Pages:
Author

Topic: [ANN][GAP] Gapcoin - Prime Gap Search - New Math Algo - CPU / GPU - Zero Premine - page 99. (Read 287695 times)

legendary
Activity: 1246
Merit: 1000
We are currently testing a stratum server on port 3385 for those that can build from source.

https://github.com/noncepool/GapMiner
Code:
[2014-11-07 00:27:47] can not parse server response
legendary
Activity: 1008
Merit: 1000
We are currently testing a stratum server on port 3385 for those that can build from source.

https://github.com/noncepool/GapMiner
+1Who compiles it
full member
Activity: 149
Merit: 100
We are currently testing a stratum server on port 3385 for those that can build from source.

https://github.com/noncepool/GapMiner
sr. member
Activity: 280
Merit: 250
Thanks to riecoin my gmp is already custom compiled and tuned.  I only mentioned testnet as the diffidently was high and was increasing at a steady rate.

Interesting that you get some performance out of SIMD with the code as is.  One of the first things I did was to compile with "-ftree-vectorize -mavx2 -ftree-vectorizer-verbose=5" to see what auto-vectorising found.  It didn't find much (any?) to vectorise as most of the size of the loops aren't know at compile time (there are trick you can use however) or a non-uniform step being used.

Regards,

PS.  The fact that GCC couldn't vectorise what should be very conducive to vectorising is a good avenue to work on for speed-ups.

--
bsunau7

Its not only the vector extensions that are added, its switching between march=native to everything up to core-avx2. And O2 to O3. I dont know what exactly did the trick, has to be something thats added here.

member
Activity: 114
Merit: 10
Anyone else been keeping an eye on testnet?

Difficulty is rising pretty fast, which means some serious mining from a very small base is happening.

Regards,

--
bsunau7

I see a difficulty below 17. Thats <100k pps total.

Let me show you how to optimize the miner:


Thanks to riecoin my gmp is already custom compiled and tuned.  I only mentioned testnet as the diffidently was high and was increasing at a steady rate.

Interesting that you get some performance out of SIMD with the code as is.  One of the first things I did was to compile with "-ftree-vectorize -mavx2 -ftree-vectorizer-verbose=5" to see what auto-vectorising found.  It didn't find much (any?) to vectorise as most of the size of the loops aren't know at compile time (there are trick you can use however) or a non-uniform step being used.

Regards,

PS.  The fact that GCC couldn't vectorise what should be very conducive to vectorising is a good avenue to work on for speed-ups.

--
bsunau7
legendary
Activity: 2646
Merit: 1722
https://youtu.be/DsAVx0u9Cw4 ... Dr. WHO < KLF
Herewith, a basic guide for pool mining Gapcoin in the cloud - http://gapcoin.grn.cc/

~ Serving some text ad links across various sites to get some traffic to the guide. N.B. Its just a free domain on some free hosting.
hero member
Activity: 938
Merit: 1000
Anyone else been keeping an eye on testnet?

Difficulty is rising pretty fast, which means some serious mining from a very small base is happening.

Regards,

--
bsunau7

I see a difficulty below 17. Thats <100k pps total.

Let me show you how to optimize the miner:

My I7-4770k with 8 threads, stock-miner:
pps: 31821 / 31160  10g/h 4473 / 3429  15g/h 0 / 0

Now go on and build your own libgmp from source. When done, run the tune utility for optimum results. Copy your final libgmp.so.10 to /usr/local/lib/.

Now we're at:
pps: 34039 / 34041  10g/h 6077 / 4479  15g/h 0 / 0

Optimizing sieve size:
open src/main.cpp

line 252:
  1048576);

to
   8 * 1048576);


and line 256
  500000);

to
  15000000);

make sure to also change the shift, otherwise you wont be able to submit most blocks! You could specify it on command line, but this is safer:

line 128:
  uint16_t shift = (opts->has_shift() ?  atoi(opts->get_shift().c_str()) : 20);

change to:
  uint16_t shift = (opts->has_shift() ?  atoi(opts->get_shift().c_str()) : 23);


Now its mining at
pps: 38276 / 37068  10g/h 5156 / 5265  15g/h 0 / 0


Lets also add the tuning for Haswell AVX2:
open Makefile add set

OTFLAGS   = -march=core-avx2 -mtune=core-avx2 -O3
(if you have a haswell, of cause)

Results:
pps: 39691 / 39553  10g/h 6059 / 5635  15g/h 0 / 0


Oh yeah. 20% for free.
Thanks dcct.
sr. member
Activity: 280
Merit: 250
Anyone else been keeping an eye on testnet?

Difficulty is rising pretty fast, which means some serious mining from a very small base is happening.

Regards,

--
bsunau7

I see a difficulty below 17. Thats <100k pps total.

Let me show you how to optimize the miner:

My I7-4770k with 8 threads, stock-miner:
pps: 31821 / 31160  10g/h 4473 / 3429  15g/h 0 / 0

Now go on and build your own libgmp from source. When done, run the tune utility for optimum results. Copy your final libgmp.so.10 to /usr/local/lib/.

Now we're at:
pps: 34039 / 34041  10g/h 6077 / 4479  15g/h 0 / 0

Optimizing sieve size:
open src/main.cpp

line 252:
  1048576);

to
   8 * 1048576);


and line 256
  500000);

to
  15000000);

make sure to also change the shift, otherwise you wont be able to submit most blocks! You could specify it on command line, but this is safer:

line 128:
  uint16_t shift = (opts->has_shift() ?  atoi(opts->get_shift().c_str()) : 20);

change to:
  uint16_t shift = (opts->has_shift() ?  atoi(opts->get_shift().c_str()) : 23);


Now its mining at
pps: 38276 / 37068  10g/h 5156 / 5265  15g/h 0 / 0


Lets also add the tuning for Haswell AVX2:
open Makefile and set

OTFLAGS   = -march=core-avx2 -mtune=core-avx2 -O3
(if you have a haswell, of cause)

Results:
pps: 39691 / 39553  10g/h 6059 / 5635  15g/h 0 / 0





hero member
Activity: 673
Merit: 500
Anyone else been keeping an eye on testnet?

Difficulty is rising pretty fast, which means some serious mining from a very small base is happening.

Regards,

--
bsunau7

Could very well be the dev working on a GPU miner.  He mentioned it a few posts up.
member
Activity: 114
Merit: 10
Anyone else been keeping an eye on testnet?

Difficulty is rising pretty fast, which means some serious mining from a very small base is happening.

Regards,

--
bsunau7
hero member
Activity: 938
Merit: 1000
Why is there only one pool to mine Gap ?

doesn't that affect the mining process or the blockchain integrity ?



nonce-pool doesn't have more than 50% nethash. Don't worry.
member
Activity: 69
Merit: 10
I contacted Suchpool.pw with a request to add this coin. Would be nice to see a couple more pools pop up.
legendary
Activity: 2912
Merit: 1309
Just the short question..
is there already a p2pool ready for GAP? so that I can setup one?

full member
Activity: 352
Merit: 100
Does anyone have a list of pools they like to use, or know of pools that are reliable and you think they would like to add/are capable of adding this coin?

I'll email them, asking to add this coin, all you have to do is post the name of the pool/operator so I can google/search for it.

Atrides http://dwarfpool.com/ or http://1gh.com/

maybe if someone asks, they can add Gapcoin.

I asked http://1gh.com add gapcoin through twitter, and http://suchpool.pw but no answer.
hero member
Activity: 556
Merit: 501
Does anyone have a list of pools they like to use, or know of pools that are reliable and you think they would like to add/are capable of adding this coin?

I'll email them, asking to add this coin, all you have to do is post the name of the pool/operator so I can google/search for it.

Atrides http://dwarfpool.com/ or http://1gh.com/

maybe if someone asks, they can add Gapcoin.
hero member
Activity: 605
Merit: 500
Still no one wants to sell me 10k GAP for 20k sat instead of dumping on the exchange ? That's weird.

A scorpion asks a frog to carry him over a river. The frog is afraid of being stung during the trip, but the scorpion argues that if it stung the frog, both would sink and the scorpion would drown. The frog agrees and begins carrying the scorpion, but midway across the river the scorpion does indeed sting the frog, dooming them both. When asked why, the scorpion points out that this is its nature.

tl;dr: Dumpers gonna dump  Grin
full member
Activity: 177
Merit: 154
Does anyone have a list of pools they like to use, or know of pools that are reliable and you think they would like to add/are capable of adding this coin?

I'll email them, asking to add this coin, all you have to do is post the name of the pool/operator so I can google/search for it.

I already asked ypool, but they currently don't have the ability to add a new coin.
member
Activity: 70
Merit: 10
Activity: 350
Does anyone have a list of pools they like to use, or know of pools that are reliable and you think they would like to add/are capable of adding this coin?

I'll email them, asking to add this coin, all you have to do is post the name of the pool/operator so I can google/search for it.
full member
Activity: 177
Merit: 154
Why is there only one pool to mine Gap ?

doesn't that affect the mining process or the blockchain integrity ?



Yes, it's not a perfect situation,
but consider Gapcoin is only two weeks old.

More pools will follow.
member
Activity: 69
Merit: 10
Why is there only one pool to mine Gap ?

doesn't that affect the mining process or the blockchain integrity ?

Pages:
Jump to: