Topic: Unbeatable algorithm? (Read 676 times)

A compute bound PoW (or one using limited memory like scrypt) results in
a self-contained ASIC allowing for many generations of improvements and refinements.

With a memory-latency bound PoW requiring the use of separate DRAM memory modules,
an ASIC only needs to be able to saturate memory bandwidth, meaning fewer generations and less obsolescence.

At a given level of investment, this will also make mining much less power-hungry,
with costs shifting from power/cooling costs to investment in DRAM, which is even
more general-purpose and obsolescence proof than GPUs.

While Kokojie is correct regarding Viruses/botnets.  The problem people have with ASIC mining, and the reason their is a push for coins to be ASIC resistant also has to be weighed in.

PoW in itself is just a simplistic game.  If a coins network hash rate is 1 Mh/s from having 1000 machines mining it, it is not more secure if you have 1 Gh/s and have only 3 machines mining it.  The security of a coin comes from how dispersed it's mining is, and how many machines ($$$) it would take to overthrow it.  With GPU mining you can get literally millions of people joining in with what they already have, or for a fairly low cost.  Which keeps the coin very secure.  Someone would require millions of Video cards to attack it. 

With ASIC's, its more of a nuclear arms race.  First it's $500 to get into the game, then $1000, then $10,000.  Who knows the next generation might require $100,000 to get into the game.    Each time it jumps, more and more people get pushed out, and the more it turns into Corporations trying to outmine each other.  (First of all I thought cryptocurrencies were made to go against corporations/banks/big business LMAO).  Second of all, as things keep ramping up, eventually all the power will be in the hands of a handful of corporations, and the actual network security (Might be measured in some astronomical hashrate) won't actually be all that secure. 

The best/most secure coin possible would be to have an algorithm that is ASIC minable, but enforce that no one is to manufacture an ASIC beyond the $500 limit.  Kind of like how sports teams are bound to a pay cap, even if they could pay a lot more and make a super team.  This would make it very secure, as the hash rate would be spread out to possibly millions of people, would be the most energy efficient, and be a good deterrent for Botnets/viruses.  But we all know greed would get in the way, and that will never happen.

CPU mining is NOT preferable, as it encourages virus and botnets to mine, and soon the cryptocurrency will be associated with virus/botnet and get a bad reputation.

GPU mining is middle of the road, it's possible for virus/botnets to mine, but they can probably be used for more profitable things like paid DDoS.

ASIC mining is most preferable, as it requires specialized equipment, no point for virus/botnet to attempt mining.

The only way to make an algorithm un-ASIC minable is to ensure only a handful of coins are using the algorithm, and that the value of the coins using that algorithm are not worth enough for people to put in the time \ money into building an ASIC.

ASIC's are simplistic in nature.  They aren't supercomputers, or anything of the sort.  The name says it is.  Application Specific.  They are built for 1 thing and 1 thing only.  Where as a computer is designed to do millions of different functions, but the mining software is getting it to do a certain task.  

Unless you have millions of dollars at your disposal, you can't "make" an ASIC.  You can buy an already made ASIC chip (for SHA-256 or now Scrypt) and build yourself a miner with it.   The ASIC chip is built in a way that it can only do 1 pre-programmed calculation.  Which is what makes it so fast.  

Think of it kind of like if you ever watched the show "How is it made".  A computer would be a designing a robotic arm and programming it to do a pile of various moves to pick up a sticker and place it on a box (slow).  An ASIC would be like having a sticker wheel, where all it does all day is spin around and place a sticker in the same spot every rotation.  (Super fast).  

But once that job is finished, the sticker wheel can not be used to do another task.  It's a sticker wheel.   It can't even be used to put stickers on a different type of package, or to put stickers in a different location on the box.  

The robotic arm can run another program and put stickers anywhere else on the box, a different type of package, or even do a task totally unrelated.  

Because of (random access) memory latency.
See http://en.wikipedia.org/wiki/Random-access_memory#Memory_wall

cpu-only coin are bad, lead to botnet fest, and anyway they always become gpu-coin and then asic minable, is the natural evolution in crypto world

Why "lots of memory takes lots of time"? Will it sound reasonable to use algo other than scrypt, along with "a lot" memory.

Thanks for defining. I didn't understand the differences between them earlier but now I understand. Can you tell me what all things are included in ASIC? I just want to try making one.
Kindly,
       Muhammed Zakhir

Requiring lots of memory is good for CPU friendliness and ASIC resistance.
But using lots of memory takes lots of time, and you want PoW verification to be super fast.

So scrypt is a bad idea.

You want an assymetric PoW where only finding the proof requires a lot of memory,
while verification takes none.

Gigabytes of memory cannot (yet) be put on a self-contained ASIC. At best it will lead to an ASIC that you have to connect to much more expensive memory modules.
Then the already heavily commoditized DRAM market dictates mining costs, preventing centralization
based on access to fabrication technology and cheap power...

No way to make a CPU only coin. Botnets would rape it to death.

So the question can be like that a coin which is particularly designed to emphasize the difference between CPU mining and GPU mining will make CPU mining more efficient. Seems one existing option is the memory-based, Scrypt-N, since CPU (physical memeory in the motherboard) can always use more memory than GPU (built-in memory). From this point of view, a coin making use of algo similar to the Scrypt-N highly relying on memeory could be more effeciently mined by CPU. Correct me if I'm wrong.

hashing is algorithm.
algorithm may be completed even on http://en.wikipedia.org/wiki/Turing_machine

Firstly , the difference between a CPU and a GPU , is that the GPU can process a lot more data and a lot faster , but miss some features of a CPU. A more complete explanation can be found here.

Whenever a PoW system is made , it is primarily designed to be CPU mined.
Later , somebody creates an algorithm that allows GPUs to mine the coin too. This instantly makes CPUs useless and outdated as GPUs are much faster.
An ASIC ( Application Specific Integrated Circuit ) is a device made to only mine the coin and nothing but that coin (and other coins using the same PoW). It is highly optimized as obsolete parts are removed and it can mine the coin even faster , making GPUs outdated.

Each and every single PoW system tries to make it restricted to CPU mining , but as time passes GPU and ASIC mining becomes easier.
Every new PoW is originally a CPU-only coin , it will not boost CPU development , it will only boost development on a method to mine it on a GPU.

Not possible: CPU is just an ASIC.

Not sure if this is off-topic.

Bitcoin is based on SHA-256, initially CPU mining only, then GPU, then ASIC ...
Litecoin is based on Scrypt, initially CPU mining only, then GPU, then ASIC ...

Is there a way to create an algo to prevent ASIC, even GPU?
Mining with CPU only is really what people prefer?? (as can be seen, Bitcoin and Litecoin were targeting CPU at the very beginning)

If we can develop a CPU-only coin, probably we will boost the CPU development? 

Just a thinking.