int main()
{
char dest[5] = {0};
char src[5] = {0};
random_answer(dest,4);
BOOL res = FALSE;
do{
printf("please input your answer :");
//int i = 0;
//for(i = 0;i<4;i++)
//{
// src[i] = getchar();
//}
//getchar();
int real_count = get_src(src,4);
printf("src : %s\n",src);
if(real_count < 4)
{
printf("too short \n");
continue;
}
res = get_option(dest,src,4);
}while(!res);
printf("YOU WIN !!!!\n");
return 0;
}
static void
nv04_surface_copy_cpu(struct gl_context *ctx,
struct nouveau_surface *dst,
struct nouveau_surface *src,
int dx, int dy, int sx, int sy,
int w, int h)
{
int x, y;
get_offset_t get_dst = (dst->layout == SWIZZLED ?
get_swizzled_offset : get_linear_offset);
get_offset_t get_src = (src->layout == SWIZZLED ?
get_swizzled_offset : get_linear_offset);
void *dp, *sp;
nouveau_bo_map(dst->bo, NOUVEAU_BO_WR, context_client(ctx));
nouveau_bo_map(src->bo, NOUVEAU_BO_RD, context_client(ctx));
dp = dst->bo->map + dst->offset;
sp = src->bo->map + src->offset;
for (y = 0; y < h; y++) {
for (x = 0; x < w; x++) {
memcpy(dp + get_dst(dst, dx + x, dy + y),
sp + get_src(src, sx + x, sy + y), dst->cpp);
}
}
}
ErrorCode HTMLIFrameElement::Load()
{
ASSERT(0);
return 0;
#if 0
ASSERT(m_pFrameContent);
sysstring src = get_src();
if (src.Length())
{
TCHAR result[2048];
{
sysstring documentUrl = m_ownerDocument->get_url();
DWORD resultLen = sizeof(result);
InternetCombineUrl(W2A(documentUrl), W2A(src), result, &resultLen, 0);
}
CComBSTR url = A2W(result);
return m_pFrameContent->LoadSRC(CComQIPtr<IHlinkSite>(m_ownerDocument), url);
}
else
return S_FALSE;
#endif
}
void edge::dump (void)
{
util::scoped_lock (mtx_);
std::cout << " -------- edge "
<< get_name ()
<< " [" << get_src ().get_string () << "\t -> " << get_tgt ().get_string () << "] "
<< " (" << state_to_string (state_) << ")"
<< std::endl;
}
int add_function(char *path, char *hpath)
{
char *raw;
if ((raw = get_src(path)) == NULL)
{
free(path);
path = NULL;
return (-1);
}
print_function(raw);
fprintf(stderr, "Writed %s%s%s'functions into %s%s%s\n", GREEN, path, WHITE, GREEN, hpath, WHITE);
free(raw);
free(path);
path = NULL;
return (0);
}
bool AdminNetLinkAdd::commit(AdminState* admin) throw (admin_error)
{
AdminNetLinkBase::commit(admin);
AdminState::WriteLockRecord net_rec = get_net_write();
NetStatePtr net = boost::dynamic_pointer_cast<NetState>(net_rec.first);
NetVertex v_src;
if (!net->get_node(get_src(), v_src, *net_rec.second))
return false;
NetVertex v_dst;
if (!net->get_node(get_dst(), v_dst, *net_rec.second))
return false;
NetEdge e;
if (!net->add_edge(v_src, v_dst, e, *net_rec.second))
return false;
net->set_name(e, get_link(), *net_rec.second);
return true;
}
CImage* HTMLImageElement::GetImage()
{
if (m_pImage == NULL)
{
if (false)
{
String src = get_src();
if (src.GetLength())
{
WCHAR result[2048];
{
String documentUrl = m_ownerDocument->get_url();
// documentUrl = _bstr_t(bdocumentUrl, false);
String baseUrl = documentUrl;
ASSERT(0);
#if 0
DWORD resultLen = sizeof(result) - sizeof(result[0]);
InternetCombineUrlW(baseUrl->c_str(), src->c_str(), result, &resultLen, 0);
#endif
}
// StringA localfilename = DownloadFile(result);
m_pImage = new CImage;//LXML::g_Images.OpenImage(Convert2W(result));
m_pImage->m_url = string_copy(result);
DWORD threadId;
::CreateThread(NULL, 0, ThreadFunc, this, 0, &threadId);
}
}
}
return m_pImage;
}
Opnd
SymDispOpnd::get_disp() const
{
return get_src(1);
}
Opnd
IndexSymDispOpnd::get_addr_sym() const
{
return get_src(1);
}
o = 0;
if (!m_backupdir || !tino_file_notexistsE(rename=tmp=o=tino_file_glue_pathOi(NULL, 0, m_backupdir, tino_file_filenameptr_constO(rename))))
tmp = tino_file_backupnameNi(NULL, 0, rename);
do_rename(name, tmp);
tino_freeO(tmp);
if (o!=tmp)
tino_freeO(o);
}
static void
do_rename_backup(const char *old, const char *new)
{
const char *src;
src = get_src(old);
#if 0
XXX DOES NOT WORK XXX;
/* Try to move, skips 2 syscalls in the most common situation
*/
if (!rename(src, new))
{
verbose("rename: %s -> %s", src, new);
return;
}
#endif
/* Try to figure out what happened */
if (tino_file_notexistsE(src))
Opnd
BaseIndexScaleDispOpnd::get_disp() const
{
return get_src(3);
}
Opnd
BaseIndexScaleDispOpnd::get_scale() const
{
return get_src(2);
}
Opnd
BaseDispOpnd::get_disp() const
{
return get_src(1);
}
Opnd
IndexScaleDispOpnd::get_scale() const
{
return get_src(1);
}
Opnd
IndexScaleDispOpnd::get_index() const
{
return get_src(0);
}
Opnd
BaseIndexDispOpnd::get_disp() const
{
return get_src(2);
}
Opnd
BaseIndexOpnd::get_index() const
{
return get_src(1);
}
Opnd
BaseIndexOpnd::get_base() const
{
return get_src(0);
}
Opnd
SymDispOpnd::get_addr_sym() const
{
return get_src(0);
}
int main()
{
#ifndef __EMSCRIPTEN__
_MM_SET_FLUSH_ZERO_MODE(_MM_FLUSH_ZERO_ON);
#endif
printf ("{ \"workload\": %u, \"results\": [\n", N);
assert(N%4 == 0); // Don't care about the tail for now.
float *src = get_src();//(float*)aligned_alloc(16, N*sizeof(float));
for(int i = 0; i < N; ++i)
src[i] = (float)rand() / RAND_MAX;
float *src2 = get_src2();//(float*)aligned_alloc(16, N*sizeof(float));
for(int i = 0; i < N; ++i)
src2[i] = (float)rand() / RAND_MAX;
float *dst = get_dst();//(float*)aligned_alloc(16, N*sizeof(float));
float scalarTime;
SETCHART("load");
START();
for(int i = 0; i < N; ++i)
dst[i] = src[i];
ENDSCALAR(checksum_dst(dst), "scalar");
LS_TEST("_mm_load_ps", _mm_load_ps, 0, _mm_store_ps, 0);
LS_TEST("_mm_load_ps1", _mm_load_ps1, 1, _mm_store_ps, 0);
LS_TEST("_mm_load_ss", _mm_load_ss, 1, _mm_store_ps, 0);
LS_TEST("_mm_load1_ps", _mm_load1_ps, 1, _mm_store_ps, 0);
// _mm_loadh_pi
// _mm_loadl_pi
LS_TEST("_mm_loadr_ps", _mm_loadr_ps, 0, _mm_store_ps, 0);
LS_TEST("_mm_loadu_ps", _mm_loadu_ps, 1, _mm_store_ps, 0);
SETCHART("set");
SS_TEST("_mm_set_ps", _mm_set_ps(src[i+2], src[i+1], src[i+5], src[i+0]));
SS_TEST("_mm_set_ps1", _mm_set_ps1(src[i]));
SS_TEST("_mm_set_ss", _mm_set_ss(src[i]));
SS_TEST("_mm_set1_ps", _mm_set1_ps(src[i]));
SS_TEST("_mm_setr_ps", _mm_set_ps(src[i+2], src[i+1], src[i+5], src[i+0]));
SS_TEST("_mm_setzero_ps", _mm_setzero_ps());
SETCHART("move");
SS_TEST("_mm_move_ss", _mm_move_ss(_mm_load_ps(src+i), _mm_load_ps(src2+i)));
SS_TEST("_mm_movehl_ps", _mm_movehl_ps(_mm_load_ps(src+i), _mm_load_ps(src2+i)));
SS_TEST("_mm_movelh_ps", _mm_movelh_ps(_mm_load_ps(src+i), _mm_load_ps(src2+i)));
SETCHART("store");
LS_TEST("_mm_store_ps", _mm_load_ps, 0, _mm_store_ps, 0);
LS_TEST("_mm_store_ps1", _mm_load_ps, 0, _mm_store_ps1, 0);
LS_TEST("_mm_store_ss", _mm_load_ps, 0, _mm_store_ss, 1);
LS64_TEST("_mm_storeh_pi", _mm_load_ps, 0, _mm_storeh_pi, 1);
LS64_TEST("_mm_storel_pi", _mm_load_ps, 0, _mm_storel_pi, 1);
LS_TEST("_mm_storer_ps", _mm_load_ps, 0, _mm_storer_ps, 0);
LS_TEST("_mm_storeu_ps", _mm_load_ps, 0, _mm_storeu_ps, 1);
LS_TEST("_mm_stream_ps", _mm_load_ps, 0, _mm_stream_ps, 0);
SETCHART("arithmetic");
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] += src2[0]; dst[1] += src2[1]; dst[2] += src2[2]; dst[3] += src2[3]; } ENDSCALAR(checksum_dst(dst), "scalar add");
BINARYOP_TEST("_mm_add_ps", _mm_add_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_add_ss", _mm_add_ss, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] /= src2[0]; dst[1] /= src2[1]; dst[2] /= src2[2]; dst[3] /= src2[3]; } ENDSCALAR(checksum_dst(dst), "scalar div");
BINARYOP_TEST("_mm_div_ps", _mm_div_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_div_ss", _mm_div_ss, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] *= src2[0]; dst[1] *= src2[1]; dst[2] *= src2[2]; dst[3] *= src2[3]; } ENDSCALAR(checksum_dst(dst), "scalar mul");
BINARYOP_TEST("_mm_mul_ps", _mm_mul_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_mul_ss", _mm_mul_ss, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] -= src2[0]; dst[1] -= src2[1]; dst[2] -= src2[2]; dst[3] -= src2[3]; } ENDSCALAR(checksum_dst(dst), "scalar sub");
BINARYOP_TEST("_mm_sub_ps", _mm_sub_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_sub_ss", _mm_sub_ss, _mm_load_ps(src), _mm_load_ps(src2));
SETCHART("roots");
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = 1.f / dst[0]; dst[1] = 1.f / dst[1]; dst[2] = 1.f / dst[2]; dst[3] = 1.f / dst[3]; } ENDSCALAR(checksum_dst(dst), "scalar rcp");
UNARYOP_TEST("_mm_rcp_ps", _mm_rcp_ps, _mm_load_ps(src));
UNARYOP_TEST("_mm_rcp_ss", _mm_rcp_ss, _mm_load_ps(src));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = 1.f / sqrtf(dst[0]); dst[1] = 1.f / sqrtf(dst[1]); dst[2] = 1.f / sqrtf(dst[2]); dst[3] = 1.f / sqrtf(dst[3]); } ENDSCALAR(checksum_dst(dst), "scalar rsqrt");
UNARYOP_TEST("_mm_rsqrt_ps", _mm_rsqrt_ps, _mm_load_ps(src));
UNARYOP_TEST("_mm_rsqrt_ss", _mm_rsqrt_ss, _mm_load_ps(src));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = sqrtf(dst[0]); dst[1] = sqrtf(dst[1]); dst[2] = sqrtf(dst[2]); dst[3] = sqrtf(dst[3]); } ENDSCALAR(checksum_dst(dst), "scalar sqrt");
UNARYOP_TEST("_mm_sqrt_ps", _mm_sqrt_ps, _mm_load_ps(src));
UNARYOP_TEST("_mm_sqrt_ss", _mm_sqrt_ss, _mm_load_ps(src));
SETCHART("logical");
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = ucastf(fcastu(dst[0]) & fcastu(src2[0])); dst[1] = ucastf(fcastu(dst[1]) & fcastu(src2[1])); dst[2] = ucastf(fcastu(dst[2]) & fcastu(src2[2])); dst[3] = ucastf(fcastu(dst[3]) & fcastu(src2[3])); } ENDSCALAR(checksum_dst(dst), "scalar and");
BINARYOP_TEST("_mm_and_ps", _mm_and_ps, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = ucastf((~fcastu(dst[0])) & fcastu(src2[0])); dst[1] = ucastf((~fcastu(dst[1])) & fcastu(src2[1])); dst[2] = ucastf((~fcastu(dst[2])) & fcastu(src2[2])); dst[3] = ucastf((~fcastu(dst[3])) & fcastu(src2[3])); } ENDSCALAR(checksum_dst(dst), "scalar andnot");
BINARYOP_TEST("_mm_andnot_ps", _mm_andnot_ps, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = ucastf(fcastu(dst[0]) | fcastu(src2[0])); dst[1] = ucastf(fcastu(dst[1]) | fcastu(src2[1])); dst[2] = ucastf(fcastu(dst[2]) | fcastu(src2[2])); dst[3] = ucastf(fcastu(dst[3]) | fcastu(src2[3])); } ENDSCALAR(checksum_dst(dst), "scalar or");
BINARYOP_TEST("_mm_or_ps", _mm_or_ps, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = ucastf(fcastu(dst[0]) ^ fcastu(src2[0])); dst[1] = ucastf(fcastu(dst[1]) ^ fcastu(src2[1])); dst[2] = ucastf(fcastu(dst[2]) ^ fcastu(src2[2])); dst[3] = ucastf(fcastu(dst[3]) ^ fcastu(src2[3])); } ENDSCALAR(checksum_dst(dst), "scalar xor");
BINARYOP_TEST("_mm_xor_ps", _mm_xor_ps, _mm_load_ps(src), _mm_load_ps(src2));
SETCHART("cmp");
#ifndef __EMSCRIPTEN__ // TODO: Disabled due to https://github.com/kripken/emscripten/issues/2841
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = (dst[0] == src2[0]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[1] = (dst[1] == src2[1]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[2] = (dst[2] == src2[2]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[3] = (dst[3] == src2[3]) ? ucastf(0xFFFFFFFFU) : 0.f; } ENDSCALAR(checksum_dst(dst), "scalar cmp==");
BINARYOP_TEST("_mm_cmpeq_ps", _mm_cmpeq_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_cmpeq_ss", _mm_cmpeq_ss, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = (dst[0] >= src2[0]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[1] = (dst[1] >= src2[1]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[2] = (dst[2] >= src2[2]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[3] = (dst[3] >= src2[3]) ? ucastf(0xFFFFFFFFU) : 0.f; } ENDSCALAR(checksum_dst(dst), "scalar cmp>=");
BINARYOP_TEST("_mm_cmpge_ps", _mm_cmpge_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_cmpge_ss", _mm_cmpge_ss, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = (dst[0] > src2[0]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[1] = (dst[1] > src2[1]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[2] = (dst[2] > src2[2]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[3] = (dst[3] > src2[3]) ? ucastf(0xFFFFFFFFU) : 0.f; } ENDSCALAR(checksum_dst(dst), "scalar cmp>");
BINARYOP_TEST("_mm_cmpgt_ps", _mm_cmpgt_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_cmpgt_ss", _mm_cmpgt_ss, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = (dst[0] <= src2[0]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[1] = (dst[1] <= src2[1]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[2] = (dst[2] <= src2[2]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[3] = (dst[3] <= src2[3]) ? ucastf(0xFFFFFFFFU) : 0.f; } ENDSCALAR(checksum_dst(dst), "scalar cmp<=");
BINARYOP_TEST("_mm_cmple_ps", _mm_cmple_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_cmple_ss", _mm_cmple_ss, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = (dst[0] < src2[0]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[1] = (dst[1] < src2[1]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[2] = (dst[2] < src2[2]) ? ucastf(0xFFFFFFFFU) : 0.f; dst[3] = (dst[3] < src2[3]) ? ucastf(0xFFFFFFFFU) : 0.f; } ENDSCALAR(checksum_dst(dst), "scalar cmp<");
BINARYOP_TEST("_mm_cmplt_ps", _mm_cmplt_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_cmplt_ss", _mm_cmplt_ss, _mm_load_ps(src), _mm_load_ps(src2));
#endif
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = (!Isnan(dst[0]) && !Isnan(src2[0])) ? ucastf(0xFFFFFFFFU) : 0.f; dst[1] = (!Isnan(dst[1]) && !Isnan(src2[1])) ? ucastf(0xFFFFFFFFU) : 0.f; dst[2] = (!Isnan(dst[2]) && !Isnan(src2[2])) ? ucastf(0xFFFFFFFFU) : 0.f; dst[3] = (!Isnan(dst[3]) && !Isnan(src2[3])) ? ucastf(0xFFFFFFFFU) : 0.f; } ENDSCALAR(checksum_dst(dst), "scalar cmpord");
BINARYOP_TEST("_mm_cmpord_ps", _mm_cmpord_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_cmpord_ss", _mm_cmpord_ss, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = (Isnan(dst[0]) || Isnan(src2[0])) ? ucastf(0xFFFFFFFFU) : 0.f; dst[1] = (Isnan(dst[1]) || Isnan(src2[1])) ? ucastf(0xFFFFFFFFU) : 0.f; dst[2] = (Isnan(dst[2]) || Isnan(src2[2])) ? ucastf(0xFFFFFFFFU) : 0.f; dst[3] = (Isnan(dst[3]) || Isnan(src2[3])) ? ucastf(0xFFFFFFFFU) : 0.f; } ENDSCALAR(checksum_dst(dst), "scalar cmpunord");
BINARYOP_TEST("_mm_cmpunord_ps", _mm_cmpunord_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_cmpunord_ss", _mm_cmpunord_ss, _mm_load_ps(src), _mm_load_ps(src2));
SETCHART("max");
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = Max(dst[0], src2[0]); dst[1] = Max(dst[1], src2[1]); dst[2] = Max(dst[2], src2[2]); dst[3] = Max(dst[3], src2[3]); } ENDSCALAR(checksum_dst(dst), "scalar max");
BINARYOP_TEST("_mm_max_ps", _mm_max_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_max_ss", _mm_max_ss, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = Min(dst[0], src2[0]); dst[1] = Min(dst[1], src2[1]); dst[2] = Min(dst[2], src2[2]); dst[3] = Min(dst[3], src2[3]); } ENDSCALAR(checksum_dst(dst), "scalar min");
BINARYOP_TEST("_mm_min_ps", _mm_min_ps, _mm_load_ps(src), _mm_load_ps(src2));
BINARYOP_TEST("_mm_min_ss", _mm_min_ss, _mm_load_ps(src), _mm_load_ps(src2));
SETCHART("shuffle");
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[3] = dst[1]; dst[2] = dst[0]; dst[1] = src2[3]; dst[0] = src2[2]; } ENDSCALAR(checksum_dst(dst), "scalar shuffle");
// BINARYOP_TEST("_mm_shuffle_ps", _mm_shuffle_ps, _mm_load_ps(src), _mm_load_ps(src2));
START();
__m128 o0 = _mm_load_ps(src);
__m128 o1 = _mm_load_ps(src2);
for(int i = 0; i < N; i += 4)
o0 = _mm_shuffle_ps(o0, o1, _MM_SHUFFLE(1, 0, 3, 2));
_mm_store_ps(dst, o0);
END(checksum_dst(dst), "_mm_shuffle_ps");
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[0] = dst[2]; dst[1] = src2[2]; dst[2] = dst[3]; dst[3] = src2[3]; } ENDSCALAR(checksum_dst(dst), "scalar unpackhi_ps");
BINARYOP_TEST("_mm_unpackhi_ps", _mm_unpackhi_ps, _mm_load_ps(src), _mm_load_ps(src2));
START(); dst[0] = src[0]; dst[1] = src[1]; dst[2] = src[2]; dst[3] = src[3]; for(int i = 0; i < N; ++i) { dst[2] = dst[1]; dst[1] = dst[0]; dst[0] = src2[0]; dst[3] = src2[1]; } ENDSCALAR(checksum_dst(dst), "scalar unpacklo_ps");
BINARYOP_TEST("_mm_unpacklo_ps", _mm_unpacklo_ps, _mm_load_ps(src), _mm_load_ps(src2));
printf("]}\n");
/*
printf("Finished!\n");
printf("Total time spent in scalar intrinsics: %f msecs.\n", (double)scalarTotalTicks * 1000.0 / ticks_per_sec());
printf("Total time spent in SSE1 intrinsics: %f msecs.\n", (double)simdTotalTicks * 1000.0 / ticks_per_sec());
if (scalarTotalTicks > simdTotalTicks)
printf("SSE1 was %.3fx faster than scalar!\n", (double)scalarTotalTicks / simdTotalTicks);
else
printf("SSE1 was %.3fx slower than scalar!\n", (double)simdTotalTicks / scalarTotalTicks);
*/
#ifdef __EMSCRIPTEN__
fprintf(stderr,"User Agent: %s\n", emscripten_run_script_string("navigator.userAgent"));
printf("/*Test finished! Now please close Firefox to continue with benchmark_sse1.py.*/\n");
#endif
exit(0);
}
Opnd
BaseIndexScaleDispOpnd::get_index() const
{
return get_src(1);
}
Opnd
IndexScaleDispOpnd::get_disp() const
{
return get_src(2);
}
Opnd
BaseDispOpnd::get_base() const
{
return get_src(0);
}
static void
dump_cons_stack( int gtid, struct cons_header * p ) {
int i;
int tos = p->stack_top;
kmp_str_buf_t buffer;
__kmp_str_buf_init( & buffer );
__kmp_str_buf_print( & buffer, "+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-\n" );
__kmp_str_buf_print( & buffer, "Begin construct stack with %d items for thread %d\n", tos, gtid );
__kmp_str_buf_print( & buffer, " stack_top=%d { P=%d, W=%d, S=%d }\n", tos, p->p_top, p->w_top, p->s_top );
for ( i = tos; i > 0; i-- ) {
struct cons_data * c = & ( p->stack_data[ i ] );
__kmp_str_buf_print( & buffer, " stack_data[%2d] = { %s (%s) %d %p }\n", i, cons_text_c[ c->type ], get_src( c->ident ), c->prev, c->name );
}; // for i
__kmp_str_buf_print( & buffer, "End construct stack for thread %d\n", gtid );
__kmp_str_buf_print( & buffer, "+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-\n" );
__kmp_debug_printf( "%s", buffer.str );
__kmp_str_buf_free( & buffer );
}
Opnd
BaseIndexScaleDispOpnd::get_base() const
{
return get_src(0);
}