[XPM] Working on a GPU miner for Primecoin

[itod]

@itod

I don`t agree at all, you don`t get block time is 1min and now we have let say 20 000miners or less, with GPU we will have a lot more so more miners sharing same block numbers and less coins, and you do a math!

Also GPU use a lot power sure it will be faster but reward in coins will be small

Only pure retard would sell XPM now, or big shark dumping price and talking crap to buy low

I don't know if I'm a retard, I've converted *all* my XPM, that I've mined from the beginning, to BTC two days ago when I realized GPU miner is closer than I thought, don't have a single XPM now. I calculated and reasoned what I've posted above, if I'm wrong I will pay, if I'm right i will have the option to buy them back if I want to for the fraction of the BTC I've got when I sold them. We'll see who is right or wrong in a month or two.

[vinne81]

@itod

I don`t agree at all, you don`t get block time is 1min and now we have let say 20 000miners or less, with GPU we will have a lot more so more miners sharing same block numbers and less coins, and you do a math!

Also GPU use a lot power sure it will be faster but reward in coins will be small

Only pure retard would sell XPM now, or big shark dumping price and talking crap to buy low

Once GPU miners start, hash goes up, reward goes down. That seems like the only certainty.

I love all your conclusions: price MUST go up, price MUST go down once GPU is there. Why? Imo it can go both ways, or nothing might happen. Since HP9 was released, hash has increased a LOT. Price didn't change.

I'm selling every XPM I mine, I guess we'll see if I'm a pure retard or not.

To those who are so certain price will skyrocket, why are you not investing and buying every XPM you can get your hands on?

[vinne81]

@itod

I don`t agree at all, you don`t get block time is 1min and now we have let say 20 000miners or less, with GPU we will have a lot more so more miners sharing same block numbers and less coins, and you do a math!

Also GPU use a lot power sure it will be faster but reward in coins will be small

Only pure retard would sell XPM now, or big shark dumping price and talking crap to buy low

I don't know if I'm a retard, I've converted *all* my XPM, that I've mined from the beginning, to BTC two days ago when I realized GPU miner is closer than I thought, don't have a single XPM now. I calculated and reasoned what I've posted above, if I'm wrong I will pay, if I'm right i will have the option to buy them back if I want to for the fraction of the BTC I've got when I sold them. We'll see who is right or wrong in a month or two.

+1

[bcp19]

But if GPU mining is relatively cheap, more GPU owners will jump on it, making the difficulty higher, giving less reward, removing the relative cheapness from GPU mining.

Higher difficulty and less reward per CPU core gets irrelevant when nobody mines with CPU. If GPU mining is 25x more efficient (as was claimed somewhere in this thread), structure of the mining public will change. No more few giant miners with big expensive rented VPS farms, instead large majority of smaller miners with already owned GPU cards with only electricity costs to cover. Wouldn't that make mining cheaper, allowing everybody to sell XPM cheaper? This seams obvious to me that GPU mining will drop the XPM price significantly, maybe I'm missing some factors.

As the processing power on the network goes up block reward will go down. What happens to price of XPM when block reward is less than 5 XPM per block? or less than 2?
Just having a more efficient solominer had a big impact on the block reward. GPU mining will have a huge impact.

Correct me if im wrong but Primecoin should keep ~1minute block time with decreasing block reward as the hashing power increases. This means less XPM per day for way more hashing power. Most other coins dont change the block reward and as such have the same reward per day no matter the hash power on the network. Right now around 16,000XPM are created every day. When GPU miners come online and the block reward drops to 5xpm per block we will see about 7,000XPM per day. Less coins generated a day will make the price go up.

By my observations Primecoin can do nothing but go up as the hash power increases.

Am I wrong?

Yes, i think you are wrong. Block reward will not decrease not nearly as quick as you think, we are mislead by quick decrease from the first days of the coin to the current level of 9.4 difficulty. Let me explain why: link between the block reward (number of XPM/block) is exponential, and is described by the formula:
reward = 999 / difficulty ^2
So, for block reward to decrease to the levels you described:

5 XMP/block:
difficulty = sqrt(999/5) = 14.13506278

2 XMP/block:
difficulty = sqrt(999/2) = 22.34949663

See, wee will not reach similar levels of difficultly in the foreseeable future, I' may dare to say we will not reach them ever with only GPU mining without ASIC and BTC like adoption, no matter how super efficient that GPU miner will be.

So, no significant decrease of block reward is expected, just the re-distribution of the mined coins from big-mining, rich, VPS expensively generated coins to the small-miner, cheaply produced GPU generated coins. If you produce something with big expenses, you have to sell it high, if you produce it cheaply you can afford to sell it low. I'm pretty sure GPU mining will drop the XPM price significantly.

You are correct, it would take a lot of effort to push the block reward down to 5.  But while you noted the 25x speed increase posted earlier, that was only for thr sieve portion of the coding.  According to Sunny, the sieving and fermat testing were pretty much split 50/50 for time, so that 25x speed increase could end up only being 2-4x at this time.  Admittedly, you could change parameters to sieve deeper and get different results, but that would take a lot of trial and error.  Unlike SHA and Scrypt where a change is virtually immediate, the finding of prime chains is more random and can take days of testing to really see results.

[itod]

You are correct, it would take a lot of effort to push the block reward down to 5.  But while you noted the 25x speed increase posted earlier, that was only for thr sieve portion of the coding.  According to Sunny, the sieving and fermat testing were pretty much split 50/50 for time, so that 25x speed increase could end up only being 2-4x at this time.  Admittedly, you could change parameters to sieve deeper and get different results, but that would take a lot of trial and error.  Unlike SHA and Scrypt where a change is virtually immediate, the finding of prime chains is more random and can take days of testing to really see results.

Yes, but only sieve part is highly efficient already not because fermat testing is not implementable on the GPU, but because GPU developer hadn't enough time to finish it, he just did sieve part first. I've understood that in fermat testing gains may be as big as in sieving if not higher.

Nevertheless, such high gains wouldn't be needed at all to kick the VPS farms out of the game. I don't have the VPS farm to testify how much profit they have per $ invested at current difficulty, but I'm sure even more moderate (than 25x) GPU efficiency gains will bring VPS farms in red.

[itod]

I don't know if I'm a retard, I've converted *all* my XPM, that I've mined from the beginning, to BTC two days ago when I realized GPU miner is closer than I thought, don't have a single XPM now. I calculated and reasoned what I've posted above, if I'm wrong I will pay, if I'm right i will have the option to buy them back if I want to for the fraction of the BTC I've got when I sold them. We'll see who is right or wrong in a month or two.

I believe you forgot to calculate popularity, and increased overall economic activity of the coin due to gpu miner adoption.  I think you're going to miss out with your current position, but I also hope XPM does go down for a bit so I can buy more  .

That's good argument, if demand goes higher than it may bring the price up no matter how cheap the supply is. But this may be only temporarily, because anyone can just GPU mine them selves instead of buying XPM. Also, XPM can be as popular as LTC for instance, but there is only one king in this coin came and that is BTC, all altcoins are in the market for BTC, and that market dictates the price, not the popularity.

[woodrake]

Nevertheless, such high gains wouldn't be needed at all to kick the VPS farms out of the game. I don't have the VPS farm to testify how much profit they have per $ invested at current difficulty, but I'm sure even more moderate (than 25x) GPU efficiency gains will bring VPS farms in red.

It is a close call actually. If you own the VPS farm like I do (Memset.com) then you're only paying for the small slice of the host hardware (~$0.02/day/quad-Xeon) plus the additional power (20 Watts so ~$0.07/day/quad-xeon). I'm currently getting about 1BTC/day/50 quad-xeon hosts, or ~$2/day/quad-Xeon - it would stop being profitable after a >20-fold productivity reduction roughly. If you're paying the going rate for your VMs (even if you're using a cheap host like Memset, hint hint ) of, say, $10/month then they will become unprofitable quite soon (>6-fold productivity reduction/difficulty increase).

It also depends on the uptake by GPU miners. If many decide to switch across then yes difficulty will severely curtail the XPM profits, but since it is not technically trivial to implement there is a strong argument for there being slow and/or limited adoption of XPM GPU mining. BTW, where you can you view the total XPM network hashrate/PPS? With that, and ideally historical data, I could do some forecasting on how long XPM will be profitable to mine with VM farms.

As an aside, someone commented about selling XPM. My view as a miner is that I'm not in the business of cryptocurrency speculation so I sell everything at the current price to BTC. I do then have bots trading BTC/USD to get some upside but alt-coins have too low liquidity and too high volatility to be sensible to trade at this time (unless you've really got nothing else to do). For me its all about automation: human time is very expensive!!

Kate.

[itod]

Nevertheless, such high gains wouldn't be needed at all to kick the VPS farms out of the game. I don't have the VPS farm to testify how much profit they have per $ invested at current difficulty, but I'm sure even more moderate (than 25x) GPU efficiency gains will bring VPS farms in red.

It is a close call actually. If you own the VPS farm like I do (Memset.com) then you're only paying for the small slice of the host hardware (~$0.02/day/quad-Xeon) plus the additional power (20 Watts so ~$0.07/day/quad-xeon). I'm currently getting about 1BTC/day/50 quad-xeon hosts, or ~$2/day/quad-Xeon - it would stop being profitable after a >20-fold productivity reduction roughly. If you're paying the going rate for your VMs (even if you're using a cheap host like Memset, hint hint ) of, say, $10/month then they will become unprofitable quite soon (>6-fold productivity reduction/difficulty increase).

It also depends on the uptake by GPU miners. If many decide to switch across then yes difficulty will severely curtail the XPM profits, but since it is not technically trivial to implement there is a strong argument for there being slow and/or limited adoption of XPM GPU mining. BTW, where you can you view the total XPM network hashrate/PPS? With that, and ideally historical data, I could do some forecasting on how long XPM will be profitable to mine with VM farms.

As an aside, someone commented about selling XPM. My view as a miner is that I'm not in the business of cryptocurrency speculation so I sell everything at the current price to BTC. I do then have bots trading BTC/USD to get some upside but alt-coins have too low liquidity and too high volatility to be sensible to trade at this time (unless you've really got nothing else to do). For me its all about automation: human time is very expensive!!

Kate.

Thank you Kate for this great insight, I haven't thought VPS mining is still so highly profitable, it's amazing it could withstand 6 to 20-fold productivity reduction and still be OK. Guess in that case we may have GPU/VPS mix-up much longer than I thought, it's hard to expect GPU miner will give > 20x more reward right from the beginning.

To answer your question: there is no hashrate in XPM, only PPS and similar measurements that can not be exactly deduced from difficulty. So you can not see those figures anywhere.

[hasle2]

I don't think that 25x number means what some people think it means. It is the ratio between the speed of one 6990 chip (with 1536 stream processors/cores) and ONE core of an AMD Phenom II X6 1055T. The Phenom II X6 1055T has 6 cores so the 6990 chip (it has 2 but is a very expensive card) is 25 / 6 = 4.17 times faster overall.

The 1055T is also a kind of crappy processor so it's not a fair comparison to what most people are using. It is about 1/2 the speed of an Intel core i7 4770: 4.17 / 2 = 2.09. So the sieve part of the miner is about twice as fast as an i7 4770.

The sieve part is also much easier to speed up than the fermat test because the sieve uses simple arithmetic but the prime test needs exponentiation, which is hard to parallelize on a gpu.

[FreedomCoin]

so a gpu miner has been released for XPM but not posted in this thread?

Possibly a hoax with virus?

[smolen]

I don't think that 25x number means what some people think it means. It is the ratio between the speed of one 6990 chip (with 1536 stream processors/cores) and ONE core of an AMD Phenom II X6 1055T. The Phenom II X6 1055T has 6 cores so the 6990 chip (it has 2 but is a very expensive card) is 25 / 6 = 4.17 times faster overall.

The 1055T is also a kind of crappy processor so it's not a fair comparison to what most people are using. It is about 1/2 the speed of an Intel core i7 4770: 4.17 / 2 = 2.09. So the sieve part of the miner is about twice as fast as an i7 4770.

One GPU "shader" is 32 bit wide, Phenom II is 128 bit, i7 is 256 bit

[Miki77]

so a gpu miner has been released for XPM but not posted in this thread?

Possibly a hoax with virus?

Think twice before you say something.

[ReCat]

GPU miner is out and people are trying to keep it a secret?

[BazkieBumpercar]

Nah it's not out yet, I think. It's in alpha/beta stage, as far as I know.

[FreedomCoin]

so a gpu miner has been released for XPM but not posted in this thread?

Possibly a hoax with virus?

Think twice before you say something.

Maybe you should do the same before posting worthless comments. :-D

[Miki77]

so a gpu miner has been released for XPM but not posted in this thread?

Possibly a hoax with virus?

Think twice before you say something.

Maybe you should do the same before posting worthless comments. :-D

Maybe. You should wrote:

so a gpu miner has been released for XPM, but not by mtrlt?

Possibly a hoax with virus?

[pyromaniac]

So where is OP of this thread?

[A L I E N]

So where is OP of this thread?

Well less than 24hrs ago he said:

For the last two days, I've had some real-life related stuff take priority, I haven't gotten much work done on the miner. I should have more time in a couple of days.

[tbly13]

phree BTCBTCBTCBTC Pl0x Ty 

[litecoin_messiah]

So where is OP of this thread?

Well less than 24hrs ago he said:

For the last two days, I've had some real-life related stuff take priority, I haven't gotten much work done on the miner. I should have more time in a couple of days.

Haha, so this is not "Real-Life".

He decides to take $1,000's and then enjoy his new lifestyle before even prioritizing the community who is actually having high hopes for him...


Now he can make another bs story to premine or delay the release.