[XPM] Working on a GPU miner for Primecoin

[bcp19]

Where can u find the stats, on DIFF,Block reward  & network hash rate

There is no network hash rate, block difficulty can be found by going tot eh debug window and typing getdifficulty  Block rewards can be figured by taking 999/dif^2

[hasle2]

A list of exchanges and their current rates can be found here:

http://www.cryptocoincharts.info/#jump-xpm-btc

I am working on a primecoin charts and stats page. A zoomable chart of the difficulty since block 1 can be found here:

http://192.241.170.170/

The bottom bar can zoom by moving the sliders on either side.

I plan to add dynamic data loading, so it only grabs data enough to make a chart and not all of it, automatic real time data updating soon, and charts on different aspects.

[tinnvec]

Where can u find the stats, on DIFF,Block reward  & network hash rate

Not hash rate (really hard to determine for primecoin), but everything else: http://cryptometer.org/primecoin_90_day_charts.html

[bitwarrior]

I am working on an OpenCL GPU miner for Primecoin. I fully intend to release it open-source and for free when it's ready.

Progress reports: 1 2 3 4 5 6 7 8 9

Where do I download it?

It's not ready yet. I have a working GPU proof-of-concept, but it's not any faster than straight-up CPU mining. Donate 1 BTC or more and get early beta access a week before others! The address is in the bottom of this post. Thanks!

When will it be released?

If all goes well, there should be a closed beta on July 25th 29th okay I underestimated the amount of work a bit, give me a week! The official release will be after a week or two of beta testing.

Why should I trust you?

I am the creator of the GPU miner Reaper. I was the first person to show that GPU mining Litecoins (i.e. scrypt) is possible and also a lot more efficient than on the CPU. I've made my own GPU miners for many coins, including Bitcoin, Litecoin and Yacoin, although not all of them have been released to the public. The Primecoin GPU miner will be free and open-source.

I am also a university math student, and I have made my own implementation of the Elliptic Curve Digital Signature Algorithm used by Bitcoin (and all its forks) to make transactions. From scratch. (Yes, that includes my own bignum library.) I know a lot about algebra, prime numbers and such.

Why not CUDA?

CUDA is not cross-platform. It only works on Nvidia cards, but AMD cards are the ones with the best performance. OpenCL works on all GPUs.

What operating systems will the miner support?

Windows and Linux.

How fast is the miner going to be?

It is very difficult to say at this point. Some people have/had a bounty for a CUDA miner that is twice the speed of the CPU miner. I think that's definitely achievable.

Donate 1 BTC or more and get early beta access!

To fund the development of the miner, I would appreciate it if you could donate a sum of BTC, it will motivate me to work harder and faster, and thus you'll be able to enjoy the miner sooner. As an additional bonus, every donator of 1 BTC or more gets early access to the beta version at least a week before the miner is officially released.

Donation address: 1FKfgFQeZCpjGDn7ocgj26nk4K2TNiVJyD

After donating, please PM me the transaction ID, or paste it to this thread, so that I know who has donated.

+1

Any Updates?

[bidji29]

Someone close to mlrlt said the sieve part is now 50* quicker. (than a core of 1055T )

We don't know about the fermat test but a good news is the High Performance CPU miner by mikaelh use by default 70% of his time to do the sievepart and only 30% for the fermat test. And i use 76%, because i get slighly better result.


So even if the fermat is not as quick than the sievepart, we should still got a very good performance increase because most of the time is use to do the sieve part.

[dudeguy]

Someone close to mlrlt said the sieve part is now 50* quicker. (than a core of 1055T )

We don't know about the fermat test but a good news is the High Performance CPU miner by mikaelh use by default 70% of his time to do the sievepart and only 30% for the fermat test. And i use 76%, because i get slighly better result.


So even if the fermat is not as quick than the sievepart, we should still got a very good performance increase because most of the time is use to do the sieve part.

I'm also hearing from a few people that VPS mining will continue to be viable after the anticipated Mlrlt miner. That should skyrocket the difficulty and lower the block reward. Don't quote me on it. All the "50X faster" stuff is embellished or misunderstood. We may see 50X faster improvements in some areas, but as to total miner effectiveness "50X" is way off. 

[craslovell]

my GPUs are shuddering with excitement.

[Lytse]

Guess the miner will be 4x faster when the CPU is supported by a GPU that does the sieves. 

[yager]

We should have a gpu miner speed guessing game with some xpm up for a prize

[wantrepreneur]

having just seen this thread now;

no wonder XPM is on fire..

[r3animation]

having just seen this thread now;

no wonder XPM is on fire..

Maybe you should read the thread a bit more.

[matt4054]

Maybe you should read the thread a bit more.

Now this is pessimistic

I wonder how much the mining market affects the price. I would tend to say (much) higher difficulty leads to higher prices, until a technology shift with difficulty not catching up overfloods the market. But that is only if miners don't trust their coin and dump it, because if it succeeds, the winning strategy is HOLDING the coins and selling some of them later at higher price, when the difficulty has caught up with the technology shift, and reached equilibrium for reasonable profitability. DUMPING at low price during the "flood" is only if you don't trust Primecoin.

tl;dr: only dump your GPU mined coins at low price if you think XPM has no future!

[Tuck Fheman]

It hit a high on cryptsy at 0.009949.  Close enough...

-tb-

2013-07-19 - Sell  XPM/BTC   0.01550000

Edit : I should have read the convo above more. I see now you were not talking about all-time.

[wantrepreneur]

even without the gpu-release;

XPM is in an early enough stage to expect more botnets to join the fun sooner or later

[matt4054]

even without the gpu-release;

XPM is in an early enough stage to expect more botnets to join the fun sooner or later

This is something higher difficulty is good for.

[akspa]

Beware: The 'beta' account was created 3 days ago. Probably malware.

[reb0rn21]

come on...
This is from trusty old member:
Name:   ZcZcoin
Posts:   2
Activity:   2
Position:   Jr. Member
Date Registered:   August 06, 2013, 05:10:33 PM

[matt4054]

Beta:  https://rapidshare.com/#!download|987p4|2591061454|PrimecoindGPU-beta.rar|398|0|0|1|referer-3EE0380616B2DC14740D8ED685CB37E5

Maybe the beta of the malware for the future Primecoin botnet?

It probably won't GPU mine though... only CPU mine for someone else's profit. If not worse.

[woodrake]

mtrlt has gone a bit quiet, which is mildly troubling for those of us supporting the project. News update would be appreciated.

Kate.

[tacotime]

mtrlt has gone a bit quiet, which is mildly troubling for those of us supporting the project. News update would be appreciated.

Kate.

He's busy throwing Casascius bitcoins at strippers

(I'm sure we'll hear from him soon)