Quote from: dolor on December 26, 2013, 06:28:58 PM

Quote from: kleeck on December 26, 2013, 06:11:24 PM

Quote from: dolor on December 26, 2013, 06:00:31 PM

Quote from: kleeck on December 26, 2013, 05:44:20 PM

Quote from: dolor on December 26, 2013, 05:35:03 PM

Quote from: kleeck on December 26, 2013, 05:05:06 PM

...
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. Wink

(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? Wink

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.

This will probably be my last time engaging you. While you started out reasonable you've ventured into la-la land. I have offered you a case study, and Ken's official description of Intellihash and now you respond with this?

KNC's software optimisation did not require silicon redesign.

Quote

Ken claims that his Intellihash, while making his chips 20% Cooler^{(tm) Hasbro}, required silicon redesign & was responsible for the delays.

I don't believe he did say this. Source?

Quote

This is nonsense -- either he botched the RTL code or he didn't.

Part of the prereqs for working with eASIC is getting the RTL code verified by their engineer team. Your follow up questions infer some pretty outlandish critiques of the eASIC engineering team.

Quote

Discovering that the chips could have been 20% Cooler^{(tm) Hasbro} after tapeout is a sure sign of incompetence.

Again with the cooling? Do you have a source for this? Also, your argument of incompetence implicates eASIC - again. I don't think eASIC screwed up here. I think Ken saw an opportunity to have better tech and jumped on it. Whether it will be worth it or not remains to be seen.

Quote

None of this makes any sense.

You have constructed a complete nonsense scenario, so of course it makes no sense.

Your whole argument here is devoid of any sourcing or facts. You assume that Intellihash is cooling (if this has been stated then I apologize, I don't believe it has, though). You have constructed a textbook argument ad absurdum, which is usually the start of full-on, factless trolling.