Quote from: 2112 on May 09, 2017, 06:53:19 PM

Quote from: delicopsch56 on May 09, 2017, 05:02:56 PM

I don't want to dig much deeper, this old dog just flat didn't think about the multiple cpu\core overhead at all from a hardware design perspective.

Quote from: delicopsch56 on May 09, 2017, 08:07:25 AM

Kinda noobie question, I'm still on board with SCAM SCAM SCAM but I don't follow processor dev for years, and I claim no EE, I think I get the basics of tough miniaturization issue advances on these chips at a high level.

You may not like my answer, but I will be short and frank.

The primary reason has of your inability to understand has nothing to do with old age, not following the recent trends, etc.

You've simply received a horrible education and know nothing about the digital technology advances from made in the middle of 1950 decade.

You seem to only be aware of https://en.wikipedia.org/wiki/Von_Neumann_architecture first published in 1945 and you seem to try to translate everything into it, even if clearly the implementation uses different conceptual model. Bitcoin mining is a perfect example of problem better handled by https://en.wikipedia.org/wiki/Mealy_machine (from 1955) or https://en.wikipedia.org/wiki/Moore_machine (from 1956). Any discussion involving concepts like: caches, branches, CPU, I/O, threads, cores, etc. only shows that the person writing it doesn't know the technological advances from the middle of the previous century. In my school these are discussed in the 2nd or 3rd semester of education, literally during couple of of first lectures in the digital logic design (both theory and lab practice)

The primary advances in the power efficiency of the Bitcoin miner were:

1) to implement it as fixed program Moore machine on an FPGA. The FPGA device is itself reprogrammable, so it is still wasteful
2) to have the same fixed program Moore machine implemented without the waste of supporting reprogramming and take advantage of the fact that Bitcoin's 2*SHA256 is essentially self-testing, so even the standard chip-testing circuitry is not required.

Personally, I see no point of discussing advanced electrical engineering stuff without understanding of the basics.

When I was in school it was a common understanding that students with absolutely no contact with any computer are doing noticeably better than students who gained experience of computers via some horrible "home computers" programmed in BASIC with plentitude of GOTOs. There was this seminal paper "GOTO Considered Harmful" written in 1968 by Edseger Dijkstra and published same year by the Communications of the ACM.

I presume that you (and other otherwise educated people) suffer from some version of the above problem: lack of proper basic education in computer architecture. Sometimes I wonder how those people graduated with any real degree (not from a degree mill). But then I have to remind myself that nowadays there are plenty of accredited, real "humanistic/psychological/human-oriented" educational institutions that do grant real degrees.

Don't mind MOST OF your reply much, actually. No, I didn't have a horrible education, I opted out of CS in engineering school because I was cognizant that I was hitting a wall in hard engineering and wasn't trying hard enough in digital design, physics, calculus, etc, too busy having fun, and I had no $. Reality check, almost everyone does hit a wall at some point where they realize they are just faking it if they continue to try to advance, learning how to do it pretty well but have stopped total comprehension.

Did a different career for 5 years, went back and did CS business degree and had a very successful career in business software dev. I pretty much quit giving a crap about hardware while mainframe programming for 13 years, because I was riding the fastest horse in the hardware world then, hardware not by me, but smokin hot MVS with serious capacity hardware for a many billions company and I focused on delivering high quality for my end users, and that went really well as I could actually communicate with them and deliver quickly and on target and with a low margin of error without 5 intermediaries like today's email writer management layers of bullshit overhead.

If ya saw no point, why did you reply? Did you read all my posts, or was it just time to be a dick to the new guy?