[ANN][RIC] Riecoin: constellations POW *CPU* HARD FORK successful, world record - page 227.

otila

sr. member

Activity: 336

Merit: 250

xptminer fixes:

EnterCriticalSection copy&paste bug

Code:

diff --git a/xptMiner/global.h b/xptMiner/global.h
index 90858c8..61b2702 100644
--- a/xptMiner/global.h
+++ b/xptMiner/global.h
@@ -25,7 +25,7 @@
#define strcpy_s(dest,val,src) strncopy(dest,src,val)
#define __debugbreak(); raise(SIGTRAP);
#define CRITICAL_SECTION pthread_mutex_t
-#define EnterCriticalSection(Section) pthread_mutex_unlock(Section)
+#define EnterCriticalSection(Section) pthread_mutex_lock(Section)
#define LeaveCriticalSection(Section) pthread_mutex_unlock(Section)
#define InitializeCriticalSection(Section) pthread_mutex_init(Section, NULL)
#define DestroyCriticalSection(Section) pthread_mutex_destroy(Section)

(I had also added DestroyCriticalSection)
also, it would be a good idea to check the return values of pthread_*.
xptClient_free could have DestroyCriticalSection for cs_workAccess and cs_shareSubmit.
why cs_workAccess is not being used, no EnterCriticalSection / LeaveCriticalSection?

another mem leak, saves a couple of GiB a day:

Code:

diff --git a/xptMiner/main.cpp b/xptMiner/main.cpp
index 258377f..a17586c 100644
--- a/xptMiner/main.cpp
+++ b/xptMiner/main.cpp
@@ -663,7 +663,9 @@ void xptMiner_xptQueryWorkLoop()
{
// initiate new connection
EnterCriticalSection(&cs_xptClient);
+ xptClient_free(xptClient);
xptClient = xptMiner_initateNewXptConnectionObject();
+
if(minerSettings.requestTarget.donationPercent > 0.1f)
{
//xptClient_addDeveloperFeeEntry(xptClient, "MK6n2VZZBpQrqpP9rtzsC9PRi5t1qsWuGc", getFeeFromDouble(minerSettings.requestTarget.donationPercent / 2.0));

also SIGPIPE should be ignored, I have it done in my version with sigaction
http://stackoverflow.com/questions/108183/how-to-prevent-sigpipes-or-handle-them-properly

makefile could be fixed to handle dependencies, gcc -MM *.cpp *.c

Supercomputing

sr. member

Activity: 278

Merit: 250

Quote from: dga on February 21, 2014, 10:50:16 PM

If you're going that route also - have you modified the rminerd in a clean way to handle 256 bit nonce results coming from the scan function?

This would be a great patch to get gatra to accept for the reference implementation. My own hack is just a silly quick fix, or I'd submit it. It's a core function needed for a lot of the ways to build a very fast scanner. I'd be delighted to share some of the work of improving the infrastructure for all of us poking at faster miners.

-Dave

It should resemble the following if you are using the GMP library, sorry for the typos if any.

// calculate offset
mpz_sub(z_rop, z_N, z_target);
mpz_add_ui(z_rop, z_rop, n);

uint32 size = mpz_sizeinbase(z_rop,2);
printf("\nDATA NONCE Size = %u\n", size);

// submit work
for( uint32 j=0; j_mp_size; j++)
{
memcpy(&pdata[20+2*j], &z_rop->_mp_d[j], Cool

;
pdata[20+2*j] = swab32(pdata[20+2*j]);
pdata[20+2*j+1] = swab32(pdata[20+2*j+1]);
}

return 1;

bsunau7

member

Activity: 114

Merit: 10

Quote from: dga on February 21, 2014, 10:50:16 PM

Quote from: Supercomputing on February 21, 2014, 01:38:47 PM

Quote from: jh00 on February 21, 2014, 11:59:23 AM

I don't really believe that Gatra has had his own optimized version of the miner. Even if he did, he wouldn't be able to mine a majority of starting blocks unless he owns a super large botnet.
This was the first fair coin launch that I saw and I think this is why people feel like it was 'hard to mine'. Only 200k Riecoin were created within the first three days. In comparison, Primecoin generated 600.000 XPM within the same timespan. Everyone complaining should just move on and insta-mine the next coin.

Regarding the low value, Poloniex and all the other small exchanges don't really have any buy support at the moment. The value can only go down there. Just take a look at Poloniex stats:
Sell Orders Total: 24345.77892931 RIC
Buy Orders Total: 2.35857895 BTC
There are only 300.000 RIC in existence. This means Riecoin only has a market cap of around 200 000 USD. Even the worst of the clone coins have more than that, it is only a matter of time until value shoots up. If you believe in Riecoin it would be a good time to actually buy more RIC and wait for the big exchanges to add it.

I completely agree that this coin was fairly launched. People in my position had the time to study the proof-of-work and start work on their own miner implementations in advance. As for me, I did not have the time until last weekend. I spent about 4 hours implementing my own optimized miner and it is only slightly faster than jh00's implementation.

I will continue to mine both Riecoin and Primecoin. And at this time, I am much more interested in optimizing the miner than I am interested in the coin's current market value - this coin is only a few days old.

So far, my implementation uses 16,777,216 sieving primes with a 240-bit primorial. This weekend, if time permits, I will search for better pseudo-primorials.

If you're going that route also - have you modified the rminerd in a clean way to handle 256 bit nonce results coming from the scan function?

This would be a great patch to get gatra to accept for the reference implementation. My own hack is just a silly quick fix, or I'd submit it. It's a core function needed for a lot of the ways to build a very fast scanner. I'd be delighted to share some of the work of improving the infrastructure for all of us poking at faster miners.

-Dave

Agreed, a (slightly less than) 32bit nonce is feeling pretty small at high rates. With a 240-bit primorial I would be surprised if you got many (if any) 6-chain candidates within the exiting nonce. But I've not spent any time looking at the issues in using a bigger nonce (still focusing on correctness of code).

Regards,

--
bsunau7

dga

hero member

Activity: 737

Merit: 511

Quote from: gatra on February 21, 2014, 11:21:00 PM

I did it for 64 bit nonces, which seems enough. Will publish during the weekend, but I think my miner is not faster than xptMiner...

Depends on the approach you're taking. If you want to use a stupidly big primorial, as both supercomputer and I seem to be doing, then you need 256.

There's actually an interesting gap between the spec and the protocol implementation, btw: The spec says that you can find a nonce that fits within the number of 0's added to the end to pad to the required difficulty, but the protocol implementation seems to only permit a 256 bit nonce (unless that's just the miner?).

Which one is correct, out of curiosity?

Thanks for a lot of fun, btw. I really like this one. Though I'd also love to quibble about the time-to-find-block estimator. *grin*

-Dave

cryptonewbie

full member

Activity: 434

Merit: 100

Quote from: glongsword on February 21, 2014, 11:16:56 PM

Quote from: cryptonewbie on February 21, 2014, 11:11:47 PM

I may be a little late the to the party, but would someone mind giving me a brief update of where things stand? Is the development team active, mining etc.?

Thanks!

The developer just released a new wallet and is working on P2Pool and eventually a new GPU miner, and the difficulty has been about the same for a few days despite the price on exchanges falling substantially.

Thanks! So, everything seems to be looking up? So many coins to sift through that I overlooked this one. Looks intriguing.

gatra

hero member

Activity: 583

Merit: 505

CTO @ Flixxo, Riecoin dev

Quote from: dga on February 21, 2014, 10:50:16 PM

Quote from: Supercomputing on February 21, 2014, 01:38:47 PM

Quote from: jh00 on February 21, 2014, 11:59:23 AM

I don't really believe that Gatra has had his own optimized version of the miner. Even if he did, he wouldn't be able to mine a majority of starting blocks unless he owns a super large botnet.
This was the first fair coin launch that I saw and I think this is why people feel like it was 'hard to mine'. Only 200k Riecoin were created within the first three days. In comparison, Primecoin generated 600.000 XPM within the same timespan. Everyone complaining should just move on and insta-mine the next coin.

Regarding the low value, Poloniex and all the other small exchanges don't really have any buy support at the moment. The value can only go down there. Just take a look at Poloniex stats:
Sell Orders Total: 24345.77892931 RIC
Buy Orders Total: 2.35857895 BTC
There are only 300.000 RIC in existence. This means Riecoin only has a market cap of around 200 000 USD. Even the worst of the clone coins have more than that, it is only a matter of time until value shoots up. If you believe in Riecoin it would be a good time to actually buy more RIC and wait for the big exchanges to add it.

I completely agree that this coin was fairly launched. People in my position had the time to study the proof-of-work and start work on their own miner implementations in advance. As for me, I did not have the time until last weekend. I spent about 4 hours implementing my own optimized miner and it is only slightly faster than jh00's implementation.

I will continue to mine both Riecoin and Primecoin. And at this time, I am much more interested in optimizing the miner than I am interested in the coin's current market value - this coin is only a few days old.

So far, my implementation uses 16,777,216 sieving primes with a 240-bit primorial. This weekend, if time permits, I will search for better pseudo-primorials.

If you're going that route also - have you modified the rminerd in a clean way to handle 256 bit nonce results coming from the scan function?

This would be a great patch to get gatra to accept for the reference implementation. My own hack is just a silly quick fix, or I'd submit it. It's a core function needed for a lot of the ways to build a very fast scanner. I'd be delighted to share some of the work of improving the infrastructure for all of us poking at faster miners.

-Dave

I did it for 64 bit nonces, which seems enough. Will publish during the weekend, but I think my miner is not faster than xptMiner...

glongsword

full member

Activity: 314

Merit: 100

Quote from: cryptonewbie on February 21, 2014, 11:11:47 PM

I may be a little late the to the party, but would someone mind giving me a brief update of where things stand? Is the development team active, mining etc.?

Thanks!

The developer just released a new wallet and is working on P2Pool and eventually a new GPU miner, and the difficulty has been about the same for a few days despite the price on exchanges falling substantially.

cryptonewbie

full member

Activity: 434

Merit: 100

I may be a little late the to the party, but would someone mind giving me a brief update of where things stand? Is the development team active, mining etc.?

Thanks!

northranger79510

sr. member

Activity: 308

Merit: 250

Riecoin and Huntercoin to rule all!

What is up with the crazy sell wall...its like sell/sell/sell/sell/sell/sell/sell/sell/sell/sell/sell//sell/sell/buy/sell/sell/sell/sell/sell/sell/sell/sell/sell/sell

dga

hero member

Activity: 737

Merit: 511

Quote from: Supercomputing on February 21, 2014, 01:38:47 PM

Quote from: jh00 on February 21, 2014, 11:59:23 AM

I don't really believe that Gatra has had his own optimized version of the miner. Even if he did, he wouldn't be able to mine a majority of starting blocks unless he owns a super large botnet.
This was the first fair coin launch that I saw and I think this is why people feel like it was 'hard to mine'. Only 200k Riecoin were created within the first three days. In comparison, Primecoin generated 600.000 XPM within the same timespan. Everyone complaining should just move on and insta-mine the next coin.

Regarding the low value, Poloniex and all the other small exchanges don't really have any buy support at the moment. The value can only go down there. Just take a look at Poloniex stats:
Sell Orders Total: 24345.77892931 RIC
Buy Orders Total: 2.35857895 BTC
There are only 300.000 RIC in existence. This means Riecoin only has a market cap of around 200 000 USD. Even the worst of the clone coins have more than that, it is only a matter of time until value shoots up. If you believe in Riecoin it would be a good time to actually buy more RIC and wait for the big exchanges to add it.

I completely agree that this coin was fairly launched. People in my position had the time to study the proof-of-work and start work on their own miner implementations in advance. As for me, I did not have the time until last weekend. I spent about 4 hours implementing my own optimized miner and it is only slightly faster than jh00's implementation.

I will continue to mine both Riecoin and Primecoin. And at this time, I am much more interested in optimizing the miner than I am interested in the coin's current market value - this coin is only a few days old.

So far, my implementation uses 16,777,216 sieving primes with a 240-bit primorial. This weekend, if time permits, I will search for better pseudo-primorials.

If you're going that route also - have you modified the rminerd in a clean way to handle 256 bit nonce results coming from the scan function?

This would be a great patch to get gatra to accept for the reference implementation. My own hack is just a silly quick fix, or I'd submit it. It's a core function needed for a lot of the ways to build a very fast scanner. I'd be delighted to share some of the work of improving the infrastructure for all of us poking at faster miners.

-Dave

northranger79510

sr. member

Activity: 308

Merit: 250

Riecoin and Huntercoin to rule all!

Hey guys,

I made a Facebook Page but forgot about it for a while until now. I'd appreciate it if you guys like the page as I will try to actively post something daily about Riecoin.

https://www.facebook.com/pages/Riecoin/283596855127718

physalis

sr. member

Activity: 392

Merit: 250

http://ric.candypool.net/ seems down

http://ypool.net/ i get a malware alert - wtf?

Is there any good, trusted pool for this coin?
Any special mining software I should use, or just the one that comes with the client? There's no GPU miner (yet), correct?

Supercomputing

sr. member

Activity: 278

Merit: 250

@gatra

riecoin-qt.exe -testnet -checkpoints=0

I am still receiving a lot of rejected blocks on testnet.

But real network is OK.

2014-02-21 22:45:58 proof-of-work found
hash: bdaec03fca54a9a4d4ce60359449cef240359eb28f75f22e531d75b958545852
diff compacted: 0204d200 nOffset: 000000000000000000000000000000000000000000000000000000d91c0c3111
2014-02-21 22:45:58 CBlock(hash=e22bf43abe95b90d1609f5998e3454bc50af66d01a15a8c1cd1b1f3ea479d4ac, ver=2, hashPrevBlock=a1073f4956beaae2d4a8539229938e416e163706ff06ee456d2aa88faa754fbc, hashMerkleRoot=55aff6537c123147b683c5d10e54d071cb12b202d58bf28f6eea2bfe4d48c3c6, nTime=1393022696, nBitsCompact=0x0204d200, nOffset=000000000000000000000000000000000000000000000000000000d91c0c3111, vtx=13)
.
.
.
vMerkleTree: df3081d3ac f651321817 b58731aceb d6e8d5fb7b 3aaa48e28b c1e74c76a5 4746335373 1a8182b95d b5eebd8794 c53e6b4171 13a58ad1d9 735cfdf8b4 01b856315e 8ae5a147b4 3da266c409 28da4d73d5 af3493b86c 4ad31215c0 6fb9a72b04 f4b64ccf50 bb3d2e57aa c697499e1d ceb72cbf0f 96d475664b b9a501a561 a5ec2ce527 55aff6537c
2014-02-21 22:45:58 generated 50.0012

northranger79510

sr. member

Activity: 308

Merit: 250

Riecoin and Huntercoin to rule all!

Just bought some more Riecoins =D Hopefully I did not make a bad decision.

gatra

hero member

Activity: 583

Merit: 505

CTO @ Flixxo, Riecoin dev

Quote from: Supercomputing on February 21, 2014, 05:18:21 PM

@gatra

I am now receiving the following error message with your latest Windows client while on testnet:
block with too much proof-of-work

However, the original Windows client works fine, any ideas?

maybe because I didn't put any valid checkpoint for testnet...
please try running riecoind with "-checkpoints=0" besides "-testnet"

Supercomputing

sr. member

Activity: 278

Merit: 250

@gatra

Debug Info: DATA_NONCE Size = 31-bit or 4 bytes
[2014-02-21 15:20:52] accepted: 0/61 (0.00%), 0.00 khash/s (booooo)

Supercomputing

sr. member

Activity: 278

Merit: 250

@gatra

I am now receiving the following error message with your latest Windows client while on testnet:
block with too much proof-of-work

However, the original Windows client works fine, any ideas?

Kenta

sr. member

Activity: 420

Merit: 250

Quote from: gatra on February 21, 2014, 03:39:53 PM

New client version released, with more uniqueness: "the first coin with an official release addressing the transaction malleability problem".

Much better! Thanks!! Cool

gatra

hero member

Activity: 583

Merit: 505

CTO @ Flixxo, Riecoin dev

Quote from: northranger79510 on February 21, 2014, 03:44:20 PM

Thanks for the update! Grin

Boost of confidence right there. But any future plans? Maybe incorporate coin mixer or something? Just curious. Sorry if it sounds rude.

Of course I'd like to add as many features as possible, but it's been less than 2 weeks and there is no p2pool yet. Let's give it time to mature.

Maybe you (or anyone) could start a poll or a discussion on wanted features for the future, so I could use it as a guide. Current plans are p2pool first and block explorer and gpuminer next.

northranger79510

sr. member

Activity: 308

Merit: 250

Riecoin and Huntercoin to rule all!

Thanks for the update! Grin

Boost of confidence right there. But any future plans? Maybe incorporate coin mixer or something? Just curious. Sorry if it sounds rude.

Topic: [ANN][RIC] Riecoin: constellations POW *CPU* HARD FORK successful, world record - page 227. (Read 685214 times)

Topic: [ANN][RIC] Riecoin: constellations POW CPU HARD FORK successful, world record - page 227. (Read 685214 times)