RTL code or he didn't. Discovering that the chips could have been 20% Cooler(tm) Hasbro after tapeout is a sure sign of incompetence. So is delaying the product by months to achieve 20% higher hashrate -- during which time the difficulty climbed over 100%.
None of this makes any sense. Time to wake up.