BitShares PTS (formerly ProtoShares) Mandatory Upgrade & Snapshot Announcement

[Etlase2]

The point of CPU-only is the millions mass your coin can gain in the market since anyone can download and get some coin. Mass leads to network effects, which leads to competing against Bitcoin effectively, given Bitcoin only has 350,000 users.

This is not a well-thought argument. A CPU coin is likely to be dominated by expensive CPUs and/or CPU farms that are available for time purchase. Those with low-end CPUs will be forced to upgrade to remain relevant, but the simpler and perhaps more financially sound option is to purchase the currency in lieu of mining it. This is the *exact same* scenario with GPUs. And it also makes sense economically because if everyone is a miner, no one is demanding to trade fiat for it, and all value must enter the coin via burning electricity.

[1714851309]

1714851309

[1714851309]

1714851309

[1714851309]

1714851309

[AnonyMint]

The bounty isn't for whether or not GPUs can mine it, the bounty is for whether or not there is a non-linear speed increase available without the memory tradeoff. If GPUs end up being faster, big deal, it isn't going to be an order of magnitude faster unless you have an order of magnitude increase in the memory bandwidth or in the memory available. At least, that is the presumption by my understanding.

^
That

Except that is a wrong assumption if I am correct in my understanding.

[AnonyMint]

The point of CPU-only is the millions mass your coin can gain in the market since anyone can download and get some coin. Mass leads to network effects, which leads to competing against Bitcoin effectively, given Bitcoin only has 350,000 users.

This is not a well-thought argument. A CPU coin is likely to be dominated by expensive CPUs and/or CPU farms that are available for time purchase. Those with low-end CPUs will be forced to upgrade to remain relevant, but the simpler and perhaps more financially sound option is to purchase the currency in lieu of mining it. This is the *exact same* scenario with GPUs. And it also makes sense economically because if everyone is a miner, no one is demanding to trade fiat for it, and all value must enter the coin via burning electricity.

I guess you haven't read this:

1. With CPU-only mining, I expect difficulty to exceed any reasonable
level that can be returned on investment, because millions of people will
do it because they easily can (download and click) and they won't correlate
to their insignificant increase in electric bill nor the cost of the PC they
already own applied to their part-time, hobbyist mining.

Meaning there won't be sufficient return for serious miners after the initial launch period. Only the masses will bother once the masses are the majority of miners.  

That was my most clever insight.

Call this the "lottery effect". People play the lottery even though it is a horrible return-on-investment.

[Luckybit]

I can't work with bytemaster because he and I don't agree. He wants to make a socialist coin where everyone receives their fair share just by sitting on their coins, and I would want to make a capitalist coin which doesn't steal (ahem redistribute) value to give to the collective of coin owners. The coin owners are already profiting from the rise in value of the coin.

I do talk with others behind the scenes.

Are you saying he doesn't believe in the survival of the fittest?
He doesn't believe in the market?

That doesn't seem to be what his documents say on his websites. What gives you the idea that its a socialist coin? I never heard of that before.

The public discussions he and I had (in his bitcointalk threads) about how everyone gets what is akin to an interest payment, and how everyone who owns the domain name coin gets a payment. He claims to be Austrian economics based, yet apparently he and I disagree about what that entails. I'd rather not rehash that here. The prior discussions are available. My summary is inaccurate I am sure. Better to go read the prior discussions.

I never knew that. Maybe these coins have inherent value afterall.
How is it socialism though if we had to mine it (work for it) or buy it?

Since when was interest socialism?

The only problem with interest is where does that money come from? I wouldn't say it's socialism because it's more like property owners charging rent, but I suppose you could see it as a sort of tax but isn't that the same as what Bitcoin miners do with transaction fees?

[AnonyMint]

I can't work with bytemaster because he and I don't agree. He wants to make a socialist coin where everyone receives their fair share just by sitting on their coins, and I would want to make a capitalist coin which doesn't steal (ahem redistribute) value to give to the collective of coin owners. The coin owners are already profiting from the rise in value of the coin.

I do talk with others behind the scenes.

Are you saying he doesn't believe in the survival of the fittest?
He doesn't believe in the market?

That doesn't seem to be what his documents say on his websites. What gives you the idea that its a socialist coin? I never heard of that before.

The public discussions he and I had (in his bitcointalk threads) about how everyone gets what is akin to an interest payment, and how everyone who owns the domain name coin gets a payment. He claims to be Austrian economics based, yet apparently he and I disagree about what that entails. I'd rather not rehash that here. The prior discussions are available. My summary is inaccurate I am sure. Better to go read the prior discussions.

I never knew that. Maybe these coins have inherent value afterall.
How is it socialism though if we had to mine it (work for it) or buy it?

Since when was interest socialism?

I will not rehash that discussion here. I don't want to get off on an economics debate here.

[digitalindustry]

The limiting factor on the CPU is not memory bandwidth, rather memory latency which is less than 1 GB per second because the memory latency of main memory (if outside of L3) is several hundred clock cycles. With only 8 hardware hyperthreads, that latency isn't entirely masked away.

Whereas, the GPU can run 1000s of hardware threads (not software threads!) which masks away the latency and hits the memory bandwidth as the limit.

This is why the GPU blows away the CPU.

A hardware thread has its own copy of registers so there is nearly no cost to blocking the thread on memory access, so another thread can run which was blocked and is ready to proceed.

I think an optimized algorithm, either for CPU or GPU is going to need really detailed knowledge of memory latency issues - more knowledge than I have. You might be onto something with this - I don't have enough knowledge in this area to know if or what level of improvement managing memory latency could offer. You'll have to wait for BM's response.

I have that knowledge and have done all the research.

Then use your 30+ years of coding and knowledge , build a miner , give it to NO one, mine with it yourself, gain a monopoly and achieve any objective you would want by owning a disproportionate amount of the entity.

[AnonyMint]

Then use your 30+ years of coding and knowledge , build a miner , give it to NO one, mine with it yourself, gain a monopoly and achieve any objective you would want by owning a disproportionate amount of the entity.

I already explained upthread why I don't have an incentive to do so. There is an opportunity cost to my time. I would rather spend the time creating a better coin. I will earn much more that way, than consuming my limited time writing a miner for a coin I don't believe in.

[Etlase2]

Meaning there won't be sufficient return for serious miners after the initial launch period. Only the masses will bother once the masses are the majority of miners.  

You are banking on the irrationality of the masses. Not necessarily a bad bet, but it is a square peg treatment again that involves wasting energy.

That was my most clever insight.

It is less clever than giving that value away without requiring the token of wasted energy.

[BitThink]

I've been mining around 24 hours with 30hpm. Uptill now I find that I've mined around 10 stale blocks and 0 valid block. Maybe it is because of my network connection? I have only 12 connections.

I think a low hanging optimization we can do is to stop current searching once the network has found a new block.

[AnonyMint]

Meaning there won't be sufficient return for serious miners after the initial launch period. Only the masses will bother once the masses are the majority of miners.  

You are banking on the irrationality of the masses. Not necessarily a bad bet, but it is a square peg treatment again that involves wasting energy.

That was my most clever insight.

It is less clever than giving that value away without requiring the token of wasted energy.

It is not entirely irrational. People spend on hobbies. And they get some coin with this hobby or curiosity. And they might get it anonymously, so it has a value higher than the FX value. Also the future value may justify the ROI. Alas it is a complex dynamic, very difficult to characterize entirely.

It is the best I can think of for PoW; I claim much better than Bitcoin. You know well from our many discussions that I don't believe PoS and variants such as your Decrits can be secure, because I believe they all suffer from the order of the block peer selection not being driven by a random entropy that PoW provides. But I am not claiming that I am omniscient on this, and I wish you the best possible results in proving me wrong. You know well that I respect your intellect.

And note that the mix of ranges of irrationality of the masses is perhaps a gradient in this case, thus it might not be too far from reasonable, thus perhaps one could still obtain virgin coins anonymously from mining by reasonably overpaying.

Also I have suggested that those who are very serious, can install micro-hydropower on a small stream and perhaps they get a reasonable ROI that way.

Please note that the profitability of the initial mining would not be adversely affected by this observation. The mad gold rush would still apply.

[bahamapascal]

OH, that looks bad...so is the code broken? Even thogh difficulty went up the block time is still way too fast 

My calculations are that in the 2016 blocks before the diff adjustment, blocks were coming in at 31 seconds.

Now, in the past 300 blocks, it's 1 every 48 seconds.

The diff adjustment was by a factor of 4, the maximum the code allows for.

If we continue at this pace, we'll have another difficulty adjustment in 50 hours, again by a factor of 4, probably bringing 90 second blocks. Then 4 days later, another 4x difficulty adjustment, which will bring us into line for 5 minute blocks.

Ah, ok...now that spunds good
Lol, though it souds like this coin is going to be fully mined within a couple of week
Simmilar to BTC, hashe power increases faster then blockdifficulty

By the way, I found one more block After 49 hours now, I found 2 blocks (with 17 hpm) 

[tombcoin]

Hi,
for 4 times when i found a block, client got stuck and block got orphaned.
I had to restart client of course it didnt sync anymore.
It happend on 2 different (64bit) PCs. One with Win7, i5 and 16 GB Ram and another with Win2008R2 XEON and 12 GB RAM.
No other stuff running on it. Also with different setgenerate settings (1, 2 and 3)

Any suggestions to avoid this strange behaviour?

Mining for more than 24 hours an still no valid block found.

lg t.

[adoalli]

because too many computer is mining it

[NineLives]

Morning..

I have a question please..

I'm mining 3x Instances and I'm still not peeking 100% usage on 2x Xeon 3.3.  Its on 88%.

1)  Shall I run another instant and peak it?
2)  Or its better to not peak and allow it to work through?

Ta very much.

PS:

Does this look right:

Code:

{
"blocks" : 4800,
"currentblocksize" : 1000,
"currentblocktx" : 0,
"difficulty" : 0.00000382,
"errors" : "",
"generate" : false,
"genproclimit" : 3,
"hashespermin" : 4.32148659,
"pooledtx" : 1,
"testnet" : false
}

[cryptrol]

Morning..

I have a question please..

I'm mining 3x Instances and I'm still not peeking 100% usage on 2x Xeon 3.3.  Its on 88%.

1)  Shall I run another instant and peak it?
2)  Or its better to not peak and allow it to work through?

Are you memory limited ? I was getting 2.3 hpm running a single instance on a i3, so this is pretty bad for a dual xeon.

[Sy]

Morning..

I have a question please..

I'm mining 3x Instances and I'm still not peeking 100% usage on 2x Xeon 3.3.  Its on 88%.

1)  Shall I run another instant and peak it?
2)  Or its better to not peak and allow it to work through?

Are you memory limited ? I was getting 2.3 hpm running a single instance on a i3, so this is pretty bad for a dual xeon.

Xeons seem to perform kinda bad, i get 44 hpm on 32 cores (2.2 ghz each) in 32 threads (getting less with 64) thats 1.375 hpm per thread vs 2.5 hpm per thread on an ivy i3

[cryptrol]

Since the main momentum implementation is so memory intensive, the faster the memory the better the results.
I am testing a configuration with some overclocked memory ATM.

[AnonyMint]

Since the main momentum implementation is so memory intensive, the faster the memory the better the results.
I am testing a configuration with some overclocked memory ATM.

Specifically I believe the CAS latency, yet they may not be the only factor in the random access latency seen from the CPU:

http://en.wikipedia.org/wiki/CAS_latency

[digitalindustry]

Luckybit forget about bitshares , and the simplr principle of interest is misunderstood.

The " lottery" effect is well known , and enters into sociological behavior,  I dont disagree with you Anonymint,  but I doubt you can catch the credit for it.

In the end , lets instead state this plainly :

There is not a massive incentives to rush for that 5k bounty because human behavior states that waiting is potentially more profitable,  if the algorithm has "sudo protection" or the perception,  there may be much larger reward to create a miner and try to monopolize the entity.

Having said that, time and information equals all, the benefits of there being an attempt/attempts will prove or disprove the degree at which the algorithm functions in the market in relation to its competitors.  

Bitshares the idea has proven Adam Smith ultimately correct.

The perception of this idea led to this algorithm.  

Having said that , this is an algorithm,  and just that, I will only state that your " lottry" principal is not fleshed out and missing some aspects.

[NineLives]

Morning..

I have a question please..

I'm mining 3x Instances and I'm still not peeking 100% usage on 2x Xeon 3.3.  Its on 88%.

1)  Shall I run another instant and peak it?
2)  Or its better to not peak and allow it to work through?

Are you memory limited ? I was getting 2.3 hpm running a single instance on a i3, so this is pretty bad for a dual xeon.

Xeons seem to perform kinda bad, i get 44 hpm on 32 cores (2.2 ghz each) in 32 threads (getting less with 64) thats 1.375 hpm per thread vs 2.5 hpm per thread on an ivy i3

Interesting..

Its a Xeon 5470 (which is a bit old) 4 cores per CPU.  24GB DDR.

Its a new install (Windows .  Two instances running.  Should it really be allot more?