Topic: [ANN][DASH] Dash (dash.org) | First Self-Funding Self-Governing Crypto Currency - page 6930. (Read 9723782 times)

Yep..


Personally I don't believe in the concept of ASIC-proof. Commercial ASICs maybe, but the real issue is not a commercial designer (as far as I'm concerned) but NSA. A commercial designer has a few million dollars worth of investment capacity with doubtful rewards so miners can be "ok" knowing that they won't deal with an ASIC implementation anytime soon. But an NSA contractor can get billions so that the NSA can stay ahead of the game. USA has a full-spectrum dominance strategy and that means no cost is too big in order to be ahead of the game.

If I were the NSA I would order an ASIC platform with modular design that you can fit like 20 different chips on it for every possible hash. So if a coin gets generated that has a combination of hashA, hashB...hashZ, then I place the corresponding asic-coprocessors on the modular board and I've just got myself an ASIC-raper for that coin.

Everything is balanced in the end of the day. Coins like bitcoin which have multiple asics mining for it by individuals are more safe from a 51% NSA control because they have the same tools at their disposal to counter it, but mining is virtually dead for cpu/gpu.

Scrypts may be GPU only for the time being but if NSA tomorrow reveals an ASIC+ram combo that rapes scrypt, then scrypt will be in trouble until the ordinary miners can get their hands on similar equipment.

Complicated coins with like 10 hashes are much more mining-resistant for commercial asic solutions but at the same time gives the "agencies" a far larger headstart due to their budget to create custom implementations that the private sector won't have for years.

They already have, and they want to open the door to GPU miners, because 90 percent of the crypto miners use GPUs.  If they want this coin to last more than a few weeks this needs to be done. 

The move to GPU is a natural progression, but, we need one that is accessable to everyone, just to level the playing field, i dont have any gpus that would be do able to use, but this is the path that lies ahead for darkcoin.





on another node,

im not able to update and help out as i usually do, but will try to pop in once in a while, i have 2 tasks that are the most important right now, work, and lotterymining

There may be some tiny specific function, that is difficult to calculate on GPU but is not a problem for CPU.

Surely developers know better what they are doing.

Also, profitability would depend on initial costs - the price of the CPU and the other required equipment (specific motherboard, etc).

One way to move "ASIC-proof" a currency might also be to link it's hashing to a task that people already perform. Imagine that you somehow increased your hash value by playing Angry Birds... Or serving as an IRC hub and your hash value was linked to your user count.

Obviously, these are conceptual, not actual, examples

I think this is an algorithmic task. Let the developers say their word.

Yeah, at that level ok. These are larger scale operations that not many people are involved in. For the individual with his one or two PCs its a different matter altogether.


I am not certain of the GPU-concentration dynamic, except through aggregated individual power in multipools. GPU farms are difficult to build and manage, so GPU power is mostly decentralized to simple users who mine with a couple of cards. I believe GPU concentration will never reach the same level as CPU concentration (data centers, etc).

But even so the whole issue of the developers deciding cpu or gpu is kind of irrelevant as it's not really their choice. It's more about the choice of certain programmers who will or won't make a gpu-implementation and whether they will share this or not.

A coin creator can, at most, make the life difficult for a gpu chip, asic-designer or quantum-computer designer by selecting (during the coin creation) hashes which are hard to speed up in the aforementioned hardware. But now we are past the coin creation stage, so...

Yes, but we must have the technology to strike back on someone or group that is using a gpu based miner.  I'm not saying release the miner, but keep it stock piled.

WMD

+100

Don't lose this niche. The type of distribution it encourages also fits the concept behind DRK much more nicely.

As I said, entry level for efficient GPU mining is much higher than for CPU. It's possible to rent a couple of servers for CPU mining at affordable price or simply use available computers. No need to buy it all. Initial cost of GPU farm is pretty large (and yes, it includes CPU as well) and it must be placed somewhere. I wouldn't count power as important parameter as some regions have pretty low prices.

Also, as discussed previously, we are interested in mining power to be more distributed while with GPUs it gets more concentrated. I believe this would lead to price instability which, in its turn, suppresses effective use of the coin. Anyone, except speculators, wants this? The more stable the price, the more rapid adoption of the coin would be.

That is why I believe that it is essential to keep it CPU only and prevent GPU mining as much as possible.

Yah, I'm just bummed because if this is so, they're not sharing   But why should they?

I really don't get the GPU hatred and the "for the few" argument. Most GPUs are cheaper than high-end i5/i7 etc (not to mention xeons) and used cards like 5850s/6870s with decent hashrates, can be found for 70-90 euros or near the 100 dollar mark. Their only problem is their power consumption and heat dissipation.

ASICs are really for the few (and Quantum Computers for even fewer). GPUs and CPUs, well, these are not. And it's better if everyone can have access to a gpu miner than someone mining selfishly with it.

As for the profitability, it will all depend on how accelerated it all becomes and whether the GPU is better suited for doing other coins, like scrypt. It's also a question of watts consumed per khash. If a cpu can consume, say, 30w and give 20% of what a 300w GPU can give, then the CPU is more power efficient and is well in the game with whatever hashrate it produces.

Argh, crap! Better make it unprofitable to mine on GPUs! Heed my call!

I would think that would cost a fortune.  I think it's more likely that someone or several people have figured out the gpu miner and are keeping it to themselves. 

Cheers! much appreciated, good to know

I thought that was rented mining time from a network..? I know that with 120 8-core linux instances I can do 30 mh/s so that would mean they have like 500 8-core cpu instances. or something crazy. Either that or GPU mining.

Well, some people must have a GPU miner and aren't sharing  over at dark.coinmine.pl, they have over 15 users doing up to 140 mh/s

sigh

lol I think you are right, and I'm in the same boat!

This might force me to learn C++

Awesome InternetApe!

Yes, that's the right one Matt, you're all set!

Hammerhedd, I think it has something to do with the x-ll algorithm not being properly retrofitted, LOL.  I'm no programmer, so I don't know how to expain

Any idea why it might compile and run correctly (seemingly) but not have any hashes accepted? I did get one slightly odd error "autogen.sh: command not found" but it seemed to run after that.