[XPM] Working on a GPU miner for Primecoin

[bcp19]

I'm having a hard time figuring out which direction price is going to shoot when this GPU miner is released.  Will the price tank because there will be an abundance of new coins on the market (which somebody will surely dump for a shitty low price), or will price increase with the new interest in the coin that this release will generate?

This is probably a crystal ball question, but I'm sure somebody in here has some insight.  Either way, it would most likely be completely useless for me to donate if I only have one 7950, right?

Depends on speed.  We know sieving is running about 25x, but how much of the CPU time was used for sieving and how much for fermat testing?  Let x = CPU sieve time and y = CPU Fermat time.  If x=y (50/50 split) then GPU = 2% for 52% of normal time = 92.3% speed increase.  If x=9y (90/10 split) then GPU = 3.6% for 13.6% of normal time = 635% speed increase.  if 9x=y (10/90 split) GPU = .4% fo 90.4% of normal time = 10.6% speed increase.  These are extremely rough figures, communication times between GPU/CPU will also apply.  Maybe he'll solve Fermat testing on the GPU and get a full 2500% speed increase.  Short answer: too many current unknowns.

[redphlegm]

Block reward and difficulty are recalculated every block but there is a limit on the speed at which reward changes happen. We saw this with the release (and subsequent quick release) of hpx. I personally expect that there will be a short-term flood of coins for perhaps the first week (with block reward honing in on the new difficulty level) and then an eventual stabilization and rise from the low. The question is, though, if current prices are actually rational. Any depression due to the release of GPU may translate to long-term effects simply because it "should" have been lower than .006x anyway.

The short answer is "nobody really knows but some people are probably going to make a good bit of dough on this".

[oxfeeefeee]

Depends on speed.  We know sieving is running about 25x, but how much of the CPU time was used for sieving and how much for fermat testing?  Let x = CPU sieve time and y = CPU Fermat time.  If x=y (50/50 split) then GPU = 2% for 52% of normal time = 92.3% speed increase.  If x=9y (90/10 split) then GPU = 3.6% for 13.6% of normal time = 635% speed increase.  if 9x=y (10/90 split) GPU = .4% fo 90.4% of normal time = 10.6% speed increase.  These are extremely rough figures, communication times between GPU/CPU will also apply.  Maybe he'll solve Fermat testing on the GPU and get a full 2500% speed increase.  Short answer: too many current unknowns.

I did a similar calculation, but then Koooooj@reddit pointed out that I was wrong. Because with a much powerful sieve, CPU gets less test to do for fixed number of output.

for example (with made up numbers):
before: 10000Numbers ---sieve--> 100Numbers ---test---> 1Number
after  : 10000Numbers ---sieve--> 10Numbers   ---test---> 1Number

this way, even we don't speed up the execution of Fermat test, the test time is down to 10%.

[bcp19]

Depends on speed.  We know sieving is running about 25x, but how much of the CPU time was used for sieving and how much for fermat testing?  Let x = CPU sieve time and y = CPU Fermat time.  If x=y (50/50 split) then GPU = 2% for 52% of normal time = 92.3% speed increase.  If x=9y (90/10 split) then GPU = 3.6% for 13.6% of normal time = 635% speed increase.  if 9x=y (10/90 split) GPU = .4% fo 90.4% of normal time = 10.6% speed increase.  These are extremely rough figures, communication times between GPU/CPU will also apply.  Maybe he'll solve Fermat testing on the GPU and get a full 2500% speed increase.  Short answer: too many current unknowns.

I did a similar calculation, but then Koooooj@reddit pointed out that I was wrong. Because with a much powerful sieve, CPU gets less test to do for fixed number of output.

for example (with made up numbers):
before: 10000Numbers ---sieve--> 100Numbers ---test---> 1Number
after  : 10000Numbers ---sieve--> 10Numbers   ---test---> 1Number

this way, even we don't speed up the execution of Fermat test, the test time is down to 10%.

But he's comparing apples to oranges.  If sieveing speed is increased 25x, then you need different parameters to sieve deeper.  Also, where does he get these numbers?  10000 seems extremely small, 100 is 1%, so we are to believe the CPU sieve is 99% effective?  The GPU sieve is 99.9% effective?  There's a lot of questions that need to be asked here:
What is he using for h₀?
What is he using for c?
What is he using for e? 
What is the first sieving prime? 
What is he using for P_s?

I'm sorry, but those number he's tossing out look made up, with no proof to back them up. 

[oxfeeefeee]

Depends on speed.  We know sieving is running about 25x, but how much of the CPU time was used for sieving and how much for fermat testing?  Let x = CPU sieve time and y = CPU Fermat time.  If x=y (50/50 split) then GPU = 2% for 52% of normal time = 92.3% speed increase.  If x=9y (90/10 split) then GPU = 3.6% for 13.6% of normal time = 635% speed increase.  if 9x=y (10/90 split) GPU = .4% fo 90.4% of normal time = 10.6% speed increase.  These are extremely rough figures, communication times between GPU/CPU will also apply.  Maybe he'll solve Fermat testing on the GPU and get a full 2500% speed increase.  Short answer: too many current unknowns.

I did a similar calculation, but then Koooooj@reddit pointed out that I was wrong. Because with a much powerful sieve, CPU gets less test to do for fixed number of output.

for example (with made up numbers):
before: 10000Numbers ---sieve--> 100Numbers ---test---> 1Number
after  : 10000Numbers ---sieve--> 10Numbers   ---test---> 1Number

this way, even we don't speed up the execution of Fermat test, the test time is down to 10%.

But he's comparing apples to oranges.  If sieveing speed is increased 25x, then you need different parameters to sieve deeper.  Also, where does he get these numbers?  10000 seems extremely small, 100 is 1%, so we are to believe the CPU sieve is 99% effective?  The GPU sieve is 99.9% effective?  There's a lot of questions that need to be asked here:
What is he using for h₀?
What is he using for c?
What is he using for e? 
What is the first sieving prime? 
What is he using for P_s?

I'm sorry, but those number he's tossing out look made up, with no proof to back them up. 

And I made up those number just to explain the idea, not him.
This is just another possibility, GPU could still be as slow as 2x

[alatvian]

Please, let us use it at least week, but better month, before it goes public. I'm one of the investors.

I'm pretty sure "investor" is by far not the proper term for your status.  Even VC "investors" know that the majority of their "investments" will turn up bust.

[r3animation]

Closed Beta has been postponed?

[eCoinomist]

Interesting progress thus far.

I'm still weary of donation at this stage. But once it's out, and (providing no one else has released anything else before it - we will know this if there is sudden difficulty spike in the network, currently is at 9.27), some decent donation amount will come from me.

[itod]

I'm having a hard time figuring out which direction price is going to shoot when this GPU miner is released.  Will the price tank because there will be an abundance of new coins on the market (which somebody will surely dump for a shitty low price), or will price increase with the new interest in the coin that this release will generate?

It seams rather obvious that GPU miners will be able to produce XPM cheaper than VPS farms used today. People are closely monitoring the profitability ratio of altcoins/bitcoin, and if it is rather high (like 200% for some coins) they mine altcoins and trade them for BTC. As more GPUs get into the game, more VPSs will get out. Current price is guarded by the price of VPS hosting, when that wall breaks and cheap GPU generated XPM flood the market I suppose price has to go down.

[woodrake]

I'm having a hard time figuring out which direction price is going to shoot when this GPU miner is released.  Will the price tank because there will be an abundance of new coins on the market (which somebody will surely dump for a shitty low price), or will price increase with the new interest in the coin that this release will generate?

It seams rather obvious that GPU miners will be able to produce XPM cheaper than VPS farms used today. People are closely monitoring the profitability ratio of altcoins/bitcoin, and if it is rather high (like 200% for some coins) they mine altcoins and trade them for BTC. As more GPUs get into the game, more VPSs will get out. Current price is guarded by the price of VPS hosting, when that wall breaks and cheap GPU generated XPM flood the market I suppose price has to go down.

This is a persistent misconception; all cryptocoins are generated at a static rate. The mining capacity/hashrate thrown at it has no real impact on the price of the coin, because mining hashrate has no net impact on generation rate - that is why difficulty goes up as you apply more computational power (total network hashrate).

The only thing affecting XPM price now is speculation. The only thing which will ensure longevity of its value is market support (ie. a XPM economy).

Kate.

[bidji29]

This is a persistent misconception; all cryptocoins are generated at a static rate. The mining capacity/hashrate thrown at it has no real impact on the price of the coin, because mining hashrate has no net impact on generation rate - that is why difficulty goes up as you apply more computational power (total network hashrate).

The only thing affecting XPM price now is speculation. The only thing which will ensure longevity of its value is market support (ie. a XPM economy).

Kate.

XPM is not generated at a static rate. Because the reward goes down as difficulty goes up. Right now there is 18k coin a day. But with GPU miner out, it can easily go 10K/day or even 5k.

[fluffypony]

This is a persistent misconception; all cryptocoins are generated at a static rate. The mining capacity/hashrate thrown at it has no real impact on the price of the coin, because mining hashrate has no net impact on generation rate - that is why difficulty goes up as you apply more computational power (total network hashrate).

The only thing affecting XPM price now is speculation. The only thing which will ensure longevity of its value is market support (ie. a XPM economy).

Kate.

ALL cryptocoins?

[Dsfyu]

People keep forgetting how the reward for mining a block is generated. With primecoin it is anything but a static reward and within the span of a day it can fluctuate from about 11.75 to 11.55 (roughly the change over the last day). Primecoin block rewards dynamically change with the difficulty - the equation for it is 999/difficulty^2 if I remember correctly, so as the difficulty goes up the reward per block plummets.

[bcp19]

I'm having a hard time figuring out which direction price is going to shoot when this GPU miner is released.  Will the price tank because there will be an abundance of new coins on the market (which somebody will surely dump for a shitty low price), or will price increase with the new interest in the coin that this release will generate?

It seams rather obvious that GPU miners will be able to produce XPM cheaper than VPS farms used today. People are closely monitoring the profitability ratio of altcoins/bitcoin, and if it is rather high (like 200% for some coins) they mine altcoins and trade them for BTC. As more GPUs get into the game, more VPSs will get out. Current price is guarded by the price of VPS hosting, when that wall breaks and cheap GPU generated XPM flood the market I suppose price has to go down.

This is a persistent misconception; all cryptocoins are generated at a static rate. The mining capacity/hashrate thrown at it has no real impact on the price of the coin, because mining hashrate has no net impact on generation rate - that is why difficulty goes up as you apply more computational power (total network hashrate).

The only thing affecting XPM price now is speculation. The only thing which will ensure longevity of its value is market support (ie. a XPM economy).

Kate.

Ah yes, the "All coins are created equal" speech.  They're not.  All you have to do is look at http://cryptometer.org/primecoin_90_day_charts.html and you will see how wrong you are.  Day 9, 893% of target blocks generated.  Over 190,000 coins were minted day 9, or roughly 6.6% of the current supply.  Personally, I think Sunny made a mistake in his difficulty adjustment.  Humongeous bot nets were thrown at this coin and it took over 10 days before the difficulty caught up enough to get the target block rate under 120%.  24 days * 1440 blocks per day = 34,560.  We're up over 97,000!  If the GPU is anywhere near as powerful as people think, we'll be over 200,000 blocks for sure and likely over 300,000 before difficulty catches up.

[bcp19]

Depends on speed.  We know sieving is running about 25x, but how much of the CPU time was used for sieving and how much for fermat testing?  Let x = CPU sieve time and y = CPU Fermat time.  If x=y (50/50 split) then GPU = 2% for 52% of normal time = 92.3% speed increase.  If x=9y (90/10 split) then GPU = 3.6% for 13.6% of normal time = 635% speed increase.  if 9x=y (10/90 split) GPU = .4% fo 90.4% of normal time = 10.6% speed increase.  These are extremely rough figures, communication times between GPU/CPU will also apply.  Maybe he'll solve Fermat testing on the GPU and get a full 2500% speed increase.  Short answer: too many current unknowns.

I did a similar calculation, but then Koooooj@reddit pointed out that I was wrong. Because with a much powerful sieve, CPU gets less test to do for fixed number of output.

for example (with made up numbers):
before: 10000Numbers ---sieve--> 100Numbers ---test---> 1Number
after  : 10000Numbers ---sieve--> 10Numbers   ---test---> 1Number

this way, even we don't speed up the execution of Fermat test, the test time is down to 10%.

But he's comparing apples to oranges.  If sieveing speed is increased 25x, then you need different parameters to sieve deeper.  Also, where does he get these numbers?  10000 seems extremely small, 100 is 1%, so we are to believe the CPU sieve is 99% effective?  The GPU sieve is 99.9% effective?  There's a lot of questions that need to be asked here:
What is he using for h₀?
What is he using for c?
What is he using for e?  
What is the first sieving prime?  
What is he using for P_s?

I'm sorry, but those number he's tossing out look made up, with no proof to back them up.  

And I made up those number just to explain the idea, not him.
This is just another possibility, GPU could still be as slow as 2x

Ah yes, the old scientific "I pulled these numbers out of my ass" theorum.  Great, thanks for playing... NOT!

[usahero]

Many of those coins minted on day 9 were dumped at prices 0.002-0.003. If you didn't buy then, you missed your chance.

[hasle2]

Depends on speed.  We know sieving is running about 25x, but how much of the CPU time was used for sieving and how much for fermat testing?  Let x = CPU sieve time and y = CPU Fermat time.  If x=y (50/50 split) then GPU = 2% for 52% of normal time = 92.3% speed increase.  If x=9y (90/10 split) then GPU = 3.6% for 13.6% of normal time = 635% speed increase.  if 9x=y (10/90 split) GPU = .4% fo 90.4% of normal time = 10.6% speed increase.  These are extremely rough figures, communication times between GPU/CPU will also apply.  Maybe he'll solve Fermat testing on the GPU and get a full 2500% speed increase.  Short answer: too many current unknowns.

I did a similar calculation, but then Koooooj@reddit pointed out that I was wrong. Because with a much powerful sieve, CPU gets less test to do for fixed number of output.

for example (with made up numbers):
before: 10000Numbers ---sieve--> 100Numbers ---test---> 1Number
after  : 10000Numbers ---sieve--> 10Numbers   ---test---> 1Number

this way, even we don't speed up the execution of Fermat test, the test time is down to 10%.

But he's comparing apples to oranges.  If sieveing speed is increased 25x, then you need different parameters to sieve deeper.  Also, where does he get these numbers?  10000 seems extremely small, 100 is 1%, so we are to believe the CPU sieve is 99% effective?  The GPU sieve is 99.9% effective?  There's a lot of questions that need to be asked here:
What is he using for h₀?
What is he using for c?
What is he using for e?  
What is the first sieving prime?  
What is he using for P_s?

I'm sorry, but those number he's tossing out look made up, with no proof to back them up.  

And I made up those number just to explain the idea, not him.
This is just another possibility, GPU could still be as slow as 2x

Ah yes, the old scientific "I pulled these numbers out of my ass" theorum.  Great, thanks for playing... NOT!

On another note, he is comparing one gpu chip on a 6990 to one core on a Phenom II X6 1055T. The 1055T is a relatively weak processor and has 6 cores. So that 25x improvement turns into 25/6 = 4.2. Dividing again by 1.5 because the 1055T is weak we get 4.2/1.5 = 2.8. Not bad for this early in the game but also not a huge game changer either.

As bcp19 has been saying it also depends a lot on how much difference a fast sieve makes and how effectively the overall algorithm can make use of the gpu sieving.  From what I have seen from posts by Sunny and Mikaelh it sounds like sieving is in the ballpark of 50% of the work. Only time will tell how it plays out on the gpu.

Regardless of all that if mtrly releases a working gpu miner that can match or beat an i7 3770 in the next week I will be impressed.

The 1055T compared to some other chips:

http://www.cpubenchmark.net/cpu.php?cpu=AMD+Phenom+II+X6+1055T

[itod]

I'm having a hard time figuring out which direction price is going to shoot when this GPU miner is released.  Will the price tank because there will be an abundance of new coins on the market (which somebody will surely dump for a shitty low price), or will price increase with the new interest in the coin that this release will generate?

It seams rather obvious that GPU miners will be able to produce XPM cheaper than VPS farms used today. People are closely monitoring the profitability ratio of altcoins/bitcoin, and if it is rather high (like 200% for some coins) they mine altcoins and trade them for BTC. As more GPUs get into the game, more VPSs will get out. Current price is guarded by the price of VPS hosting, when that wall breaks and cheap GPU generated XPM flood the market I suppose price has to go down.

This is a persistent misconception; all cryptocoins are generated at a static rate. The mining capacity/hashrate thrown at it has no real impact on the price of the coin, because mining hashrate has no net impact on generation rate - that is why difficulty goes up as you apply more computational power (total network hashrate).

The only thing affecting XPM price now is speculation. The only thing which will ensure longevity of its value is market support (ie. a XPM economy).

Kate.

Kate, you've completely missed my point, and said couple of false claims. I won't argue about the things you are wrong, but I'll underline my point once more: If you got hold of something at high price you can't afford to sell it low and keep your head above the water. If you paid (relatively) high price for VPS hosting, you have to sell XPS high to cover your costs and earn something. But, if you've generated your coins on the cheap with your GPU, you can sell them much cheaper and be happy. So, I'll be very surprised if price of XPM want go down fast immediately after the GPU miner is released to the masses and VPS farms are shut down.

[bitwarrior]

I wonder if the beta version of the gpu miner has been already distributed to a few donators..there is a recent increase in difficulty "9.32315743" as of last checked. Or it might be the result of the recent HP9 roll-out. Or it might be the combination of the two.

[dudeguy]

I wonder if the beta version of the gpu miner has been already distributed to a few donators..there is a recent increase in difficulty "9.32315743" as of last checked. Or it might be the result of the recent HP9 roll-out. Or it might be the combination of the two.

I would expect it is the result of both and possibly some other circumstances. It may not seem like a massive jump in difficulty but from what I gather we don't have many more new miners, so any new jump in difficulty should indicate better mining software the vast majority of the time.