BitShares PTS (formerly ProtoShares) Mandatory Upgrade & Snapshot Announcement - page 54.

AnonyMint

hero member

Activity: 518

Merit: 521

Quote from: testz on November 07, 2013, 01:56:14 AM

Sorry, but I can't not to answer to @AnonyMint

:
If you experienced programmer you should know that "the devil's in the details" and you should trust only proven working solutions. If you don't have time, left it, someone will make the working GPU miner for this algo and get his a well-deserved award.

PS: If you so busy, probably you get salary and you don't need this reward. IMHO

Someone can go implement what I have explained. If they want to send me a tip to be considerate that is good honor. Correct I don't need the money. I am looking to see the honor (or ethics) of the people involved.

Another factor is the requirements of the bounty. They don't allow for the GPU speedup I have explained. It isn't about reducing the memory on the GPU to less than that of the CPU, it is about running more threads on the same memory.

testz

legendary

Activity: 1764

Merit: 1018

Sorry, but I can't not to answer to @AnonyMint

:
If you experienced programmer you should know that "the devil's in the details" and you should trust only proven working solutions. If you don't have time, left it, someone will make the working GPU miner for this algo and get his a well-deserved award.

PS: If you so busy, probably you get salary and you don't need this reward. IMHO

AnonyMint

hero member

Activity: 518

Merit: 521

Quote from: FreeTrade on November 07, 2013, 01:51:18 AM

Quote from: AnonyMint on November 07, 2013, 01:40:56 AM

The GPU can compute 1000s of lookups on that memory in parallel, thus obviating the random memory latency to be bound at the superior memory bandwidth of the GPU.

I'm not an expert in this area, but as far as I know the GPU has a wider bus, so can do more memory access (x4), but is slower so can do less (x0.25). So I think you could get comparable speeds out of a high end GPU as for a high end processor.

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.

Luckybit

hero member

Activity: 714

Merit: 510

Quote from: Stinky_Pete on November 06, 2013, 09:59:04 PM

Quote from: Luckybit on November 06, 2013, 08:45:39 PM

I believe someone is already GPU mining it and planning to dump thousands of coins.

And your evidence for this is what?

People kept believing that Primecoin was being GPU mined, without any evidence. And still no-one has admitted to being able to mine it on GPUs, the two well publicised efforts both ending in failure.

Primecoin wasn't made by a corporate entity with funding. I'm not saying Bitshares is a bad idea, I like the idea. I'm saying this new algorithm seemed to be untested and their parameters I immediately knew were right.

But hey, if you manage to get some coins thats fantastic. I just don't see those coins going for a lot right now. I think the smart person will be the one who mines now and holds onto them until they convert them to Bitshares.

Do not sell a single coin. Each coin may be worth something decent but not for the next 6 months. Bitshares needs time to develop the decentralized exchange which will be powered by this coin.

FreeTrade

legendary

Activity: 1428

Merit: 1030

Quote from: AnonyMint on November 07, 2013, 01:40:56 AM

GPUs aren't limited to 1GB of memory, many have closer to 6 - 10GB now.

Yes - it's not about the amount of memory, GPU's can compare there with the current implementation.

Quote from: AnonyMint on November 07, 2013, 01:40:56 AM

The salient issue (unless I've misunderstood the algorithm) is that he asserts a CPU can use a hash table but a GPU can't, thus he asserts the GPU needs N ^ 2 memory. I don't see why the GPU can't use the same amount of memory as the CPU?

I think hashtables are a distraction from the point - which is . . .

Quote from: AnonyMint on November 07, 2013, 01:40:56 AM

The GPU can compute 1000s of lookups on that memory in parallel, thus obviating the random memory latency to be bound at the superior memory bandwidth of the GPU.

I'm not an expert in this area, but as far as I know the GPU has a wider bus, so can do more memory access (x4), but is slower so can do less (x0.25). So I think you could get comparable speeds out of a high end GPU as for a high end processor.

AnonyMint

hero member

Activity: 518

Merit: 521

Quote from: FreeTrade on November 07, 2013, 01:37:47 AM

Quote from: AnonyMint on November 07, 2013, 12:57:56 AM

bytemaster still hasn't responded further to my logic that the GPU can mine much faster (specifically he hasn't stated why he believes a hash table can't also be implemented on the GPU). He said he is sick.

GPU's will be able to generate the SHA512 hashes quicker, no doubt. But the GPU resistance relies on the GPU not being able to store them. So if your GPU is generating 10GB/s but can only store 1GB/s, there's no advantage to an i7 that generates 1GB/s and stores 1GB/s. Whether that's through the medium of a hashmap or not is not important.

ByteMaster really wants to know if there is a way to beat this with GPU's - that's why he's put up the bounty.

GPUs aren't limited to 1GB of memory, many have closer to 6 - 10GB now.

The salient issue (unless I've misunderstood the algorithm) is that he asserts a CPU can use a hash table but a GPU can't, thus he asserts the GPU needs N ^ 2 memory. I don't see why the GPU can't use the same amount of memory as the CPU? The GPU can compute 1000s of lookups on that memory in parallel, thus obviating the random memory latency to be bound at the superior memory bandwidth of the GPU.

FreeTrade

legendary

Activity: 1428

Merit: 1030

Quote from: AnonyMint on November 07, 2013, 12:57:56 AM

bytemaster still hasn't responded further to my logic that the GPU can mine much faster (specifically he hasn't stated why he believes a hash table can't also be implemented on the GPU). He said he is sick.

GPU's will be able to generate the SHA512 hashes quicker, no doubt. But the GPU resistance relies on the GPU not being able to store them. So if your GPU is generating 10GB/s but can only store 1GB/s, there's no advantage to an i7 that generates 1GB/s and stores 1GB/s. Whether that's through the medium of a hashmap or not is not important.

ByteMaster really wants to know if there is a way to beat this with GPU's - that's why he's put up the bounty.

AnonyMint

hero member

Activity: 518

Merit: 521

Quote from: digitalindustry on November 07, 2013, 01:08:12 AM

Anonymint I'd say two things :

1. If you can implement your theory or get someone to you will have much more credibility.

I don't need to implement to know I am correct. And I don't expect enough future value in this coin to justify implementing. Or perhaps we can rephrase that as I have limited time and I would prefer to focus it on the type of coin I want to see. And I don't trust bytemaster to pay out the $5000. And $5000 is peanuts and isn't really worth my time any way.

Quote from: digitalindustry on November 07, 2013, 01:08:12 AM

2. If you can build a better PoW algorithm or think you can , work with these people or others .

I don't need to work with anyone to create a coin, if I decide to do so. I have 30 years of programming experience with several million user commercial successes.

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.

Quote from: digitalindustry on November 07, 2013, 01:08:12 AM

Now , does a CPU algorithm matter , no I think you will find that it's the dispersion Anonymint , notice these reports of the variances between cloud services and single systems.

If this is real an consistent I think you will find that is important.

Clouds don't usually have GPUs, so it would seem to me that CPU-only is what drove the interest of servers to mine here. If a GPU is doing 100x more blocks than a Xeon, then why bother mining with your CPU.

Variance is an orthogonal issue. Having pools built in from the beginning is an interesting suggestion.

digitalindustry

hero member

Activity: 798

Merit: 1000

‘Try to be nice’

Quote from: AnonyMint on November 07, 2013, 12:57:56 AM

Quote from: billotronic on November 06, 2013, 09:25:09 PM

Quote from: Luckybit on November 06, 2013, 08:45:39 PM

Quote from: ?? on ??

Quote from: Luckybit on November 06, 2013, 08:36:11 PM

Its time to stop mining. This coin is not profitable to mine.

There isn't even an established market price.
But we know the drill, mine the coin while people figure out it exists / how to mine it.

Whatever the price is, you wont get very main coins. You'll be lucky if you can get 100 and that is around half a Bitcoin. It's profitable now but it wont be once difficulty goes up.

Quote from: AnonyMint on November 06, 2013, 08:37:41 PM

Quote from: Luckybit on November 06, 2013, 08:36:11 PM

Its time to stop mining. This coin is not profitable to mine.

How can you make that conclusion when you don't know the future value of the coin?

Doesn't it largely depend on whether the coin will be truly CPU-only or not?

Or are you saying you can buy the coin for less than the cost to mine it?

I believe someone is already GPU mining it and planning to dump thousands of coins. I would not count on it being profitable to mine. Where are all the coins going? Seems no one has more than a few hundred. The people who get thousands will be the professional mining companies like Ciphermine.

Agreed, much if nothing else than because that is how the crypto scene works these days. That being said, it is at least refreshing to see something different algo wise even if the 'cpu-only' claim stands up so long.

Let's not kid ourselves, this is just another scam coin BUT might be worth a fun ride in the end

bytemaster still hasn't responded to my logic that the GPU can mine much faster. He said he is sick.

Having a new algorithm is not necessarily good, if it is not based on well vetted cryptography.

My analysis is that Litecoin's Scrypt parameters are better than this new PoW. At least Litecoin reduced the speedup on the GPU to roughly 10 at most, compared to roughly 100 for Bitcoin.

And I know why and I know how to make a CPU-only coin, where the GPU will be 10 times slower.

P.S. when you say it is a scam coin, are you implying the creator is scamming or just that the algorithm doesn't appear to be working as intending?

What I really want to know is if a CPU-only coin is what everybody really wants? I believe so, because I see Litecoin's value has declined since it has been mined by GPUs after an initial runup in price before that.

Anonymint I'd say two things :

1. If you can implement your theory or get someone to you will have much more credibility.

2. If you can build a better PoW algorithm or think you can , work with these people or others .

Now , does a CPU algorithm matter , no I think you will find that it's the dispersion Anonymint , notice these reports of the variances between cloud services and single systems.

If this is real an consistent I think you will find that is important.

digitalindustry

hero member

Activity: 798

Merit: 1000

‘Try to be nice’

Quote from: hiddenvalleyranch on November 07, 2013, 12:34:34 AM

Yeah I've found 5 blocks on my i5-4670k, only one block on about 10 cloud servers.

Interesting and impressive.

AnonyMint

hero member

Activity: 518

Merit: 521

Quote from: billotronic on November 06, 2013, 09:25:09 PM

Quote from: Luckybit on November 06, 2013, 08:45:39 PM

Quote from: ?? on ??

Quote from: Luckybit on November 06, 2013, 08:36:11 PM

Its time to stop mining. This coin is not profitable to mine.

There isn't even an established market price.
But we know the drill, mine the coin while people figure out it exists / how to mine it.

Whatever the price is, you wont get very main coins. You'll be lucky if you can get 100 and that is around half a Bitcoin. It's profitable now but it wont be once difficulty goes up.

Quote from: AnonyMint on November 06, 2013, 08:37:41 PM

Quote from: Luckybit on November 06, 2013, 08:36:11 PM

Its time to stop mining. This coin is not profitable to mine.

How can you make that conclusion when you don't know the future value of the coin?

Doesn't it largely depend on whether the coin will be truly CPU-only or not?

Or are you saying you can buy the coin for less than the cost to mine it?

I believe someone is already GPU mining it and planning to dump thousands of coins. I would not count on it being profitable to mine. Where are all the coins going? Seems no one has more than a few hundred. The people who get thousands will be the professional mining companies like Ciphermine.

Agreed, much if nothing else than because that is how the crypto scene works these days. That being said, it is at least refreshing to see something different algo wise even if the 'cpu-only' claim stands up so long.

Let's not kid ourselves, this is just another scam coin BUT might be worth a fun ride in the end

bytemaster still hasn't responded further to my logic that the GPU can mine much faster (specifically he hasn't stated why he believes a hash table can't also be implemented on the GPU). He said he is sick.

Having a new algorithm is not necessarily good, if it is not based on well vetted cryptography.

My analysis is that Litecoin's Scrypt parameters are better than this new PoW. At least Litecoin reduced the speedup on the GPU to roughly 10 at most, compared to roughly 100 for Bitcoin.

And I know why and I know how to make a CPU-only coin, where the GPU will be 10 times slower.

P.S. when you say it is a scam coin, are you implying the creator is scamming or just that the algorithm doesn't appear to be working as intending? I assume the latter.

What I really want to know is if a CPU-only coin is what everybody really wants? I believe so, because I see Litecoin's value has declined since it has been mined by GPUs after an initial runup in price before that.

hiddenvalleyranch

member

Activity: 103

Merit: 10

Yeah I've found 5 blocks on my i5-4670k, only one block on about 10 cloud servers.

wildraven

member

Activity: 100

Merit: 10

Quote from: bahamapascal on November 06, 2013, 06:05:18 PM

Quote from: wildraven on November 06, 2013, 05:55:51 PM

Quote from: bahamapascal on November 06, 2013, 05:52:52 PM

Quote from: wildraven on November 06, 2013, 05:10:18 PM

Quote from: bahamapascal on November 06, 2013, 04:59:51 PM

Anyone on here using setgenerate 4 and found blocks? I am starting to wonder if setgenerate 4 cant find any blocks, which might be the reason I only found one block so far (and its good possibel I had setgenerate 3 when I found it).

setgenerate 4 works for me

Hmm, ok...and you have alrady found blocks with it?

yes I have, but from time to time it crashes and I have to restart it but it works...

Hmm, then I realy don´t know why I am not geting any blocks ;(
Thanks for the info

after the diff change I have not found anything

digitalindustry

hero member

Activity: 798

Merit: 1000

‘Try to be nice’

Quote from: Hilux74 on November 07, 2013, 12:17:37 AM

Lots of luck with this coin makes it fun. 3 blocks today on my system with a Celeron G1610 and 0 blocks on the system with the i7-2600k.

Ok , now that's interesting .

Hilux74

legendary

Activity: 912

Merit: 1000

Lots of luck with this coin makes it fun. 3 blocks today on my system with a Celeron G1610 and 0 blocks on the system with the i7-2600k.

ryen123

sr. member

Activity: 292

Merit: 250

Quote from: Bitcoin2Cheval on November 06, 2013, 10:48:23 PM

do you have fresh nodes valid?
i'm in trouble to make connections
thanks

My currently connected peers.

180.183.204.204
146.185.173.206
54.219.166.105
188.165.216.59
182.236.161.243
192.241.240.44
162.243.34.76
123.110.21.18

Bitcoin2Cheval

newbie

Activity: 19

Merit: 0

do you have fresh nodes valid?
i'm in trouble to make connections
thanks

Buratino

legendary

Activity: 1151

Merit: 1003

Quote from: rwessels on November 06, 2013, 10:02:52 PM

I have had 43 cores mining this coin for 36 hours now and have only found 2 blocks....

It shows that PTS has botnet and cloud mining resistance. PTS is decentralized more than enough.

FreeTrade

legendary

Activity: 1428

Merit: 1030

Quote from: bahamapascal on November 06, 2013, 06:48:41 PM

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

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.

rwessels

member

Activity: 102

Merit: 10

I have had 43 cores mining this coin for 36 hours now and have only found 2 blocks....

Topic: BitShares PTS (formerly ProtoShares) Mandatory Upgrade & Snapshot Announcement - page 54. (Read 218428 times)