//removes the children of any node with less than limit work
void SolverPNS2::garbage_collect(PNSNode * node){
PNSNode * child = node->children.begin();
PNSNode * end = node->children.end();
for( ; child != end; child++){
if(child->terminal()){ //solved
//log heavy nodes?
PLUS(nodes, -child->dealloc(ctmem));
}else if(child->work < gclimit){ //low work, ignore solvedness since it's trivial to re-solve
PLUS(nodes, -child->dealloc(ctmem));
}else if(child->children.num() > 0){
garbage_collect(child);
}
}
}
void main(void)
{
int delay;
char ch;
delay = DEF_TIME;
m.load(word_list);
if (setup())
return;
while(kbhit())
getch();
do
{
while (! kbhit())
message(delay);
ch = getch();
if (MINUS(ch) &&
(delay >= MIN_TIME))
delay -= INC_TIME;
if (PLUS(ch) &&
(delay <= MAX_TIME))
delay += INC_TIME;
}
while (ch == '-' || ch == '+');
cleanup();
cputs("\n\n\n(C) 1995 N.A.A. MATHEWSON");
}
void AgentMCTS::create_children_simple(const Board & board, Node * node){
assert(node->children.empty());
node->children.alloc(board.moves_avail(), ctmem);
Node * child = node->children.begin(),
* end = node->children.end();
MoveIterator moveit(board, prunesymmetry);
int nummoves = 0;
for(; !moveit.done() && child != end; ++moveit, ++child){
*child = Node(*moveit);
nummoves++;
}
if(prunesymmetry)
node->children.shrink(nummoves); //shrink the node to ignore the extra moves
else //both end conditions should happen in parallel
assert(moveit.done() && child == end);
PLUS(nodes, node->children.num());
}
bool AgentPNS::PNSThread::pns(const Board & board, Node * node, int depth, uint32_t tp, uint32_t td){
iters++;
if(agent->maxdepth < depth)
agent->maxdepth = depth;
if(node->children.empty()){
if(node->terminal())
return true;
if(agent->ctmem.memalloced() >= agent->memlimit)
return false;
if(!node->children.lock())
return false;
int numnodes = board.moves_avail();
CompactTree<Node>::Children temp;
temp.alloc(numnodes, agent->ctmem);
unsigned int i = 0;
unsigned int seen = 0;
for(MoveIterator move(board); !move.done(); ++move){
int outcome = solve1ply(move.board(), seen);
unsigned int pd = 1;
temp[i] = Node(*move).outcome(outcome, board.toplay(), agent->ties, pd);
i++;
}
PLUS(agent->nodes, i);
temp.shrink(i); //if symmetry, there may be extra moves to ignore
node->children.swap(temp);
assert(temp.unlock());
PLUS(agent->nodes_seen, seen);
updatePDnum(node);
return true;
}
bool mem;
do{
Node * child = node->children.begin(),
* child2 = node->children.begin(),
* childend = node->children.end();
uint32_t tpc, tdc;
if(agent->df){
for(Node * i = node->children.begin(); i != childend; i++){
if(i->refdelta() <= child->refdelta()){
child2 = child;
child = i;
}else if(i->refdelta() < child2->refdelta()){
child2 = i;
}
}
tpc = min(INF32/2, (td + child->phi - node->delta));
tdc = min(tp, (uint32_t)(child2->delta*(1.0 + agent->epsilon) + 1));
}else{
tpc = tdc = 0;
for(Node * i = node->children.begin(); i != childend; i++)
if(child->refdelta() > i->refdelta())
child = i;
}
Board next = board;
next.move(child->move);
child->ref();
uint64_t itersbefore = iters;
mem = pns(next, child, depth + 1, tpc, tdc);
child->deref();
PLUS(child->work, iters - itersbefore);
if(updatePDnum(node) && !agent->df)
break;
}while(!agent->timeout && mem && (!agent->df || (node->phi < tp && node->delta < td)));
return mem;
}
static void salsa20_wordtobyte(u8 output[64],const u32 input[16])
{
u32 x[16];
unsigned int i;
for (i = 0;i < 16;++i) x[i] = input[i];
for (i = 20;i > 0;i -= 2) {
x[ 4] = XOR(x[ 4],ROTATE(PLUS(x[ 0],x[12]), 7));
x[ 8] = XOR(x[ 8],ROTATE(PLUS(x[ 4],x[ 0]), 9));
x[12] = XOR(x[12],ROTATE(PLUS(x[ 8],x[ 4]),13));
x[ 0] = XOR(x[ 0],ROTATE(PLUS(x[12],x[ 8]),18));
x[ 9] = XOR(x[ 9],ROTATE(PLUS(x[ 5],x[ 1]), 7));
x[13] = XOR(x[13],ROTATE(PLUS(x[ 9],x[ 5]), 9));
x[ 1] = XOR(x[ 1],ROTATE(PLUS(x[13],x[ 9]),13));
x[ 5] = XOR(x[ 5],ROTATE(PLUS(x[ 1],x[13]),18));
x[14] = XOR(x[14],ROTATE(PLUS(x[10],x[ 6]), 7));
x[ 2] = XOR(x[ 2],ROTATE(PLUS(x[14],x[10]), 9));
x[ 6] = XOR(x[ 6],ROTATE(PLUS(x[ 2],x[14]),13));
x[10] = XOR(x[10],ROTATE(PLUS(x[ 6],x[ 2]),18));
x[ 3] = XOR(x[ 3],ROTATE(PLUS(x[15],x[11]), 7));
x[ 7] = XOR(x[ 7],ROTATE(PLUS(x[ 3],x[15]), 9));
x[11] = XOR(x[11],ROTATE(PLUS(x[ 7],x[ 3]),13));
x[15] = XOR(x[15],ROTATE(PLUS(x[11],x[ 7]),18));
x[ 1] = XOR(x[ 1],ROTATE(PLUS(x[ 0],x[ 3]), 7));
x[ 2] = XOR(x[ 2],ROTATE(PLUS(x[ 1],x[ 0]), 9));
x[ 3] = XOR(x[ 3],ROTATE(PLUS(x[ 2],x[ 1]),13));
x[ 0] = XOR(x[ 0],ROTATE(PLUS(x[ 3],x[ 2]),18));
x[ 6] = XOR(x[ 6],ROTATE(PLUS(x[ 5],x[ 4]), 7));
x[ 7] = XOR(x[ 7],ROTATE(PLUS(x[ 6],x[ 5]), 9));
x[ 4] = XOR(x[ 4],ROTATE(PLUS(x[ 7],x[ 6]),13));
x[ 5] = XOR(x[ 5],ROTATE(PLUS(x[ 4],x[ 7]),18));
x[11] = XOR(x[11],ROTATE(PLUS(x[10],x[ 9]), 7));
x[ 8] = XOR(x[ 8],ROTATE(PLUS(x[11],x[10]), 9));
x[ 9] = XOR(x[ 9],ROTATE(PLUS(x[ 8],x[11]),13));
x[10] = XOR(x[10],ROTATE(PLUS(x[ 9],x[ 8]),18));
x[12] = XOR(x[12],ROTATE(PLUS(x[15],x[14]), 7));
x[13] = XOR(x[13],ROTATE(PLUS(x[12],x[15]), 9));
x[14] = XOR(x[14],ROTATE(PLUS(x[13],x[12]),13));
x[15] = XOR(x[15],ROTATE(PLUS(x[14],x[13]),18));
}
for (i = 0;i < 16;++i) x[i] = PLUS(x[i],input[i]);
for (i = 0;i < 16;++i) U32TO8_LITTLE(output + 4 * i,x[i]);
}
PEG_FUNC(parse_peg_open);
PEG_FUNC(parse_peg_close);
PEG_FUNC(parse_peg_dot);
PEG_FUNC(parse_peg_spacing);
PEG_FUNC(parse_peg_comment);
PEG_FUNC(parse_peg_space);
PEG_FUNC(parse_peg_eol);
PEG_FUNC(parse_peg_eof);
PEG_FUNC(parse_peg_hide);
/*************************************************/
/* grammar */
PEG_PARSE(parse_peg_grammar, PEG_GRAMMAR_NODE, L"PEG grammar", SEQ(T(parse_peg_spacing),
SEQ(PLUS(T(parse_peg_rule)),
T(parse_peg_eof))));
PEG_PARSE(parse_peg_rule, PEG_RULE_NODE, L"PEG rule", SEQ(T(parse_peg_identifier),
SEQ(QUES(T(parse_peg_literal)),
SEQ(T(parse_peg_left_arrow),
T(parse_peg_disjunction)))));
PEG_PARSE(parse_peg_disjunction, PEG_DISJ_NODE, L"expression", SEQ(T(parse_peg_conjunction), STAR(SEQ(T(parse_peg_slash), T(parse_peg_conjunction)))));
PEG_PARSE(parse_peg_conjunction, PEG_CONJ_NODE, L"term", PLUS(T(parse_peg_prefix_exp)));
PEG_PARSE(parse_peg_prefix_exp, PEG_PREFIX_EXP_NODE, L"term", SEQ(QUES(DISJ(T(parse_peg_and), DISJ(T(parse_peg_not), T(parse_peg_hide)))), T(parse_peg_suffix_exp)));
PEG_PARSE(parse_peg_suffix_exp, PEG_SUFFIX_EXP_NODE, L"term", SEQ(T(parse_peg_term),
QUES(DISJ(DISJ(T(parse_peg_question),
static void salsa20_wordtobyte(u8 output[64], const u32 input[16])
{
u32 x0, x1, x2, x3, x4, x5, x6, x7,
x8, x9, x10, x11, x12, x13, x14, x15;
int i;
x0 = input[0];
x1 = input[1];
x2 = input[2];
x3 = input[3];
x4 = input[4];
x5 = input[5];
x6 = input[6];
x7 = input[7];
x8 = input[8];
x9 = input[9];
x10 = input[10];
x11 = input[11];
x12 = input[12];
x13 = input[13];
x14 = input[14];
x15 = input[15];
for( i=20; i>0; i-=2 ) {
QUARTERROUND( x0, x4, x8,x12);
QUARTERROUND( x1, x5, x9,x13);
QUARTERROUND( x2, x6,x10,x14);
QUARTERROUND( x3, x7,x11,x15);
QUARTERROUND( x0, x5,x10,x15);
QUARTERROUND( x1, x6,x11,x12);
QUARTERROUND( x2, x7, x8,x13);
QUARTERROUND( x3, x4, x9,x14);
}
x0 = PLUS(x0, input[0]);
x1 = PLUS(x1, input[1]);
x2 = PLUS(x2, input[2]);
x3 = PLUS(x3, input[3]);
x4 = PLUS(x4, input[4]);
x5 = PLUS(x5, input[5]);
x6 = PLUS(x6, input[6]);
x7 = PLUS(x7, input[7]);
x8 = PLUS(x8, input[8]);
x9 = PLUS(x9, input[9]);
x10 = PLUS(x10, input[10]);
x11 = PLUS(x11, input[11]);
x12 = PLUS(x12, input[12]);
x13 = PLUS(x13, input[13]);
x14 = PLUS(x14, input[14]);
x15 = PLUS(x15, input[15]);
U32TO8_LITTLE(output + 0,x0);
U32TO8_LITTLE(output + 4,x1);
U32TO8_LITTLE(output + 8,x2);
U32TO8_LITTLE(output + 12,x3);
U32TO8_LITTLE(output + 16,x4);
U32TO8_LITTLE(output + 20,x5);
U32TO8_LITTLE(output + 24,x6);
U32TO8_LITTLE(output + 28,x7);
U32TO8_LITTLE(output + 32,x8);
U32TO8_LITTLE(output + 36,x9);
U32TO8_LITTLE(output + 40,x10);
U32TO8_LITTLE(output + 44,x11);
U32TO8_LITTLE(output + 48,x12);
U32TO8_LITTLE(output + 52,x13);
U32TO8_LITTLE(output + 56,x14);
U32TO8_LITTLE(output + 60,x15);
}
bool SolverPNS2::SolverThread::pns(const Board & board, PNSNode * node, int depth, uint32_t tp, uint32_t td){
iters++;
if(solver->maxdepth < depth)
solver->maxdepth = depth;
if(node->children.empty()){
if(node->terminal())
return true;
if(solver->ctmem.memalloced() >= solver->memlimit)
return false;
if(!node->children.lock())
return false;
int numnodes = board.movesremain();
CompactTree<PNSNode>::Children temp;
temp.alloc(numnodes, solver->ctmem);
PLUS(solver->nodes, numnodes);
if(solver->lbdist)
dists.run(&board);
int i = 0;
for(Board::MoveIterator move = board.moveit(true); !move.done(); ++move){
int outcome, pd;
if(solver->ab){
Board next = board;
next.move(*move);
pd = 0;
outcome = (solver->ab == 1 ? solve1ply(next, pd) : solve2ply(next, pd));
PLUS(solver->nodes_seen, pd);
}else{
outcome = board.test_win(*move);
pd = 1;
}
if(solver->lbdist && outcome < 0)
pd = dists.get(*move);
temp[i] = PNSNode(*move).outcome(outcome, board.toplay(), solver->ties, pd);
i++;
}
temp.shrink(i); //if symmetry, there may be extra moves to ignore
node->children.swap(temp);
assert(temp.unlock());
PLUS(solver->nodes_seen, i);
updatePDnum(node);
return true;
}
bool mem;
do{
PNSNode * child = node->children.begin(),
* child2 = node->children.begin(),
* childend = node->children.end();
uint32_t tpc, tdc;
if(solver->df){
for(PNSNode * i = node->children.begin(); i != childend; i++){
if(i->refdelta() <= child->refdelta()){
child2 = child;
child = i;
}else if(i->refdelta() < child2->refdelta()){
child2 = i;
}
}
tpc = min(INF32/2, (td + child->phi - node->delta));
tdc = min(tp, (uint32_t)(child2->delta*(1.0 + solver->epsilon) + 1));
}else{
tpc = tdc = 0;
for(PNSNode * i = node->children.begin(); i != childend; i++)
if(child->refdelta() > i->refdelta())
child = i;
}
Board next = board;
next.move(child->move);
child->ref();
uint64_t itersbefore = iters;
mem = pns(next, child, depth + 1, tpc, tdc);
child->deref();
PLUS(child->work, iters - itersbefore);
if(updatePDnum(node) && !solver->df)
break;
}while(!solver->timeout && mem && (!solver->df || (node->phi < tp && node->delta < td)));
return mem;
}
#define X() 1 #define PLUS() + #define MINUS() - #define DOT() . #define GREATER() > #define LESS() < //R #line 23 "t_9_003.cpp" X()2 //R 1 2 STRINGIZE( X()2 ) //R "12" //R X() 2 //R 1 2 STRINGIZE( X() 2 ) //R "1 2" //R PLUS()MINUS() //R +- STRINGIZE( PLUS()MINUS() ) //R "+-" //R PLUS()PLUS() //R + + STRINGIZE( PLUS()PLUS() ) //R "++" //R MINUS()MINUS() //R - - STRINGIZE( MINUS()MINUS() ) //R "--" //R DOT()DOT()DOT() //R .. . STRINGIZE( DOT()DOT()DOT() ) //R "..." // the following are regressions reported by Stefan Seefeld //R #line 43 "t_9_003.cpp" GREATER()GREATER() //R > > STRINGIZE( GREATER()GREATER() ) //R ">>" //R
/* SIMD format-compatible scalar code. This should be replaced by
more efficient code, but so far this is a proof of concept. */
static void salsa20_wordtobyte_tr(u8 output[64],const u32 input[16])
{
u32 x[16];
int i;
for (i = 0;i < 16;++i) x[tr[i]] = input[tr[i]];
for (i = 20;i > 0;i -= 2) {
x[tr[ 4]] = XOR(x[tr[ 4]],ROTATE(PLUS(x[tr[ 0]],x[tr[12]]), 7));
x[tr[ 8]] = XOR(x[tr[ 8]],ROTATE(PLUS(x[tr[ 4]],x[tr[ 0]]), 9));
x[tr[12]] = XOR(x[tr[12]],ROTATE(PLUS(x[tr[ 8]],x[tr[ 4]]),13));
x[tr[ 0]] = XOR(x[tr[ 0]],ROTATE(PLUS(x[tr[12]],x[tr[ 8]]),18));
x[tr[ 9]] = XOR(x[tr[ 9]],ROTATE(PLUS(x[tr[ 5]],x[tr[ 1]]), 7));
x[tr[13]] = XOR(x[tr[13]],ROTATE(PLUS(x[tr[ 9]],x[tr[ 5]]), 9));
x[tr[ 1]] = XOR(x[tr[ 1]],ROTATE(PLUS(x[tr[13]],x[tr[ 9]]),13));
x[tr[ 5]] = XOR(x[tr[ 5]],ROTATE(PLUS(x[tr[ 1]],x[tr[13]]),18));
x[tr[14]] = XOR(x[tr[14]],ROTATE(PLUS(x[tr[10]],x[tr[ 6]]), 7));
x[tr[ 2]] = XOR(x[tr[ 2]],ROTATE(PLUS(x[tr[14]],x[tr[10]]), 9));
x[tr[ 6]] = XOR(x[tr[ 6]],ROTATE(PLUS(x[tr[ 2]],x[tr[14]]),13));
x[tr[10]] = XOR(x[tr[10]],ROTATE(PLUS(x[tr[ 6]],x[tr[ 2]]),18));
x[tr[ 3]] = XOR(x[tr[ 3]],ROTATE(PLUS(x[tr[15]],x[tr[11]]), 7));
x[tr[ 7]] = XOR(x[tr[ 7]],ROTATE(PLUS(x[tr[ 3]],x[tr[15]]), 9));
x[tr[11]] = XOR(x[tr[11]],ROTATE(PLUS(x[tr[ 7]],x[tr[ 3]]),13));
x[tr[15]] = XOR(x[tr[15]],ROTATE(PLUS(x[tr[11]],x[tr[ 7]]),18));
x[tr[ 1]] = XOR(x[tr[ 1]],ROTATE(PLUS(x[tr[ 0]],x[tr[ 3]]), 7));
x[tr[ 2]] = XOR(x[tr[ 2]],ROTATE(PLUS(x[tr[ 1]],x[tr[ 0]]), 9));
x[tr[ 3]] = XOR(x[tr[ 3]],ROTATE(PLUS(x[tr[ 2]],x[tr[ 1]]),13));
x[tr[ 0]] = XOR(x[tr[ 0]],ROTATE(PLUS(x[tr[ 3]],x[tr[ 2]]),18));
x[tr[ 6]] = XOR(x[tr[ 6]],ROTATE(PLUS(x[tr[ 5]],x[tr[ 4]]), 7));
x[tr[ 7]] = XOR(x[tr[ 7]],ROTATE(PLUS(x[tr[ 6]],x[tr[ 5]]), 9));
x[tr[ 4]] = XOR(x[tr[ 4]],ROTATE(PLUS(x[tr[ 7]],x[tr[ 6]]),13));
x[tr[ 5]] = XOR(x[tr[ 5]],ROTATE(PLUS(x[tr[ 4]],x[tr[ 7]]),18));
x[tr[11]] = XOR(x[tr[11]],ROTATE(PLUS(x[tr[10]],x[tr[ 9]]), 7));
x[tr[ 8]] = XOR(x[tr[ 8]],ROTATE(PLUS(x[tr[11]],x[tr[10]]), 9));
x[tr[ 9]] = XOR(x[tr[ 9]],ROTATE(PLUS(x[tr[ 8]],x[tr[11]]),13));
x[tr[10]] = XOR(x[tr[10]],ROTATE(PLUS(x[tr[ 9]],x[tr[ 8]]),18));
x[tr[12]] = XOR(x[tr[12]],ROTATE(PLUS(x[tr[15]],x[tr[14]]), 7));
x[tr[13]] = XOR(x[tr[13]],ROTATE(PLUS(x[tr[12]],x[tr[15]]), 9));
x[tr[14]] = XOR(x[tr[14]],ROTATE(PLUS(x[tr[13]],x[tr[12]]),13));
x[tr[15]] = XOR(x[tr[15]],ROTATE(PLUS(x[tr[14]],x[tr[13]]),18));
}
for (i = 0;i < 16;++i) x[tr[i]] = PLUS(x[tr[i]],input[tr[i]]);
for (i = 0;i < 16;++i) U32TO8_LITTLE(output + 4 * i,x[tr[i]]);
}
void ECRYPT_encrypt_bytes(ECRYPT_ctx *x,const u8 *m,u8 *c,u32 bytes)
{
u32 x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15;
u32 j0, j1, j2, j3, j4, j5, j6, j7, j8, j9, j10, j11, j12, j13, j14, j15;
u8 *ctarget;
u8 tmp[64];
int i;
if (!bytes) return;
j0 = x->input[0];
j1 = x->input[1];
j2 = x->input[2];
j3 = x->input[3];
j4 = x->input[4];
j5 = x->input[5];
j6 = x->input[6];
j7 = x->input[7];
j8 = x->input[8];
j9 = x->input[9];
j10 = x->input[10];
j11 = x->input[11];
j12 = x->input[12];
j13 = x->input[13];
j14 = x->input[14];
j15 = x->input[15];
for (;;) {
if (bytes < 64) {
for (i = 0;i < bytes;++i) tmp[i] = m[i];
m = tmp;
ctarget = c;
c = tmp;
}
x0 = j0;
x1 = j1;
x2 = j2;
x3 = j3;
x4 = j4;
x5 = j5;
x6 = j6;
x7 = j7;
x8 = j8;
x9 = j9;
x10 = j10;
x11 = j11;
x12 = j12;
x13 = j13;
x14 = j14;
x15 = j15;
for (i = 8;i > 0;i -= 2) {
x4 = XOR( x4,ROTATE(PLUS( x0,x12), 7));
x8 = XOR( x8,ROTATE(PLUS( x4, x0), 9));
x12 = XOR(x12,ROTATE(PLUS( x8, x4),13));
x0 = XOR( x0,ROTATE(PLUS(x12, x8),18));
x9 = XOR( x9,ROTATE(PLUS( x5, x1), 7));
x13 = XOR(x13,ROTATE(PLUS( x9, x5), 9));
x1 = XOR( x1,ROTATE(PLUS(x13, x9),13));
x5 = XOR( x5,ROTATE(PLUS( x1,x13),18));
x14 = XOR(x14,ROTATE(PLUS(x10, x6), 7));
x2 = XOR( x2,ROTATE(PLUS(x14,x10), 9));
x6 = XOR( x6,ROTATE(PLUS( x2,x14),13));
x10 = XOR(x10,ROTATE(PLUS( x6, x2),18));
x3 = XOR( x3,ROTATE(PLUS(x15,x11), 7));
x7 = XOR( x7,ROTATE(PLUS( x3,x15), 9));
x11 = XOR(x11,ROTATE(PLUS( x7, x3),13));
x15 = XOR(x15,ROTATE(PLUS(x11, x7),18));
x1 = XOR( x1,ROTATE(PLUS( x0, x3), 7));
x2 = XOR( x2,ROTATE(PLUS( x1, x0), 9));
x3 = XOR( x3,ROTATE(PLUS( x2, x1),13));
x0 = XOR( x0,ROTATE(PLUS( x3, x2),18));
x6 = XOR( x6,ROTATE(PLUS( x5, x4), 7));
x7 = XOR( x7,ROTATE(PLUS( x6, x5), 9));
x4 = XOR( x4,ROTATE(PLUS( x7, x6),13));
x5 = XOR( x5,ROTATE(PLUS( x4, x7),18));
x11 = XOR(x11,ROTATE(PLUS(x10, x9), 7));
x8 = XOR( x8,ROTATE(PLUS(x11,x10), 9));
x9 = XOR( x9,ROTATE(PLUS( x8,x11),13));
x10 = XOR(x10,ROTATE(PLUS( x9, x8),18));
x12 = XOR(x12,ROTATE(PLUS(x15,x14), 7));
x13 = XOR(x13,ROTATE(PLUS(x12,x15), 9));
x14 = XOR(x14,ROTATE(PLUS(x13,x12),13));
x15 = XOR(x15,ROTATE(PLUS(x14,x13),18));
}
x0 = PLUS(x0,j0);
x1 = PLUS(x1,j1);
x2 = PLUS(x2,j2);
x3 = PLUS(x3,j3);
x4 = PLUS(x4,j4);
x5 = PLUS(x5,j5);
x6 = PLUS(x6,j6);
x7 = PLUS(x7,j7);
x8 = PLUS(x8,j8);
x9 = PLUS(x9,j9);
x10 = PLUS(x10,j10);
x11 = PLUS(x11,j11);
x12 = PLUS(x12,j12);
x13 = PLUS(x13,j13);
x14 = PLUS(x14,j14);
x15 = PLUS(x15,j15);
x0 = XOR(x0,U8TO32_LITTLE(m + 0));
x1 = XOR(x1,U8TO32_LITTLE(m + 4));
x2 = XOR(x2,U8TO32_LITTLE(m + 8));
x3 = XOR(x3,U8TO32_LITTLE(m + 12));
x4 = XOR(x4,U8TO32_LITTLE(m + 16));
x5 = XOR(x5,U8TO32_LITTLE(m + 20));
x6 = XOR(x6,U8TO32_LITTLE(m + 24));
x7 = XOR(x7,U8TO32_LITTLE(m + 28));
x8 = XOR(x8,U8TO32_LITTLE(m + 32));
x9 = XOR(x9,U8TO32_LITTLE(m + 36));
x10 = XOR(x10,U8TO32_LITTLE(m + 40));
x11 = XOR(x11,U8TO32_LITTLE(m + 44));
x12 = XOR(x12,U8TO32_LITTLE(m + 48));
x13 = XOR(x13,U8TO32_LITTLE(m + 52));
x14 = XOR(x14,U8TO32_LITTLE(m + 56));
x15 = XOR(x15,U8TO32_LITTLE(m + 60));
j8 = PLUSONE(j8);
if (!j8) {
j9 = PLUSONE(j9);
/* stopping at 2^70 bytes per nonce is user's responsibility */
}
U32TO8_LITTLE(c + 0,x0);
U32TO8_LITTLE(c + 4,x1);
U32TO8_LITTLE(c + 8,x2);
U32TO8_LITTLE(c + 12,x3);
U32TO8_LITTLE(c + 16,x4);
U32TO8_LITTLE(c + 20,x5);
U32TO8_LITTLE(c + 24,x6);
U32TO8_LITTLE(c + 28,x7);
U32TO8_LITTLE(c + 32,x8);
U32TO8_LITTLE(c + 36,x9);
U32TO8_LITTLE(c + 40,x10);
U32TO8_LITTLE(c + 44,x11);
U32TO8_LITTLE(c + 48,x12);
U32TO8_LITTLE(c + 52,x13);
U32TO8_LITTLE(c + 56,x14);
U32TO8_LITTLE(c + 60,x15);
if (bytes <= 64) {
if (bytes < 64) {
for (i = 0;i < bytes;++i) ctarget[i] = c[i];
}
x->input[8] = j8;
x->input[9] = j9;
return;
}
bytes -= 64;
c += 64;
m += 64;
}
}
void dfsch_salsa20_get_keystream_block(dfsch_salsa20_state_t* st,
uint8_t output[64]){
uint32_t x[16];
int i;
for (i = 0;i < 16;++i) x[i] = st->input[i];
for (i = 20;i > 0;i -= 2) {
x[ 4] = XOR(x[ 4],ROTATE(PLUS(x[ 0],x[12]), 7));
x[ 8] = XOR(x[ 8],ROTATE(PLUS(x[ 4],x[ 0]), 9));
x[12] = XOR(x[12],ROTATE(PLUS(x[ 8],x[ 4]),13));
x[ 0] = XOR(x[ 0],ROTATE(PLUS(x[12],x[ 8]),18));
x[ 9] = XOR(x[ 9],ROTATE(PLUS(x[ 5],x[ 1]), 7));
x[13] = XOR(x[13],ROTATE(PLUS(x[ 9],x[ 5]), 9));
x[ 1] = XOR(x[ 1],ROTATE(PLUS(x[13],x[ 9]),13));
x[ 5] = XOR(x[ 5],ROTATE(PLUS(x[ 1],x[13]),18));
x[14] = XOR(x[14],ROTATE(PLUS(x[10],x[ 6]), 7));
x[ 2] = XOR(x[ 2],ROTATE(PLUS(x[14],x[10]), 9));
x[ 6] = XOR(x[ 6],ROTATE(PLUS(x[ 2],x[14]),13));
x[10] = XOR(x[10],ROTATE(PLUS(x[ 6],x[ 2]),18));
x[ 3] = XOR(x[ 3],ROTATE(PLUS(x[15],x[11]), 7));
x[ 7] = XOR(x[ 7],ROTATE(PLUS(x[ 3],x[15]), 9));
x[11] = XOR(x[11],ROTATE(PLUS(x[ 7],x[ 3]),13));
x[15] = XOR(x[15],ROTATE(PLUS(x[11],x[ 7]),18));
x[ 1] = XOR(x[ 1],ROTATE(PLUS(x[ 0],x[ 3]), 7));
x[ 2] = XOR(x[ 2],ROTATE(PLUS(x[ 1],x[ 0]), 9));
x[ 3] = XOR(x[ 3],ROTATE(PLUS(x[ 2],x[ 1]),13));
x[ 0] = XOR(x[ 0],ROTATE(PLUS(x[ 3],x[ 2]),18));
x[ 6] = XOR(x[ 6],ROTATE(PLUS(x[ 5],x[ 4]), 7));
x[ 7] = XOR(x[ 7],ROTATE(PLUS(x[ 6],x[ 5]), 9));
x[ 4] = XOR(x[ 4],ROTATE(PLUS(x[ 7],x[ 6]),13));
x[ 5] = XOR(x[ 5],ROTATE(PLUS(x[ 4],x[ 7]),18));
x[11] = XOR(x[11],ROTATE(PLUS(x[10],x[ 9]), 7));
x[ 8] = XOR(x[ 8],ROTATE(PLUS(x[11],x[10]), 9));
x[ 9] = XOR(x[ 9],ROTATE(PLUS(x[ 8],x[11]),13));
x[10] = XOR(x[10],ROTATE(PLUS(x[ 9],x[ 8]),18));
x[12] = XOR(x[12],ROTATE(PLUS(x[15],x[14]), 7));
x[13] = XOR(x[13],ROTATE(PLUS(x[12],x[15]), 9));
x[14] = XOR(x[14],ROTATE(PLUS(x[13],x[12]),13));
x[15] = XOR(x[15],ROTATE(PLUS(x[14],x[13]),18));
}
for (i = 0;i < 16;++i) x[i] = PLUS(x[i],st->input[i]);
for (i = 0;i < 16;++i) U32TO8_LITTLE(output + 4 * i,x[i]);
st->input[8]++;
if (!st->input[8]){
st->input[9]++;
}
}
bool Player::PlayerUCT::create_children(Board & board, Node * node, int toplay){
if(!node->children.lock())
return false;
if(player->dists || player->detectdraw){
dists.run(&board, (player->dists > 0), (player->detectdraw ? 0 : toplay));
if(player->detectdraw){
// assert(node->outcome == -3);
node->outcome = dists.isdraw(); //could be winnable by only one side
if(node->outcome == 0){ //proven draw, neither side can influence the outcome
node->bestmove = *(board.moveit()); //just choose the first move since all are equal at this point
node->children.unlock();
return true;
}
}
}
CompactTree<Node>::Children temp;
temp.alloc(board.movesremain(), player->ctmem);
int losses = 0;
Node * child = temp.begin(),
* end = temp.end(),
* loss = NULL;
Board::MoveIterator move = board.moveit(player->prunesymmetry);
int nummoves = 0;
for(; !move.done() && child != end; ++move, ++child){
*child = Node(*move);
if(player->minimax){
child->outcome = board.test_win(*move);
if(player->minimax >= 2 && board.test_win(*move, 3 - board.toplay()) > 0){
losses++;
loss = child;
}
if(child->outcome == toplay){ //proven win from here, don't need children
node->outcome = child->outcome;
node->proofdepth = 1;
node->bestmove = *move;
node->children.unlock();
temp.dealloc(player->ctmem);
return true;
}
}
if(player->knowledge)
add_knowledge(board, node, child);
nummoves++;
}
if(player->prunesymmetry)
temp.shrink(nummoves); //shrink the node to ignore the extra moves
else //both end conditions should happen in parallel
assert(move.done() && child == end);
//Make a macro move, add experience to the move so the current simulation continues past this move
if(losses == 1){
Node macro = *loss;
temp.dealloc(player->ctmem);
temp.alloc(1, player->ctmem);
macro.exp.addwins(player->visitexpand);
*(temp.begin()) = macro;
}else if(losses >= 2){ //proven loss, but at least try to block one of them
node->outcome = 3 - toplay;
node->proofdepth = 2;
node->bestmove = loss->move;
node->children.unlock();
temp.dealloc(player->ctmem);
return true;
}
if(player->dynwiden > 0) //sort in decreasing order by knowledge
sort(temp.begin(), temp.end(), sort_node_know);
PLUS(player->nodes, temp.num());
node->children.swap(temp);
assert(temp.unlock());
return true;
}
static void
_salsa20_block(int rounds, uint32_t *input, uint8_t *output)
{
uint32_t x0, x1, x2, x3, x4, x5, x6, x7;
uint32_t x8, x9, x10, x11, x12, x13, x14, x15;
uint8_t i;
x0 = input[0];
x1 = input[1];
x2 = input[2];
x3 = input[3];
x4 = input[4];
x5 = input[5];
x6 = input[6];
x7 = input[7];
x8 = input[8];
x9 = input[9];
x10 = input[10];
x11 = input[11];
x12 = input[12];
x13 = input[13];
x14 = input[14];
x15 = input[15];
for (i = rounds; i > 0; i -= 2) {
/* Column round */
x4 = XOR ( x4, ROTATE (PLUS ( x0,x12), 7));
x8 = XOR ( x8, ROTATE (PLUS ( x4, x0), 9));
x12 = XOR (x12, ROTATE (PLUS ( x8, x4),13));
x0 = XOR ( x0, ROTATE (PLUS (x12, x8),18));
x9 = XOR ( x9, ROTATE (PLUS ( x5, x1), 7));
x13 = XOR (x13, ROTATE (PLUS ( x9, x5), 9));
x1 = XOR ( x1, ROTATE (PLUS (x13, x9),13));
x5 = XOR ( x5, ROTATE (PLUS ( x1,x13),18));
x14 = XOR (x14, ROTATE (PLUS (x10, x6), 7));
x2 = XOR ( x2, ROTATE (PLUS (x14,x10), 9));
x6 = XOR ( x6, ROTATE (PLUS ( x2,x14),13));
x10 = XOR (x10, ROTATE (PLUS ( x6, x2),18));
x3 = XOR ( x3, ROTATE (PLUS (x15,x11), 7));
x7 = XOR ( x7, ROTATE (PLUS ( x3,x15), 9));
x11 = XOR (x11, ROTATE (PLUS ( x7, x3),13));
x15 = XOR (x15, ROTATE (PLUS (x11, x7),18));
/* Row round */
x1 = XOR ( x1, ROTATE (PLUS ( x0, x3), 7));
x2 = XOR ( x2, ROTATE (PLUS ( x1, x0), 9));
x3 = XOR ( x3, ROTATE (PLUS ( x2, x1),13));
x0 = XOR ( x0, ROTATE (PLUS ( x3, x2),18));
x6 = XOR ( x6, ROTATE (PLUS ( x5, x4), 7));
x7 = XOR ( x7, ROTATE (PLUS ( x6, x5), 9));
x4 = XOR ( x4, ROTATE (PLUS ( x7, x6),13));
x5 = XOR ( x5, ROTATE (PLUS ( x4, x7),18));
x11 = XOR (x11, ROTATE (PLUS (x10, x9), 7));
x8 = XOR ( x8, ROTATE (PLUS (x11,x10), 9));
x9 = XOR ( x9, ROTATE (PLUS ( x8,x11),13));
x10 = XOR (x10, ROTATE (PLUS ( x9, x8),18));
x12 = XOR (x12, ROTATE (PLUS (x15,x14), 7));
x13 = XOR (x13, ROTATE (PLUS (x12,x15), 9));
x14 = XOR (x14, ROTATE (PLUS (x13,x12),13));
x15 = XOR (x15, ROTATE (PLUS (x14,x13),18));
}
x0 = PLUS (x0, input[0]);
x1 = PLUS (x1, input[1]);
x2 = PLUS (x2, input[2]);
x3 = PLUS (x3, input[3]);
x4 = PLUS (x4, input[4]);
x5 = PLUS (x5, input[5]);
x6 = PLUS (x6, input[6]);
x7 = PLUS (x7, input[7]);
x8 = PLUS (x8, input[8]);
x9 = PLUS (x9, input[9]);
x10 = PLUS (x10, input[10]);
x11 = PLUS (x11, input[11]);
x12 = PLUS (x12, input[12]);
x13 = PLUS (x13, input[13]);
x14 = PLUS (x14, input[14]);
x15 = PLUS (x15, input[15]);
U32TO8_LITTLE (output + 0, x0);
U32TO8_LITTLE (output + 4, x1);
U32TO8_LITTLE (output + 8, x2);
U32TO8_LITTLE (output + 12, x3);
U32TO8_LITTLE (output + 16, x4);
U32TO8_LITTLE (output + 20, x5);
U32TO8_LITTLE (output + 24, x6);
U32TO8_LITTLE (output + 28, x7);
U32TO8_LITTLE (output + 32, x8);
U32TO8_LITTLE (output + 36, x9);
U32TO8_LITTLE (output + 40, x10);
U32TO8_LITTLE (output + 44, x11);
U32TO8_LITTLE (output + 48, x12);
U32TO8_LITTLE (output + 52, x13);
U32TO8_LITTLE (output + 56, x14);
U32TO8_LITTLE (output + 60, x15);
/* Increment block counter */
input[8] = PLUSONE (input[8]);
if (!input[8]) {
input[9] = PLUSONE (input[9]);
/* stopping at 2^70 bytes per nonce is user's responsibility */
}
}
