I was running 0.15.0.1 in -blocksonly mode and then issued a tx. Obviously it can go nowhere. So I did an abandontransaction and got an error (-5).

Restarted the client in both blocksonly and normal mode later, erased mempool.dat, etc etc, abandontransaction still won't work. I guess I'll have to rescan, but I shouldn't, right?

[1-2m]

Launching 0.12 bitcoin-qt in linux with -dbcache=2048

SSD = around 1-2m to see wallet
mechanical = around 4-5m to see wallet

...similar very slow times with -dbcache=1024.

Launching 0.12 bitcoin-qt with  -dbcache=0

SSD and mechanical = 10-15 secs at most.

-dbcache=0 works way faster, betraying some kind of problem in the way caching works.

Is it possible to have a --bench option (perhaps even a GUI frontend in the qt) for the (non-mining) CPU-intensive tasks in order to evaluate custom builds, software improvements and regressions as the software evolves? Performance bugs and cross-platform performance issues will also be spotted faster.

It will also allow hardware sites to include bitcoin related benchmarks in their sites as well as allowing people to choose the right hardware for the task of running a full node.

I just realized something about RBF - please correct any misunderstandings I may have, I'm not a btc developer.

We are all operating under the assumption that a quantum computer doesn't exist - but I'm not so sure that there isn't, or that there won't be pretty soon. In any case, the greatest safeguard we have against such a possibility is storing coins in addresses that have no prior spending in them. So when a QC is on the loose, the best one can do is to spend the full amount in order to not leave any coins behind for priv. key extrapolation and hacking. But this is based on a first seen-first serve scenario. RBF would allow an attacker to see a transaction, extrapolate the priv key and issue a respending with a higher fee, hijacking one's money.

So, from what I understand, RBF reduces the futureproofing / quantum resistance of BTC. Is there a way where it can only be implemented by honoring the initial transactions, plus a higher fee - for those desiring to jump the queue - and reject any other attempts to change the destination?

In August, a post appeared in the dev mailing list, signed as Satoshi Nakamoto, with an email of Satoshi Nakamoto (not spoofed), from an email that has not been known to have been hacked, which contained the following message and warning regarding the attempted XT coup:

I have been following the recent block size debates through the mailing list.  I had hoped the debate would resolve and that a fork proposal would achieve widespread consensus.  However with the formal release of Bitcoin XT 0.11A, this looks unlikely to happen, and so I am forced to share my concerns about this very dangerous fork.

The developers of this pretender-Bitcoin claim to be following my original vision, but nothing could be further from the truth.  When I designed Bitcoin, I designed it in such a way as to make future modifications to the consensus rules difficult without near unanimous agreement.  Bitcoin was designed to be protected from the influence of charismatic leaders, even if their name is Gavin Andresen, Barack Obama, or Satoshi Nakamoto.  Nearly everyone has to agree on a change, and they have to do it without being forced or pressured into it.  By doing a fork in this way, these developers are violating the "original vision" they claim to honour.

They use my old writings to make claims about what Bitcoin was supposed to be.  However I acknowledge that a lot has changed since that time, and new knowledge has been gained that contradicts some of my early opinions.  For example I didn't anticipate pooled mining and its effects on the security of the network.  Making Bitcoin a competitive monetary system while also preserving its security properties is not a trivial problem, and we should take more time to come up with a robust solution.  I suspect we need a better incentive for users to run nodes instead of relying solely on altruism.

If two developers can fork Bitcoin and succeed in redefining what "Bitcoin" is, in the face of widespread technical criticism and through the use of populist tactics, then I will have no choice but to declare Bitcoin a failed project.  Bitcoin was meant to be both technically and socially robust.  This present situation has been very disappointing to watch unfold.

Satoshi Nakamoto

https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2015-August/010238.html

Obviously, the context is still massively relevant to the new "classic" fork attempt as the ingredients are the same.

If the identity of the poster is legitimate, then there is a real risk that a fork might force Satoshi to declare Bitcoin a failed project due to the coup in governance. That will official #REKT bitcoin and crypto in general, no matter which part of the argument anyone supports.

Some say Satoshi needed to sign the mail but there is not a single instance where Satoshi signed anything. Plus if he did he would also be doing the same thing he condemns regarding authority influence. So his message being semi-official / semi-unofficial, can act as a warning, but also as a non-decisive influence.

I want to see the community's weighing of that particular risk, if a fork occurs: Do you expect Satoshi to come out and make an official obituary of BTC if BTC forks - which would make Hearn's (or the other 80something obituaries) pale in comparison?

We know that "change" in real life are useful because you give ten dollars and you get back change. However in Bitcoin you can send a precise amount of coins, so change is not really "necessary" - not even as an option. It's not needed and adds bloat out of nowhere.

Some say "change increase privacy so that's why it was placed in there". Surely, a protocol as transparent as bitcoin, wouldn't increase its privacy by any significant amount through change (not to mention that change can be linked during future spending). This is stuff that even a script can put together, deanonymizing transactions.

And why, if it is privacy-related, wouldn't one be able to control change spending so as to not be linked together?

So there has to be something else here that Satoshi saw.

I'm thinking it may be related to Quantum-Computing resistance. By moving the amount to the recipient + change to a new address (which hasn't yet published its public key), a good portion of the network's money will remain uncrackable by a quantum computer as the QC won't know the public key to extrapolate the private key.

If control of change is going to be implemented in future versions of Bitcoin, this quantum-resistance could be broken. Perhaps it should also be accompanied by a change in the private/public key algorithm to a quantum-resistant one.

Or, alternatively, introduce a button in the wallet that places one's funds into "quantum storage" - aggregating them automatically in a single address with no spends (that prevent QC cracking). One could even checkbox something like "automatic quantum storage" so that when one wants to spend money, one amount would go to the destination and the other would go to a new address with zero spends. Thus change control won't affect the principle of QC-resistance, if that's the rationale of Satoshi.

Suppose we are a few years ahead and that we have already discovered AI that exceeds human intelligence. How would that effect today's cryptocurrencies? Would they be able to compete with AI?

What I have in mind is this:

Right now dumb if-then-else algorithms dictate the operation of PoW and PoS crypto.

What if we eliminate the need for PoW / PoS and all 51% risks, by replacing these systems with a trusted party which is actually ...not human, but rather an intelligence which acts as the "operator" of the whole network and runs as a decentralized intelligence.

This AI operator will have no politics or agenda, so in that regard "he"/"she" would be as trusted as an if-then-else dumb algo, if not more (because there is no 51% threat anymore). So trust is not something that would bother anyone really.

This AI operator would conduct and verify all transactions, ensure network resilience, load balance the network in an intelligent way, protect it from DDOS on its own by detecting attack patterns, distribute the blockchain over the cloud so that no 15 TB wallets are needed and much more stuff that could be added due to its super-intelligence. For example personal authentication by presenting yourself to the "operator" as an additional safeguard to one's password where the operator can understand you based on your face, voice, etc.

The question is this: How futureproof would Bitcoin be in such a scenario of an intelligent currency network that does everything one would dream of with very few of today's limitations? Or how would it be, from the developers perspective, that in, say, 15-20 years later from Bitcoin's inception, the BTC would look like a dinosaur that needs "updating"... What then? Would the devs feel obligated to keep Satoshi's dream alive by remaining true to that vision he had, or would any discussion of updating the core to an intelligent network be faced with a church-like attitude of "blasphemy"?

This thought experiment has implications because it also puts into question what we know about BTC right now... stuff like "its characteristics will be these and these by year X" could be totally invalid. (I'm not necessarily talking about coin numbers, but mining could be affected if the PoW is redundant as a dinosaur-type-of-way to verify the blockchain).

There may be quite a dilemma into how to stand true to what was once started as X or evolve into the much better Z alternative. I know the date of this happening seems too far into the future but you'll recall this post when the first intelligent cryptocurrency is issued to compete with BTC. Questions will start to be raised about how to deal with the situation and whether to adapt or "remain true".

Or maybe the tables are turned and the criticism is on the AI side: Who will update the AI? Isn't the AI's update a form of centralization and control of the network? Or if the AI is "fixed" from its inception and non-upgradeable, what would stop someone making a better currency in another 10 years when AI is even more advanced and has more capabilities, or even self-improvement-capabilities that will allow it to evolve on its own?

We'll see how it goes...

I've tried some small OpenCL improvements that got my card from 255 khash to 262khash. While it's a minor improvement I think it could be quite significant for GPU farms (2.6% = 26 cards working for free in a 1000 GPU farm).

Now, to the chase, this is the modified scrypt130511.cl (which you must use to replace the scrypt130511.cl or other similar scryptxxxxx.cl of the cgminer 3.7.2 source directory)

Code:

/*-
 * Copyright 2009 Colin Percival, 2011 ArtForz, 2011 pooler, 2012 mtrlt
 * 2012-2013 Con Kolivas.
 * All rights reserved.
 *
 *"Unrolled edition" v1.00 for 2.6% gain (AMD 5xxx) by AlexGR (2014)
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 * This file was originally written by Colin Percival as part of the Tarsnap
 * online backup system.
 */

__constant uint ES[2] = { 0x00FF00FF, 0xFF00FF00 };
__constant uint K[] = {
	0x428a2f98U,
	0x71374491U,
	0xb5c0fbcfU,
	0xe9b5dba5U,
	0x3956c25bU,
	0x59f111f1U,
	0x923f82a4U,
	0xab1c5ed5U,
	0xd807aa98U,
	0x12835b01U,
	0x243185beU, // 10
	0x550c7dc3U,
	0x72be5d74U,
	0x80deb1feU,
	0x9bdc06a7U,
	0xe49b69c1U,
	0xefbe4786U,
	0x0fc19dc6U,
	0x240ca1ccU,
	0x2de92c6fU,
	0x4a7484aaU, // 20
	0x5cb0a9dcU,
	0x76f988daU,
	0x983e5152U,
	0xa831c66dU,
	0xb00327c8U,
	0xbf597fc7U,
	0xc6e00bf3U,
	0xd5a79147U,
	0x06ca6351U,
	0x14292967U, // 30
	0x27b70a85U,
	0x2e1b2138U,
	0x4d2c6dfcU,
	0x53380d13U,
	0x650a7354U,
	0x766a0abbU,
	0x81c2c92eU,
	0x92722c85U,
	0xa2bfe8a1U,
	0xa81a664bU, // 40
	0xc24b8b70U,
	0xc76c51a3U,
	0xd192e819U,
	0xd6990624U,
	0xf40e3585U,
	0x106aa070U,
	0x19a4c116U,
	0x1e376c08U,
	0x2748774cU,
	0x34b0bcb5U, // 50
	0x391c0cb3U,
	0x4ed8aa4aU,
	0x5b9cca4fU,
	0x682e6ff3U,
	0x748f82eeU,
	0x78a5636fU,
	0x84c87814U,
	0x8cc70208U,
	0x90befffaU,
	0xa4506cebU, // 60
	0xbef9a3f7U,
	0xc67178f2U,
	0x98c7e2a2U,
	0xfc08884dU,
	0xcd2a11aeU,
	0x510e527fU,
	0x9b05688cU,
	0xC3910C8EU,
	0xfb6feee7U,
	0x2a01a605U, // 70
	0x0c2e12e0U,
	0x4498517BU,
	0x6a09e667U,
	0xa4ce148bU,
	0x95F61999U,
	0xc19bf174U,
	0xBB67AE85U,
	0x3C6EF372U,
	0xA54FF53AU,
	0x1F83D9ABU, // 80
	0x5BE0CD19U,
	0x5C5C5C5CU,
	0x36363636U,
	0x80000000U,
	0x000003FFU,
	0x00000280U,
	0x000004a0U,
	0x00000300U
};

#define rotl(x,y) rotate(x,y)
#define Ch(x,y,z) bitselect(z,y,x)
#define Maj(x,y,z) Ch((x^z),y,z)

#define EndianSwap(n) (rotl(n & ES[0], 24U)|rotl(n & ES[1], 8U))

#define Tr2(x)		(rotl(x, 30U) ^ rotl(x, 19U) ^ rotl(x, 10U))
#define Tr1(x)		(rotl(x, 26U) ^ rotl(x, 21U) ^ rotl(x, 7U))
#define Wr2(x)		(rotl(x, 25U) ^ rotl(x, 14U) ^ (x>>3U))
#define Wr1(x)		(rotl(x, 15U) ^ rotl(x, 13U) ^ (x>>10U))

#define RND(a, b, c, d, e, f, g, h, k)	\
	h += Tr1(e); 			\
	h += Ch(e, f, g); 		\
	h += k;				\
	d += h;				\
	h += Tr2(a); 			\
	h += Maj(a, b, c);

void SHA256(uint4*restrict state0,uint4*restrict state1, const uint4 block0, const uint4 block1, const uint4 block2, const uint4 block3)
{
	uint4 S0 = *state0;
	uint4 S1 = *state1;
	
#define A S0.x
#define B S0.y
#define C S0.z
#define D S0.w
#define E S1.x
#define F S1.y
#define G S1.z
#define H S1.w

	uint4 W[4];

	W[ 0].x = block0.x;
	RND(A,B,C,D,E,F,G,H, W[0].x+ K[0]);
	W[ 0].y = block0.y;
	RND(H,A,B,C,D,E,F,G, W[0].y+ K[1]);
	W[ 0].z = block0.z;
	RND(G,H,A,B,C,D,E,F, W[0].z+ K[2]);
	W[ 0].w = block0.w;
	RND(F,G,H,A,B,C,D,E, W[0].w+ K[3]);

	W[ 1].x = block1.x;
	RND(E,F,G,H,A,B,C,D, W[1].x+ K[4]);
	W[ 1].y = block1.y;
	RND(D,E,F,G,H,A,B,C, W[1].y+ K[5]);
	W[ 1].z = block1.z;
	RND(C,D,E,F,G,H,A,B, W[1].z+ K[6]);
	W[ 1].w = block1.w;
	RND(B,C,D,E,F,G,H,A, W[1].w+ K[7]);

	W[ 2].x = block2.x;
	RND(A,B,C,D,E,F,G,H, W[2].x+ K[8]);
	W[ 2].y = block2.y;
	RND(H,A,B,C,D,E,F,G, W[2].y+ K[9]);
	W[ 2].z = block2.z;
	RND(G,H,A,B,C,D,E,F, W[2].z+ K[10]);
	W[ 2].w = block2.w;
	RND(F,G,H,A,B,C,D,E, W[2].w+ K[11]);

	W[ 3].x = block3.x;
	RND(E,F,G,H,A,B,C,D, W[3].x+ K[12]);
	W[ 3].y = block3.y;
	RND(D,E,F,G,H,A,B,C, W[3].y+ K[13]);
	W[ 3].z = block3.z;
	RND(C,D,E,F,G,H,A,B, W[3].z+ K[14]);
	W[ 3].w = block3.w;
	RND(B,C,D,E,F,G,H,A, W[3].w+ K[76]);

	W[ 0].x += Wr1(W[ 3].z) + W[ 2].y + Wr2(W[ 0].y);
	RND(A,B,C,D,E,F,G,H, W[0].x+ K[15]);

	W[ 0].y += Wr1(W[ 3].w) + W[ 2].z + Wr2(W[ 0].z);
	RND(H,A,B,C,D,E,F,G, W[0].y+ K[16]);

	W[ 0].z += Wr1(W[ 0].x) + W[ 2].w + Wr2(W[ 0].w);
	RND(G,H,A,B,C,D,E,F, W[0].z+ K[17]);

	W[ 0].w += Wr1(W[ 0].y) + W[ 3].x + Wr2(W[ 1].x);
	RND(F,G,H,A,B,C,D,E, W[0].w+ K[18]);

	W[ 1].x += Wr1(W[ 0].z) + W[ 3].y + Wr2(W[ 1].y);
	RND(E,F,G,H,A,B,C,D, W[1].x+ K[19]);

	W[ 1].y += Wr1(W[ 0].w) + W[ 3].z + Wr2(W[ 1].z);
	RND(D,E,F,G,H,A,B,C, W[1].y+ K[20]);

	W[ 1].z += Wr1(W[ 1].x) + W[ 3].w + Wr2(W[ 1].w);
	RND(C,D,E,F,G,H,A,B, W[1].z+ K[21]);

	W[ 1].w += Wr1(W[ 1].y) + W[ 0].x + Wr2(W[ 2].x);
	RND(B,C,D,E,F,G,H,A, W[1].w+ K[22]);

	W[ 2].x += Wr1(W[ 1].z) + W[ 0].y + Wr2(W[ 2].y);
	RND(A,B,C,D,E,F,G,H, W[2].x+ K[23]);

	W[ 2].y += Wr1(W[ 1].w) + W[ 0].z + Wr2(W[ 2].z);
	RND(H,A,B,C,D,E,F,G, W[2].y+ K[24]);

	W[ 2].z += Wr1(W[ 2].x) + W[ 0].w + Wr2(W[ 2].w);
	RND(G,H,A,B,C,D,E,F, W[2].z+ K[25]);

	W[ 2].w += Wr1(W[ 2].y) + W[ 1].x + Wr2(W[ 3].x);
	RND(F,G,H,A,B,C,D,E, W[2].w+ K[26]);

	W[ 3].x += Wr1(W[ 2].z) + W[ 1].y + Wr2(W[ 3].y);
	RND(E,F,G,H,A,B,C,D, W[3].x+ K[27]);

	W[ 3].y += Wr1(W[ 2].w) + W[ 1].z + Wr2(W[ 3].z);
	RND(D,E,F,G,H,A,B,C, W[3].y+ K[28]);

	W[ 3].z += Wr1(W[ 3].x) + W[ 1].w + Wr2(W[ 3].w);
	RND(C,D,E,F,G,H,A,B, W[3].z+ K[29]);

	W[ 3].w += Wr1(W[ 3].y) + W[ 2].x + Wr2(W[ 0].x);
	RND(B,C,D,E,F,G,H,A, W[3].w+ K[30]);

	W[ 0].x += Wr1(W[ 3].z) + W[ 2].y + Wr2(W[ 0].y);
	RND(A,B,C,D,E,F,G,H, W[0].x+ K[31]);

	W[ 0].y += Wr1(W[ 3].w) + W[ 2].z + Wr2(W[ 0].z);
	RND(H,A,B,C,D,E,F,G, W[0].y+ K[32]);

	W[ 0].z += Wr1(W[ 0].x) + W[ 2].w + Wr2(W[ 0].w);
	RND(G,H,A,B,C,D,E,F, W[0].z+ K[33]);

	W[ 0].w += Wr1(W[ 0].y) + W[ 3].x + Wr2(W[ 1].x);
	RND(F,G,H,A,B,C,D,E, W[0].w+ K[34]);

	W[ 1].x += Wr1(W[ 0].z) + W[ 3].y + Wr2(W[ 1].y);
	RND(E,F,G,H,A,B,C,D, W[1].x+ K[35]);

	W[ 1].y += Wr1(W[ 0].w) + W[ 3].z + Wr2(W[ 1].z);
	RND(D,E,F,G,H,A,B,C, W[1].y+ K[36]);

	W[ 1].z += Wr1(W[ 1].x) + W[ 3].w + Wr2(W[ 1].w);
	RND(C,D,E,F,G,H,A,B, W[1].z+ K[37]);

	W[ 1].w += Wr1(W[ 1].y) + W[ 0].x + Wr2(W[ 2].x);
	RND(B,C,D,E,F,G,H,A, W[1].w+ K[38]);

	W[ 2].x += Wr1(W[ 1].z) + W[ 0].y + Wr2(W[ 2].y);
	RND(A,B,C,D,E,F,G,H, W[2].x+ K[39]);

	W[ 2].y += Wr1(W[ 1].w) + W[ 0].z + Wr2(W[ 2].z);
	RND(H,A,B,C,D,E,F,G, W[2].y+ K[40]);

	W[ 2].z += Wr1(W[ 2].x) + W[ 0].w + Wr2(W[ 2].w);
	RND(G,H,A,B,C,D,E,F, W[2].z+ K[41]);

	W[ 2].w += Wr1(W[ 2].y) + W[ 1].x + Wr2(W[ 3].x);
	RND(F,G,H,A,B,C,D,E, W[2].w+ K[42]);

	W[ 3].x += Wr1(W[ 2].z) + W[ 1].y + Wr2(W[ 3].y);
	RND(E,F,G,H,A,B,C,D, W[3].x+ K[43]);

	W[ 3].y += Wr1(W[ 2].w) + W[ 1].z + Wr2(W[ 3].z);
	RND(D,E,F,G,H,A,B,C, W[3].y+ K[44]);

	W[ 3].z += Wr1(W[ 3].x) + W[ 1].w + Wr2(W[ 3].w);
	RND(C,D,E,F,G,H,A,B, W[3].z+ K[45]);

	W[ 3].w += Wr1(W[ 3].y) + W[ 2].x + Wr2(W[ 0].x);
	RND(B,C,D,E,F,G,H,A, W[3].w+ K[46]);

	W[ 0].x += Wr1(W[ 3].z) + W[ 2].y + Wr2(W[ 0].y);
	RND(A,B,C,D,E,F,G,H, W[0].x+ K[47]);

	W[ 0].y += Wr1(W[ 3].w) + W[ 2].z + Wr2(W[ 0].z);
	RND(H,A,B,C,D,E,F,G, W[0].y+ K[48]);

	W[ 0].z += Wr1(W[ 0].x) + W[ 2].w + Wr2(W[ 0].w);
	RND(G,H,A,B,C,D,E,F, W[0].z+ K[49]);

	W[ 0].w += Wr1(W[ 0].y) + W[ 3].x + Wr2(W[ 1].x);
	RND(F,G,H,A,B,C,D,E, W[0].w+ K[50]);

	W[ 1].x += Wr1(W[ 0].z) + W[ 3].y + Wr2(W[ 1].y);
	RND(E,F,G,H,A,B,C,D, W[1].x+ K[51]);

	W[ 1].y += Wr1(W[ 0].w) + W[ 3].z + Wr2(W[ 1].z);
	RND(D,E,F,G,H,A,B,C, W[1].y+ K[52]);

	W[ 1].z += Wr1(W[ 1].x) + W[ 3].w + Wr2(W[ 1].w);
	RND(C,D,E,F,G,H,A,B, W[1].z+ K[53]);

	W[ 1].w += Wr1(W[ 1].y) + W[ 0].x + Wr2(W[ 2].x);
	RND(B,C,D,E,F,G,H,A, W[1].w+ K[54]);

	W[ 2].x += Wr1(W[ 1].z) + W[ 0].y + Wr2(W[ 2].y);
	RND(A,B,C,D,E,F,G,H, W[2].x+ K[55]);

	W[ 2].y += Wr1(W[ 1].w) + W[ 0].z + Wr2(W[ 2].z);
	RND(H,A,B,C,D,E,F,G, W[2].y+ K[56]);

	W[ 2].z += Wr1(W[ 2].x) + W[ 0].w + Wr2(W[ 2].w);
	RND(G,H,A,B,C,D,E,F, W[2].z+ K[57]);

	W[ 2].w += Wr1(W[ 2].y) + W[ 1].x + Wr2(W[ 3].x);
	RND(F,G,H,A,B,C,D,E, W[2].w+ K[58]);

	W[ 3].x += Wr1(W[ 2].z) + W[ 1].y + Wr2(W[ 3].y);
	RND(E,F,G,H,A,B,C,D, W[3].x+ K[59]);

	W[ 3].y += Wr1(W[ 2].w) + W[ 1].z + Wr2(W[ 3].z);
	RND(D,E,F,G,H,A,B,C, W[3].y+ K[60]);

	W[ 3].z += Wr1(W[ 3].x) + W[ 1].w + Wr2(W[ 3].w);
	RND(C,D,E,F,G,H,A,B, W[3].z+ K[61]);

	W[ 3].w += Wr1(W[ 3].y) + W[ 2].x + Wr2(W[ 0].x);
	RND(B,C,D,E,F,G,H,A, W[3].w+ K[62]);
	
#undef A
#undef B
#undef C
#undef D
#undef E
#undef F
#undef G
#undef H

	*state0 += S0;
	*state1 += S1;
}

void SHA256_fresh(uint4*restrict state0,uint4*restrict state1, const uint4 block0, const uint4 block1, const uint4 block2, const uint4 block3)
{
#define A (*state0).x
#define B (*state0).y
#define C (*state0).z
#define D (*state0).w
#define E (*state1).x
#define F (*state1).y
#define G (*state1).z
#define H (*state1).w

	uint4 W[4];

	W[0].x = block0.x;
	D= K[63] +W[0].x;
	H= K[64] +W[0].x;

	W[0].y = block0.y;
	C= K[65] +Tr1(D)+Ch(D, K[66], K[67])+W[0].y;
	G= K[68] +C+Tr2(H)+Ch(H, K[69] ,K[70]);

	W[0].z = block0.z;
	B= K[71] +Tr1(C)+Ch(C,D,K[66])+W[0].z;
	F= K[72] +B+Tr2(G)+Maj(G,H, K[73]);

	W[0].w = block0.w;
	A= K[74] +Tr1(B)+Ch(B,C,D)+W[0].w;
	E= K[75] +A+Tr2(F)+Maj(F,G,H);

	W[1].x = block1.x;
	RND(E,F,G,H,A,B,C,D, W[1].x+ K[4]);
	W[1].y = block1.y;
	RND(D,E,F,G,H,A,B,C, W[1].y+ K[5]);
	W[1].z = block1.z;
	RND(C,D,E,F,G,H,A,B, W[1].z+ K[6]);
	W[1].w = block1.w;
	RND(B,C,D,E,F,G,H,A, W[1].w+ K[7]);
	
	W[2].x = block2.x;
	RND(A,B,C,D,E,F,G,H, W[2].x+ K[8]);
	W[2].y = block2.y;
	RND(H,A,B,C,D,E,F,G, W[2].y+ K[9]);
	W[2].z = block2.z;
	RND(G,H,A,B,C,D,E,F, W[2].z+ K[10]);
	W[2].w = block2.w;
	RND(F,G,H,A,B,C,D,E, W[2].w+ K[11]);
	
	W[3].x = block3.x;
	RND(E,F,G,H,A,B,C,D, W[3].x+ K[12]);
	W[3].y = block3.y;
	RND(D,E,F,G,H,A,B,C, W[3].y+ K[13]);
	W[3].z = block3.z;
	RND(C,D,E,F,G,H,A,B, W[3].z+ K[14]);
	W[3].w = block3.w;
	RND(B,C,D,E,F,G,H,A, W[3].w+ K[76]);

	W[0].x += Wr1(W[3].z) + W[2].y + Wr2(W[0].y);
	RND(A,B,C,D,E,F,G,H, W[0].x+ K[15]);

	W[0].y += Wr1(W[3].w) + W[2].z + Wr2(W[0].z);
	RND(H,A,B,C,D,E,F,G, W[0].y+ K[16]);

	W[0].z += Wr1(W[0].x) + W[2].w + Wr2(W[0].w);
	RND(G,H,A,B,C,D,E,F, W[0].z+ K[17]);

	W[0].w += Wr1(W[0].y) + W[3].x + Wr2(W[1].x);
	RND(F,G,H,A,B,C,D,E, W[0].w+ K[18]);

	W[1].x += Wr1(W[0].z) + W[3].y + Wr2(W[1].y);
	RND(E,F,G,H,A,B,C,D, W[1].x+ K[19]);

	W[1].y += Wr1(W[0].w) + W[3].z + Wr2(W[1].z);
	RND(D,E,F,G,H,A,B,C, W[1].y+ K[20]);

	W[1].z += Wr1(W[1].x) + W[3].w + Wr2(W[1].w);
	RND(C,D,E,F,G,H,A,B, W[1].z+ K[21]);

	W[1].w += Wr1(W[1].y) + W[0].x + Wr2(W[2].x);
	RND(B,C,D,E,F,G,H,A, W[1].w+ K[22]);

	W[2].x += Wr1(W[1].z) + W[0].y + Wr2(W[2].y);
	RND(A,B,C,D,E,F,G,H, W[2].x+ K[23]);

	W[2].y += Wr1(W[1].w) + W[0].z + Wr2(W[2].z);
	RND(H,A,B,C,D,E,F,G, W[2].y+ K[24]);

	W[2].z += Wr1(W[2].x) + W[0].w + Wr2(W[2].w);
	RND(G,H,A,B,C,D,E,F, W[2].z+ K[25]);

	W[2].w += Wr1(W[2].y) + W[1].x + Wr2(W[3].x);
	RND(F,G,H,A,B,C,D,E, W[2].w+ K[26]);

	W[3].x += Wr1(W[2].z) + W[1].y + Wr2(W[3].y);
	RND(E,F,G,H,A,B,C,D, W[3].x+ K[27]);

	W[3].y += Wr1(W[2].w) + W[1].z + Wr2(W[3].z);
	RND(D,E,F,G,H,A,B,C, W[3].y+ K[28]);

	W[3].z += Wr1(W[3].x) + W[1].w + Wr2(W[3].w);
	RND(C,D,E,F,G,H,A,B, W[3].z+ K[29]);

	W[3].w += Wr1(W[3].y) + W[2].x + Wr2(W[0].x);
	RND(B,C,D,E,F,G,H,A, W[3].w+ K[30]);

	W[0].x += Wr1(W[3].z) + W[2].y + Wr2(W[0].y);
	RND(A,B,C,D,E,F,G,H, W[0].x+ K[31]);

	W[0].y += Wr1(W[3].w) + W[2].z + Wr2(W[0].z);
	RND(H,A,B,C,D,E,F,G, W[0].y+ K[32]);

	W[0].z += Wr1(W[0].x) + W[2].w + Wr2(W[0].w);
	RND(G,H,A,B,C,D,E,F, W[0].z+ K[33]);

	W[0].w += Wr1(W[0].y) + W[3].x + Wr2(W[1].x);
	RND(F,G,H,A,B,C,D,E, W[0].w+ K[34]);

	W[1].x += Wr1(W[0].z) + W[3].y + Wr2(W[1].y);
	RND(E,F,G,H,A,B,C,D, W[1].x+ K[35]);

	W[1].y += Wr1(W[0].w) + W[3].z + Wr2(W[1].z);
	RND(D,E,F,G,H,A,B,C, W[1].y+ K[36]);

	W[1].z += Wr1(W[1].x) + W[3].w + Wr2(W[1].w);
	RND(C,D,E,F,G,H,A,B, W[1].z+ K[37]);

	W[1].w += Wr1(W[1].y) + W[0].x + Wr2(W[2].x);
	RND(B,C,D,E,F,G,H,A, W[1].w+ K[38]);

	W[2].x += Wr1(W[1].z) + W[0].y + Wr2(W[2].y);
	RND(A,B,C,D,E,F,G,H, W[2].x+ K[39]);

	W[2].y += Wr1(W[1].w) + W[0].z + Wr2(W[2].z);
	RND(H,A,B,C,D,E,F,G, W[2].y+ K[40]);

	W[2].z += Wr1(W[2].x) + W[0].w + Wr2(W[2].w);
	RND(G,H,A,B,C,D,E,F, W[2].z+ K[41]);

	W[2].w += Wr1(W[2].y) + W[1].x + Wr2(W[3].x);
	RND(F,G,H,A,B,C,D,E, W[2].w+ K[42]);

	W[3].x += Wr1(W[2].z) + W[1].y + Wr2(W[3].y);
	RND(E,F,G,H,A,B,C,D, W[3].x+ K[43]);

	W[3].y += Wr1(W[2].w) + W[1].z + Wr2(W[3].z);
	RND(D,E,F,G,H,A,B,C, W[3].y+ K[44]);

	W[3].z += Wr1(W[3].x) + W[1].w + Wr2(W[3].w);
	RND(C,D,E,F,G,H,A,B, W[3].z+ K[45]);

	W[3].w += Wr1(W[3].y) + W[2].x + Wr2(W[0].x);
	RND(B,C,D,E,F,G,H,A, W[3].w+ K[46]);

	W[0].x += Wr1(W[3].z) + W[2].y + Wr2(W[0].y);
	RND(A,B,C,D,E,F,G,H, W[0].x+ K[47]);

	W[0].y += Wr1(W[3].w) + W[2].z + Wr2(W[0].z);
	RND(H,A,B,C,D,E,F,G, W[0].y+ K[48]);

	W[0].z += Wr1(W[0].x) + W[2].w + Wr2(W[0].w);
	RND(G,H,A,B,C,D,E,F, W[0].z+ K[49]);

	W[0].w += Wr1(W[0].y) + W[3].x + Wr2(W[1].x);
	RND(F,G,H,A,B,C,D,E, W[0].w+ K[50]);

	W[1].x += Wr1(W[0].z) + W[3].y + Wr2(W[1].y);
	RND(E,F,G,H,A,B,C,D, W[1].x+ K[51]);

	W[1].y += Wr1(W[0].w) + W[3].z + Wr2(W[1].z);
	RND(D,E,F,G,H,A,B,C, W[1].y+ K[52]);

	W[1].z += Wr1(W[1].x) + W[3].w + Wr2(W[1].w);
	RND(C,D,E,F,G,H,A,B, W[1].z+ K[53]);

	W[1].w += Wr1(W[1].y) + W[0].x + Wr2(W[2].x);
	RND(B,C,D,E,F,G,H,A, W[1].w+ K[54]);

	W[2].x += Wr1(W[1].z) + W[0].y + Wr2(W[2].y);
	RND(A,B,C,D,E,F,G,H, W[2].x+ K[55]);

	W[2].y += Wr1(W[1].w) + W[0].z + Wr2(W[2].z);
	RND(H,A,B,C,D,E,F,G, W[2].y+ K[56]);

	W[2].z += Wr1(W[2].x) + W[0].w + Wr2(W[2].w);
	RND(G,H,A,B,C,D,E,F, W[2].z+ K[57]);

	W[2].w += Wr1(W[2].y) + W[1].x + Wr2(W[3].x);
	RND(F,G,H,A,B,C,D,E, W[2].w+ K[58]);

	W[3].x += Wr1(W[2].z) + W[1].y + Wr2(W[3].y);
	RND(E,F,G,H,A,B,C,D, W[3].x+ K[59]);

	W[3].y += Wr1(W[2].w) + W[1].z + Wr2(W[3].z);
	RND(D,E,F,G,H,A,B,C, W[3].y+ K[60]);

	W[3].z += Wr1(W[3].x) + W[1].w + Wr2(W[3].w);
	RND(C,D,E,F,G,H,A,B, W[3].z+ K[61]);

	W[3].w += Wr1(W[3].y) + W[2].x + Wr2(W[0].x);
	RND(B,C,D,E,F,G,H,A, W[3].w+ K[62]);
	
#undef A
#undef B
#undef C
#undef D
#undef E
#undef F
#undef G
#undef H

	*state0 += (uint4)(K[73], K[77], K[78], K[79]);
	*state1 += (uint4)(K[66], K[67], K[80], K[81]);
}

__constant uint fixedW[64] =
{
	0x428a2f99,0xf1374491,0xb5c0fbcf,0xe9b5dba5,0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5,
	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3,0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf794,
	0xf59b89c2,0x73924787,0x23c6886e,0xa42ca65c,0x15ed3627,0x4d6edcbf,0xe28217fc,0xef02488f,
	0xb707775c,0x0468c23f,0xe7e72b4c,0x49e1f1a2,0x4b99c816,0x926d1570,0xaa0fc072,0xadb36e2c,
	0xad87a3ea,0xbcb1d3a3,0x7b993186,0x562b9420,0xbff3ca0c,0xda4b0c23,0x6cd8711a,0x8f337caa,
	0xc91b1417,0xc359dce1,0xa83253a7,0x3b13c12d,0x9d3d725d,0xd9031a84,0xb1a03340,0x16f58012,
	0xe64fb6a2,0xe84d923a,0xe93a5730,0x09837686,0x078ff753,0x29833341,0xd5de0b7e,0x6948ccf4,
	0xe0a1adbe,0x7c728e11,0x511c78e4,0x315b45bd,0xfca71413,0xea28f96a,0x79703128,0x4e1ef848,
};

void SHA256_fixed(uint4*restrict state0,uint4*restrict state1)
{
	uint4 S0 = *state0;
	uint4 S1 = *state1;

#define A S0.x
#define B S0.y
#define C S0.z
#define D S0.w
#define E S1.x
#define F S1.y
#define G S1.z
#define H S1.w

	RND(A,B,C,D,E,F,G,H, fixedW[0]);
	RND(H,A,B,C,D,E,F,G, fixedW[1]);
	RND(G,H,A,B,C,D,E,F, fixedW[2]);
	RND(F,G,H,A,B,C,D,E, fixedW[3]);
	RND(E,F,G,H,A,B,C,D, fixedW[4]);
	RND(D,E,F,G,H,A,B,C, fixedW[5]);
	RND(C,D,E,F,G,H,A,B, fixedW[6]);
	RND(B,C,D,E,F,G,H,A, fixedW[7]);
	RND(A,B,C,D,E,F,G,H, fixedW[8]);
	RND(H,A,B,C,D,E,F,G, fixedW[9]);
	RND(G,H,A,B,C,D,E,F, fixedW[10]);
	RND(F,G,H,A,B,C,D,E, fixedW[11]);
	RND(E,F,G,H,A,B,C,D, fixedW[12]);
	RND(D,E,F,G,H,A,B,C, fixedW[13]);
	RND(C,D,E,F,G,H,A,B, fixedW[14]);
	RND(B,C,D,E,F,G,H,A, fixedW[15]);
	RND(A,B,C,D,E,F,G,H, fixedW[16]);
	RND(H,A,B,C,D,E,F,G, fixedW[17]);
	RND(G,H,A,B,C,D,E,F, fixedW[18]);
	RND(F,G,H,A,B,C,D,E, fixedW[19]);
	RND(E,F,G,H,A,B,C,D, fixedW[20]);
	RND(D,E,F,G,H,A,B,C, fixedW[21]);
	RND(C,D,E,F,G,H,A,B, fixedW[22]);
	RND(B,C,D,E,F,G,H,A, fixedW[23]);
	RND(A,B,C,D,E,F,G,H, fixedW[24]);
	RND(H,A,B,C,D,E,F,G, fixedW[25]);
	RND(G,H,A,B,C,D,E,F, fixedW[26]);
	RND(F,G,H,A,B,C,D,E, fixedW[27]);
	RND(E,F,G,H,A,B,C,D, fixedW[28]);
	RND(D,E,F,G,H,A,B,C, fixedW[29]);
	RND(C,D,E,F,G,H,A,B, fixedW[30]);
	RND(B,C,D,E,F,G,H,A, fixedW[31]);
	RND(A,B,C,D,E,F,G,H, fixedW[32]);
	RND(H,A,B,C,D,E,F,G, fixedW[33]);
	RND(G,H,A,B,C,D,E,F, fixedW[34]);
	RND(F,G,H,A,B,C,D,E, fixedW[35]);
	RND(E,F,G,H,A,B,C,D, fixedW[36]);
	RND(D,E,F,G,H,A,B,C, fixedW[37]);
	RND(C,D,E,F,G,H,A,B, fixedW[38]);
	RND(B,C,D,E,F,G,H,A, fixedW[39]);
	RND(A,B,C,D,E,F,G,H, fixedW[40]);
	RND(H,A,B,C,D,E,F,G, fixedW[41]);
	RND(G,H,A,B,C,D,E,F, fixedW[42]);
	RND(F,G,H,A,B,C,D,E, fixedW[43]);
	RND(E,F,G,H,A,B,C,D, fixedW[44]);
	RND(D,E,F,G,H,A,B,C, fixedW[45]);
	RND(C,D,E,F,G,H,A,B, fixedW[46]);
	RND(B,C,D,E,F,G,H,A, fixedW[47]);
	RND(A,B,C,D,E,F,G,H, fixedW[48]);
	RND(H,A,B,C,D,E,F,G, fixedW[49]);
	RND(G,H,A,B,C,D,E,F, fixedW[50]);
	RND(F,G,H,A,B,C,D,E, fixedW[51]);
	RND(E,F,G,H,A,B,C,D, fixedW[52]);
	RND(D,E,F,G,H,A,B,C, fixedW[53]);
	RND(C,D,E,F,G,H,A,B, fixedW[54]);
	RND(B,C,D,E,F,G,H,A, fixedW[55]);
	RND(A,B,C,D,E,F,G,H, fixedW[56]);
	RND(H,A,B,C,D,E,F,G, fixedW[57]);
	RND(G,H,A,B,C,D,E,F, fixedW[58]);
	RND(F,G,H,A,B,C,D,E, fixedW[59]);
	RND(E,F,G,H,A,B,C,D, fixedW[60]);
	RND(D,E,F,G,H,A,B,C, fixedW[61]);
	RND(C,D,E,F,G,H,A,B, fixedW[62]);
	RND(B,C,D,E,F,G,H,A, fixedW[63]);
	
#undef A
#undef B
#undef C
#undef D
#undef E
#undef F
#undef G
#undef H
	*state0 += S0;
	*state1 += S1;
}

void shittify(uint4 B[8])
{
	uint4 tmp[4];
	tmp[0] = (uint4)(B[1].x,B[2].y,B[3].z,B[0].w);
	tmp[1] = (uint4)(B[2].x,B[3].y,B[0].z,B[1].w);
	tmp[2] = (uint4)(B[3].x,B[0].y,B[1].z,B[2].w);
	tmp[3] = (uint4)(B[0].x,B[1].y,B[2].z,B[3].w);
	
#pragma unroll
	for(uint i=0; i<4; ++i)
		B[i] = EndianSwap(tmp[i]);

	tmp[0] = (uint4)(B[5].x,B[6].y,B[7].z,B[4].w);
	tmp[1] = (uint4)(B[6].x,B[7].y,B[4].z,B[5].w);
	tmp[2] = (uint4)(B[7].x,B[4].y,B[5].z,B[6].w);
	tmp[3] = (uint4)(B[4].x,B[5].y,B[6].z,B[7].w);
	
#pragma unroll
	for(uint i=0; i<4; ++i)
		B[i+4] = EndianSwap(tmp[i]);
}

void unshittify(uint4 B[8])
{
	uint4 tmp[4];
	tmp[0] = (uint4)(B[3].x,B[2].y,B[1].z,B[0].w);
	tmp[1] = (uint4)(B[0].x,B[3].y,B[2].z,B[1].w);
	tmp[2] = (uint4)(B[1].x,B[0].y,B[3].z,B[2].w);
	tmp[3] = (uint4)(B[2].x,B[1].y,B[0].z,B[3].w);
	
#pragma unroll
	for(uint i=0; i<4; ++i)
		B[i] = EndianSwap(tmp[i]);

	tmp[0] = (uint4)(B[7].x,B[6].y,B[5].z,B[4].w);
	tmp[1] = (uint4)(B[4].x,B[7].y,B[6].z,B[5].w);
	tmp[2] = (uint4)(B[5].x,B[4].y,B[7].z,B[6].w);
	tmp[3] = (uint4)(B[6].x,B[5].y,B[4].z,B[7].w);
	
#pragma unroll
	for(uint i=0; i<4; ++i)
		B[i+4] = EndianSwap(tmp[i]);
}

void salsa(uint4 B[8])
{
	uint4 w[4];

#pragma unroll
	for(uint i=0; i<4; ++i)
		w[i] = (B[i]^=B[i+4]);

#pragma unroll
	for(uint i=0; i<4; ++i)
	{
		w[0] ^= rotl(w[3]     +w[2]     , 7U);
		w[1] ^= rotl(w[0]     +w[3]     , 9U);
		w[2] ^= rotl(w[1]     +w[0]     ,13U);
		w[3] ^= rotl(w[2]     +w[1]     ,18U);
		w[2] ^= rotl(w[3].wxyz+w[0].zwxy, 7U);
		w[1] ^= rotl(w[2].wxyz+w[3].zwxy, 9U);
		w[0] ^= rotl(w[1].wxyz+w[2].zwxy,13U);
		w[3] ^= rotl(w[0].wxyz+w[1].zwxy,18U);
	}

#pragma unroll
	for(uint i=0; i<4; ++i)
		w[i] = (B[i+4]^=(B[i]+=w[i]));

#pragma unroll
	for(uint i=0; i<4; ++i)
	{
		w[0] ^= rotl(w[3]     +w[2]     , 7U);
		w[1] ^= rotl(w[0]     +w[3]     , 9U);
		w[2] ^= rotl(w[1]     +w[0]     ,13U);
		w[3] ^= rotl(w[2]     +w[1]     ,18U);
		w[2] ^= rotl(w[3].wxyz+w[0].zwxy, 7U);
		w[1] ^= rotl(w[2].wxyz+w[3].zwxy, 9U);
		w[0] ^= rotl(w[1].wxyz+w[2].zwxy,13U);
		w[3] ^= rotl(w[0].wxyz+w[1].zwxy,18U);
	}

#pragma unroll
	for(uint i=0; i<4; ++i)
		B[i+4] += w[i];
}

#define Coord(x,y,z) x+y*(x ## SIZE)+z*(y ## SIZE)*(x ## SIZE)
#define CO Coord(z,x,y)

void scrypt_core(uint4 X[8], __global uint4*restrict lookup)
{
	shittify(X);
	const uint zSIZE = 8;
	const uint ySIZE = (1024/LOOKUP_GAP+(1024%LOOKUP_GAP>0));
	const uint xSIZE = CONCURRENT_THREADS;
	uint x = get_global_id(0)%xSIZE;

	for(uint y=0; y<1024/LOOKUP_GAP; ++y)
	{
#pragma unroll
		for(uint z=0; z<zSIZE; ++z)
			lookup[CO] = X[z];

#pragma unroll
		for(uint i=0; i<LOOKUP_GAP; ++i) 
			salsa(X);
	}
#if (LOOKUP_GAP != 1) && (LOOKUP_GAP != 2) && (LOOKUP_GAP != 4) && (LOOKUP_GAP != 8)
	{
		uint y = (1024/LOOKUP_GAP);

#pragma unroll		
		for(uint z=0; z<zSIZE; ++z)
			lookup[CO] = X[z];

#pragma unroll	
		for(uint i=0; i<1024%LOOKUP_GAP; ++i)
			salsa(X); 
	}
#endif

for (uint i=0; i<1024; ++i) 
	{
		uint4 V[8];
		uint j = X[7].x & K[85];
		uint y = (j/LOOKUP_GAP);
#pragma unroll
		for(uint z=0; z<zSIZE; ++z)
			V[z] = lookup[CO];

#if (LOOKUP_GAP == 1)
#elif (LOOKUP_GAP == 2)

		if (j&1)
			salsa(V);
#else
		uint val = j%LOOKUP_GAP;
#pragma unroll
		for (uint z=0; z<val; ++z) 
			salsa(V);
#endif

#pragma unroll
		for(uint z=0; z<zSIZE; ++z)
			X[z] ^= V[z];
		salsa(X);
	}
	unshittify(X);
}

#define SCRYPT_FOUND (0xFF)
#define SETFOUND(Xnonce) output[output[SCRYPT_FOUND]++] = Xnonce

__attribute__((reqd_work_group_size(WORKSIZE, 1, 1)))
__kernel void search(__global const uint4 * restrict input,
volatile __global uint*restrict output, __global uint4*restrict padcache,
const uint4 midstate0, const uint4 midstate16, const uint target)
{
	uint gid = get_global_id(0);
	uint4 X[8];
	uint4 tstate0, tstate1, ostate0, ostate1, tmp0, tmp1;
	uint4 data = (uint4)(input[4].x,input[4].y,input[4].z,gid);
	uint4 pad0 = midstate0, pad1 = midstate16;

	SHA256(&pad0,&pad1, data, (uint4)(K[84],0,0,0), (uint4)(0,0,0,0), (uint4)(0,0,0, K[86]));
	SHA256_fresh(&ostate0,&ostate1, pad0^ K[82], pad1^ K[82], K[82], K[82]);
	SHA256_fresh(&tstate0,&tstate1, pad0^ K[83], pad1^ K[83], K[83], K[83]);

	tmp0 = tstate0;
	tmp1 = tstate1;
	SHA256(&tstate0, &tstate1, input[0],input[1],input[2],input[3]);

#pragma unroll
	for (uint i=0; i<4; i++) 
	{
		pad0 = tstate0;
		pad1 = tstate1;
		X[i*2 ] = ostate0;
		X[i*2+1] = ostate1;

		SHA256(&pad0,&pad1, data, (uint4)(i+1,K[84],0,0), (uint4)(0,0,0,0), (uint4)(0,0,0, K[87]));
		SHA256(X+i*2,X+i*2+1, pad0, pad1, (uint4)(K[84], 0U, 0U, 0U), (uint4)(0U, 0U, 0U, K[88]));
	}
	scrypt_core(X,padcache);
	SHA256(&tmp0,&tmp1, X[0], X[1], X[2], X[3]);
	SHA256(&tmp0,&tmp1, X[4], X[5], X[6], X[7]);
	SHA256_fixed(&tmp0,&tmp1);
	SHA256(&ostate0,&ostate1, tmp0, tmp1, (uint4)(K[84], 0U, 0U, 0U), (uint4)(0U, 0U, 0U, K[88]));

	bool result = (EndianSwap(ostate1.w) <= target);


	if (result)
		SETFOUND(gid);
}

...you "make clean", "make", remove old .bin files from your cgminer dir (rm *.bin) and run cgminer again to create new .bins.

Check your current speed and intensity to see where you are at right now and then please report back with results of the new kernel. I'm curious what it does for 6xxx / 7xxx and R9's which have a different architecture than series 5xxx.

If it worked for you and you liked it, well, my donation addresses are listed below

[Bitcoin and The Crashing Game]

Bitcoin and The Crashing Game

If you are reading this after the latest “bad news” that drove Bitcoin to lose another 10% or more of its price, maybe you should be informed of the fundamentals regarding its price manipulation.

“Price manipulation? What price manipulation? Is it rigged?”

Well, yes, it is rigged.

Newsflash: EVERY SINGLE MARKET ON THE PLANET IS RIGGED.

Interest rates and monetary supply? Rigged.

Precious Metals? Rigged.

Other commodities? Rigged.

Houses and land? Rigged.

Energy? Rigged.

Everything is rigged.

The only thing different is the method employed for fixing each market. Given all this market-fixing why would anyone believe that Bitcoin is beyond manipulation?

Initially Bitcoin hadn't attracted as much attention. But since people took it seriously and it came to threaten paper currency it naturally became a target of extremely powerful interests that had the need to control it, destroy it, or make people lose faith in the wider cryptocurrency ecosystem which is like burning the entire forest instead of just cutting down a tree. The people behind those powerful interests need to see people getting “burned” and bail. When they achieve that they can pat themselves on the back for keeping the sheep in conformity to the established system that is the only one “worthy” of people's trust. This established system is their system of hierarchical power, with them sitting atop the pyramid.

So what is the method employed for fixing the Bitcoin market?

Bitcoin was genius in its inception because it “trolled” the system against itself.

1) Who could argue that Bitcoin is worthless when its benchmark is paper currency that is even more worthless: Just a paper that does not correspond to gold or silver. On top of that it is issued to infinity instead of being rare and increasingly hard to acquire.

2) How could the elite control Bitcoin through financial instruments with “futures”, “shorting” etc without de facto legalizing it? How could they do that without giving Bitcoin a tremendous legitimacy and desirability boost if it is traded in the “big” markets?

3) How could they tax it without acknowledging it as having genuine value? If it is illegal then one cannot be taxed for it. If it is taxed then everyone can say “I'm paying my taxes and I am legally owning and using this”.


Yet, the absence of #2 (financial instruments) and the convenient gray areas of #3 (whether it's legal or illegal to own, trade, etc or whether its taxable or non-taxable) is the fertile ground from where the manipulation scheme arises.

In finance the second most desirable ability is knowing what is going to happen. However the most desirable ability of all is being in a position to actually control what is going to happen.

Most people are not ideally positioned to practice either ability. They live in a state where their financial moves are intertwined with uncertainty, fear, pressure, anxiety, hope, rational and irrational expectations given one's own understanding, etc. As such they are simple “dancers” to the tune that is played by those who control what is going to happen.

Those who control politics, economy and the media can, for all intents and purposes, create a “disaster story” or an “announcement” about Bitcoin every now and then. “The politicians took the X negative decision” or “the financial authorities tighten the rules for money laundering”, “the FBI raided that exchange” etc...

The dynamic is clear: “Quick, dump your coins so we can buy them from you for pennies”. The “sheep”, living in uncertainty about the grander design, naturally feel that they have to make a decision regarding the latest “news”. Do they bail, or do they stay? Knowing not what is going to happen, they have to balance their conservative side, their hopeful side, their greedy side etc. So some will be selling and sending the price down, creating avalanche-type ripples.

This process is repeated in cycles, as a long-term strategy, where the value of the coin is degraded temporarily for two-digit percentages in order for those who want to control the cryptocurrency landscape to be able to buy more Bitcoins without driving the price to astronomical levels. Such an increase in price would also increase the desirability factor which is the total opposite of the result they are trying to create with the frequent scares.

For those in power who want to control the BTC market, a starting capital of just 100.000 Bitcoins  and only 10% drops in price for each crash-cycle is enough to assure success in the long-run.

Starting Capital: 100.000 BTCs

End of Cycle 1 of scare tactics: 110.000 BTCs (+10% gained or 10.000 BTCs by the drop-cycle for cheap).
End of Cycle 10 of scare tactics: 259.000 BTCs
End of Cycle 20 of scare tactics: 672.000 BTCs
End of Cycle 30 of scare tactics: 1.744.000 BTCs
End of Cycle 40 of scare tactics: 4.525.000 BTCs
End of Cycle 50 of scare tactics: 11.739.000 BTCs (close to the total market supply right now)

This is just an illustration of how the mechanism works – it is expected that some will not sell their coins, especially as the supply of the “weak hands” is diminished. This will necessitate a more sinister approach with an even more deliberate attack on cryptocurrencies and even more negative news that have the potential to shake even the “stronger hands” with the rationale that “others will be selling too, even if I am not selling”.

The buying cycles of those who desire to control the BTC supply are performed through a mix of cash use and cash gained from dumping coins during the upward phase. As the numbers escalate it is increasingly difficult to dump, say, 5 million BTCs and preserve the uptrend and then, after creating the crash, to come and buy back 5.5 million. That's why cash/fiat is essential.

In the end, the accrual of a large quantity of BTC leads to controlling the entire cryptocurrency ecosystem due the trading of other cryptocurrencies in BTC. The BTCs that ended up to users of other cryptocurrencies that sold out their coins for BTCs will find their way back to the market controllers, sooner or later, through the scare mechanism. In this way an inexpensive way to control BTC and the alternate cryptocurrencies is achieved.

By controlling a large share of coins in cryptocurrencies (through owning a large BTC wallet) they can make sure that they are able to smash them at will and shatter people's confidence in them. This is more easily performed in alternative cryptocurrencies which are way more volatile and fragile than Bitcoin.

If you are asking “what to do about this?” I will tell you that it is up to you to decide how you utilize this information. It is also entirely possible that if most people become aware of how this strategy works, the strategy may be adjusted to something else.

For the time being the bad-news-game is being played massively and it is also building up characteristics of psychological conformity to an anchor/trigger situation where people are conditioned to react negatively and expect the worst based on prior events (that “seem” to repeat inside the market).

In a way, people are conforming to the same anchor/trigger pattern where a dog salivates when he hears the dish where his food is served. Every cycle of bad news-market crash will calcify an even deeper emotional sense of impending doom as the “anchor” and “trigger” mechanism are reinforced. Some will feel it as a panic, others may feel it in their gut etc. Awareness of how this works on one's own self is also useful although emotions are harder to reason with.

You can bookmark or share this with anyone you like and you can also contribute your own thoughts and experiences below. You can also make informative videos on YouTube explaining the situation of how some can dominate the BTC market and the alt-crypto ecosystem by simply controlling the flow of events and news/announcements etc.

[5.5bn]

Initially I was skeptical on how a virtual currency can compete with traditional commodities like Gold and Silver. But I came to realize the following:

Existing Above Ground Gold Quantity: 170kt or 5.5bn ounces increasing by ~90mn ounces per year.
Existing Above Ground Silver Quantity: ~900kt or 29bn ounces increasing by 700mn ounces per year.

Contrast this with finite numbers of 20 or 80 million coins and you start to realize why those who have lots of money want to get the largest possible piece of the pie in something that is "so limited" in quantity, driving the prices up - and probably way more than gold. It's a "long-run" issue, in which the scarcity of Bitcoins or Litecoins will make them more "valuable". Who knows, we may witness market caps in the trillions for the virtual currencies, whatever that means for their value per unit.

Hi, I need help from a C programmer to try and optimize Pooler's cpuminer through the compiler, as my C knowledge fails me.

At the moment there is an embedded ASM optimization that is activated by IFs in the program. While these ASM optimizations provide a decent increase in speed, I need to shut these down so that I can compile this through various C compilers that can -perhaps- go faster still, hopefully without breaking the program. So what I want is step-by-step instructions on how to shut these ASM optimizations down. The target is an intel x64 (core2).