So the stake reward is 100% per block now? 
Topic: DEAD - page 44. (Read 79202 times)
And got results:
Code:
nSubsidy1 = -10780497
nSubsidy2 = 1000000000
Guess the GC code had an integer overflow in it
Like I explained, this is what happens when you use too much math and go over a 9 quintillion satoshi calculation. That's 9 with 18 zeroes. Although that's pretty hard to believe, apparently it's happening on multiple coins now too. Surprise!*sets off fireworks*
Coins ops are the worst programmers.
Satoshi didn't expect such calculations
I'm sure he did. That's what a debugger is for. Although no one could stand looking at his code until about 2000 people got their hands on it and made it actually human readable. I'm sure he didn't depend on doing calculations in the quintillions though, because he knows his math.
And got results:
Code:
nSubsidy1 = -10780497
nSubsidy2 = 1000000000
Guess the GC code had an integer overflow in it
Like I explained, this is what happens when you use too much math and go over a 9 quintillion satoshi calculation. That's 9 with 18 zeroes. Although that's pretty hard to believe, apparently it's happening on multiple coins now too. Surprise!*sets off fireworks*
Coins ops are the worst programmers.
Satoshi didn't expect such calculations
And got results:
Code:
nSubsidy1 = -10780497
nSubsidy2 = 1000000000
Guess the GC code had an integer overflow in it Wink Like I explained, this is what happens when you use too much math and go over a 9 quintillion satoshi calculation. That's 9 with 18 zeroes. Although that's pretty hard to believe, apparently it's happening on multiple coins now too. Surprise!
*sets off fireworks*
Coins ops are the worst programmers. Kiss
Edit: Or not. The two examples you put on the bottom of the quote did not match what you put in code. However, the existing code might not be fit for a large pile of coin. Or any amount. Who knows apparently
That's not the issue. The value is to be returned in satoshis (multiplied by COIN). I thought I made that perfectly clear.
So, is nSubsidy1 and nSubsidy2 equal?Can't tell if trolling or you can't figure it out =p
int64_t nRewardCoinYear;
nRewardCoinYear = MAX_MINT_PROOF_OF_STAKE; // 36.5 * COIN (main.h)
int64_t nCoinAge = 100 * COIN * 1;
int64_t nSubsidy = nCoinAge * nRewardCoinYear / 365 / COIN;
nSubsidy = 100 * COIN * 1 * 36.5 * COIN / 365 / COIN;
nSubsidy = 100 * COIN * 1 * 36.5 / 365;
Is it correct? Value is multiplied by COIN? nSubsidy in satoshi format?
ps
So, is nSubsidy1 and nSubsidy2 equal?
Sorry, thought you were totally trolling before. Yes, the bottom two nSubsidy are equal.
Thanks.
I wrote a little code to check it:
Code:
#include
#ifndef WIN32
#include
#endif
int main(void) {
int64_t COIN = 100000000;
int64_t MAX_MINT_PROOF_OF_STAKE = 36.5 * COIN;
int64_t nCoinAge = 100 * COIN * 1; // actually 100 coins one day aged
int64_t nRewardCoinYear = MAX_MINT_PROOF_OF_STAKE;
int64_t nSubsidy1 = nCoinAge * nRewardCoinYear / 365 / COIN;
int64_t nSubsidy2 = nCoinAge / COIN * nRewardCoinYear / 365;
printf("nSubsidy1= %ld\n", nSubsidy1);
printf("nSubsidy2= %ld\n", nSubsidy2);
}
And got results:
Code:
nSubsidy1 = -10780497
nSubsidy2 = 1000000000
So, results are equal for human, but computer thinks different )
Quote from: tryphe
In the first statement(one above the bottom) the earlier *COIN will be "reversed" by the later /COIN. But there's no reason to have that many operations. nSubsidy is the satoshi value of the block, so since age is the only real factor here. Usually (for neatness and speed) you want to convert to satoshis last and not keep switching between 1s(coin) and 10,000,000s(satoshi). You want something neat and readable like this:
Dividing by 10 is an easy way to multiply by 0.1 (daily age modifier towards reward) and to avoid using floating points. Then you multiply by COIN to calculate the satoshi value. This is actually exactly the same math as the existing algorithmn, but simplified, and another *COIN added. Here is the original logic and why we are dividing by 10:
nRewardCoinYear / 365 / COIN == 36.5 * COIN / 365 / COIN == 0.1
There is NO need to do all of those operations other than if you're reading the code and have no idea what the 0.1 value means, or if you are trying to do math in years instead of days, like above. Now it looks pretty retarded, doesn't it?
Keep in mind that's untested so definitely don't use that until you test it, and it might be missing something(I'm sure it is)
Hope that makes a bit more sense.
Code:
const int64_t MAX_PROOF_OF_STAKE_DAILY_DIVISOR = 10; // 10% daily
nSubsidy = nCoinAge / MAX_PROOF_OF_STAKE_DAILY_DIVISOR * COIN;
Dividing by 10 is an easy way to multiply by 0.1 (daily age modifier towards reward) and to avoid using floating points. Then you multiply by COIN to calculate the satoshi value. This is actually exactly the same math as the existing algorithmn, but simplified, and another *COIN added. Here is the original logic and why we are dividing by 10:
nRewardCoinYear / 365 / COIN == 36.5 * COIN / 365 / COIN == 0.1
There is NO need to do all of those operations other than if you're reading the code and have no idea what the 0.1 value means, or if you are trying to do math in years instead of days, like above. Now it looks pretty retarded, doesn't it?
Keep in mind that's untested so definitely don't use that until you test it, and it might be missing something(I'm sure it is)
Hope that makes a bit more sense.
Sure.
You can try this to make the interest rate higher with hard fork at block 15000
change this in main.cpp
to this
and put this in main.h where the other static const int64_t are defined. put it under them.
change this in main.cpp
Code:
// miner's coin stake reward based on coin age spent (coin-days)
int64_t GetProofOfStakeReward(int64_t nCoinAge, int64_t nFees)
{
int64_t nRewardCoinYear;
nRewardCoinYear = MAX_MINT_PROOF_OF_STAKE;
int64_t nSubsidy = nCoinAge * nRewardCoinYear / 365 / COIN;
if (fDebug && GetBoolArg("-printcreation"))
printf("GetProofOfStakeReward(): create=%s nCoinAge=%"PRId64"\n", FormatMoney(nSubsidy).c_str(), nCoinAge);
return nSubsidy + nFees;
}
Code:
// miner's coin stake reward based on coin age spent (coin-days)
int64_t GetProofOfStakeReward(int64_t nCoinAge, int64_t nFees)
{
int64_t nRewardCoinYear;
nRewardCoinYear = MAX_MINT_PROOF_OF_STAKE;
int64_t biginterest;
biginterest = BIG_INTEREST;
if(pindexBest->nHeight > 15000)
{
int64_t nSubsidy = nCoinAge * biginterest * 33 / (365 * 33 + 8);
}
else
{
int64_t nSubsidy = nCoinAge * nRewardCoinYear / 365 / COIN;
}
if (fDebug && GetBoolArg("-printcreation"))
printf("GetProofOfStakeReward(): create=%s nCoinAge=%"PRId64"\n", FormatMoney(nSubsidy).c_str(), nCoinAge);
return nSubsidy + nFees;
}
and put this in main.h where the other static const int64_t are defined. put it under them.
Code:
static const int64_t BIG_INTEREST = 3650 * CENT; // 3650% per year
That's not the issue. The value is to be returned in satoshis (multiplied by COIN). I thought I made that perfectly clear.
So, is nSubsidy1 and nSubsidy2 equal?Can't tell if trolling or you can't figure it out =p
int64_t nRewardCoinYear;
nRewardCoinYear = MAX_MINT_PROOF_OF_STAKE; // 36.5 * COIN (main.h)
int64_t nCoinAge = 100 * COIN * 1;
int64_t nSubsidy = nCoinAge * nRewardCoinYear / 365 / COIN;
nSubsidy = 100 * COIN * 1 * 36.5 * COIN / 365 / COIN;
nSubsidy = 100 * COIN * 1 * 36.5 / 365;
Is it correct? Value is multiplied by COIN? nSubsidy in satoshi format?
ps
So, is nSubsidy1 and nSubsidy2 equal?
Sorry, thought you were totally trolling before. Yes, the bottom two nSubsidy are equal. In the first statement(one above the bottom) the earlier *COIN will be "reversed" by the later /COIN. But there's no reason to have that many operations. nSubsidy is the satoshi value of the block, so since age is the only real factor here(for this coin at least), I'll leave it really simple. Usually (for neatness and speed) you want to convert to satoshis last and not keep switching between 1s(coin) and 10,000,000s(satoshi). You want something neat and readable like this:
Code:
const int64_t MAX_PROOF_OF_STAKE_DAILY_DIVISOR = 10; // 10% daily
nSubsidy = nCoinAge / MAX_PROOF_OF_STAKE_DAILY_DIVISOR * COIN;
Dividing by 10 is an easy way to multiply by 0.1 (daily age modifier towards reward) and to avoid using floating points. Then you multiply by COIN to calculate the satoshi value. This is actually exactly the same math as the existing algorithmn, but simplified, and another *COIN added. Here is the original logic and why we are dividing by 10:
nRewardCoinYear / 365 / COIN == 36.5 * COIN / 365 / COIN == 0.1
There is NO need to do all of those operations other than if you're reading the code and have no idea what the 0.1 value means, or if you are trying to do math in years instead of days, like above. Now it looks pretty retarded, doesn't it?
Keep in mind that's untested so definitely don't use that until you test it, and it might be missing something(I'm sure it is)
Hope that makes a bit more sense.
Just come back to check on thread, it is kind of embarrassing, if you invest you win you lose, if things don't work out then it really is just one of those things, the state of altcoins over recent months is shameful, lots of scams breed no confidence in coins that have credence, for example Sapience is a coin which I am in and am enthusiastic about as it has some legitimacy behind it and "could" be a game changer, ILT however was always a gamble, at the moment POS is not working for those that thought it would grant them 10% daily reward but it is working so you have the option of either sticking with it or getting out. All these stupid posts are not doing anything apart from wasting people's time, if you are unhappy then sell, if you have sold then why are you even interested, if you are still a holder then hold, go with what you think but let's be grown up and stop all this childish foolishness, it's just embarrassing!
the amount of stupidity in this thread is more impressive then the coin
well lets hope so...
Fiat = scam. Better go buy some gold.
Gold is overpriced scam
Try silver then.
Or aluminum. Fiat = scam. Better go buy some gold.
Gold is overpriced scam
shadow runner = scam
Fiat = scam. Better go buy some gold.
Gold is overpriced scam
Jump to: