Member Info for Phimx

Llewelyn:

Care to critique my post instead of posting your childish insults.

Man up sonny

Llewelyn:

All my own words I'm afraid. I suggest you try and understand a little of what you invest in.

Method C software running on the operating system of the miner performs real time assessment of different headers, we are told by the web site. The first puzzle is what does Gardin mean by different headers. The header template provided by the pool forms the core of the header to be hashed. Most fields are immutable but 2 fields are very important and mutable, the Merkle root and the 32-bit NONCE field. To form different headers these fields are changed to produce trillions of slightly different headers. The NONCE field has one job, to cycle or increment from 0 to 4.3 billion. For a 1 TH/s miner this takes 8.6 ms, a mere blink of an eye. Every time a single bit changes in the Merkle root the NONCE cycles through 4.3 billion values for that variation of the header.

Of course for multi chip ASIC rigs the time drops equivalently, the latest mining rigs can achieve 1,000 TH/s so will cycle the NONCE in 8.6 microseconds.

What happens to the header after assessment by the Oracle is complete? We know the answer to that

“ this approach avoids processing a large number of headers which have little, or no probability to generate a hash above the pool target”

In other words that header is discarded and is not hashed, as we have been led to believe from previous descriptions of Method C. How is this different from the brute force approach using “instances of Headers” ?

Again from the web site description

“ Unlike current standard BTC mining approach, based on a search, where the mining rig runs as many Tera hashes per second as possible, during the circa” 10 minutes allowance for each block, with no special criteria on which instances of the header to process first, also described as brute force”

We deduce that an instance of a header is simply a specific header, what at one time was called an input to SHA256.

Conclusion:

In a word salad, Gardin has just described the hardware method C as the software method. He claims the software method is now possible because headers are assessed and not instances but they are the same thing. The reality is Gardin didn’t understand the hashing process, and neither did I. The hardware version never needed an ASIC to operate, it had the time period whilst the 4.3 billion NONCEs were being rolled to construct a new header. For a 1 TH/s hasher that time was 8.3 ms probably sufficient for a fast microprocessor to create a new header and assess if it was worth hashing.

Gardin has continually changed the story as he has learned and covered up the flaws in his story. So much for a team of experts !

Mrhanji:

Computer demonstrations or simulations aren't a working prototype !!

Ellipses: You are correct, I have been studying the Merkle roots role in the header and the 32 bit NONCE field. Recently I came to a similar conclusion as your own. For a 1 TH/s single chip ASIC it will roll the NONCE in 4.3ms. Once back to its origin a new Merkle root has to be formed and the NONCE is rolled again. Since single chip ASICs are achieving faster hash rates and multi ASIC miners are achieving phenomenal hash rates I don't think an ESP32-3 could manage the throughput but it is just about plausible. Given some headers will be rejected I guess some dead time may occur in the hashing. I continue to learn!

It seems obvious this is not PI buying nor I doubt institutional holders buying, they don't telegraph their intentions. So it has to be the market itself and they seem to be trying to induce a pump from PIs. The question is why ?

Mrhanji: You post

" It also says :QBT’s R&D teams rely on a combined research experience of more than 200 years, which includes some of the most accredited international figures in the fields of quantum computing, cryptography, machine learning (AI), optimisation theory and FPGA and ASIC design"

I guess those are the same experts who don't understand discarding 30% of inputs to the hashing ASIC will result in a 43% hashing increase.

Pull the other one, apply commonsense

Mrhanji:

You mean those ASIC experts that failed to realise ASIC design was horrendously expensive. Gasdin strung that one out for years.

You have a choice to believe what you will and make your investment decisions accordingly. Common sense is a wonderful thing but overshadowed by naivity and greed.

Mickydj:

What open short positions ?

If you think I am part of some financial conspiracy theory report me, I welcome it with open arms.

The share price seems unusually static for this avalanche of buying!

As an aside, why does the Web site state

"QBT’s Method B, guides the search, by deciding where the most promising winning hashes are likely to be found in the 2^1024 search space."

Surely the successful hashes are found in the space containing all possible hashes I.e. 2^256 for SHA256 and Bitcoin ??

Looks like Micky is consulting with the mumbler

Mickydj: From QBT web site

It would seem from your description of Method C, it is identical to Method B. Is that correct

Method B from web site

"QBT’s Method B, guides the search, by deciding where the most promising winning hashes are likely to be found in the 2^1024 search space.

Method B performs an informed search using the degree of freedom made available different header’s parametersthereby selects portions of the search space to be explored by SHA-256. This approach avoids processing a large number of hashes, obtaining the target result, u.e. the a winning hash, in less time."

Mickydj: You post

" No Phimx we are not going to get into semantics for you to create your strawman arguments and stretch the negative narrative."

So you do not want to prove my assumed negativity false by citing RNS or publicly available information !!!!!! How very interesting ????????!!!!!

Mickydj:

There are no semantics here. I ask

Where do you get the public information from. Cite your reference or it is simply speculation and false information.

Mickydj: You post

" This assumes Method C rejects block headers outright, which it doesn’t, it biases nonce search, it doesn’t halt it, and miners don’t sit idle."

Interesting, where do you get this source of information from ????

You also admit, using my assumption, based upon the information on the QBT web site, is actually correct and far from misleading

You have recently stated you only use RNS or publicly available information ??.

Previously I've stated that Method C software running on a single slow processor and assessing each input to the SHA256, would be unable to match the throughput of the ASIC, performing TH/s. It is impossible! To add further credibility to the argument I'll introduce a second flaw in the system. Gardin states Method C Oracle checks each input and decides if it is worth hashing, if so that input is sent to the ASIC. Now essentially each input apart from the NONCE is identical !! The flaw is the NONCE is only added within the ASIC where, for each of perhaps 2,000 SHA256 identical channels, it is incremented very fast. Realising the flaws Gardin revised the Web site, it now reads the Oracle no longer assesses each input but only assesses the header itself I.e. the NONCEs now play no part, but that only leads to other issues!!

RNS 24 October 2024

"Method C's AI Oracle is able to predict the likelihood of an input to SHA-256 (the core algorithm to mine Bitcoins) to produce a winning Hash. Should the AI Oracle calculate that the current SHA-256 will not be successful in finding the winning Hash, it skips that calculation and moves on to the next input"

From QBT website:

Can anyone shed light on the description provided by Gardin on how Method C software works. From the first excerpt below he describes standard mining using instances or brute force. I think we can say an instance of the header is the header + NONCE and therefore billions are created and each one hashed. Gardin goes on to say, in the second excerpt only the header is assessed. This suggests that when received from the pool the header is either accepted or rejected before any NONCES are considered. If rejected then the miner sits idle for 10 mins until a new block header is sent by the pool. This also implies if accepted then NONCES are applied to create instances which are then hashed as normal ?? Now also since virtually each block header will contain a winning hash how would a block header be rejected ???

*********************************

Unlike current standard BTC mining approach, based on a search, where the mining rig runs as many Tera hashes per second as possible, during the circa 10 minutes allowance for each block, with no special criteria on which instances of the header to process first, also described as brute force, or blind search,

*************************************

...Unlike the hardware version of the Oracle, the software version assess headers, rather than instances of the header, hence it can be run by the processor of the control board, without causing any idle cycle of the hashing boards.

Mrhanji: You post

" 3 separate NDA's so isnt some trick or scam - shows seriousness"

Is that similar to ' revenue in 2025' or possibly ' after summer' ???

Mrhanji: You post

" Under AIM rules they must RNS any MATERIAL setback"

Who decides if a setback is material or not ?