代码拉取完成,页面将自动刷新
#include "stdafx.h"
#include "defs.h"
static void promote_tensor(void);
static int compatible(U *, U *);
//-----------------------------------------------------------------------------
//
// Called from the "eval" module to evaluate tensor elements.
//
// p1 points to the tensor operand.
//
//-----------------------------------------------------------------------------
void
eval_tensor(void)
{
int i, ndim, nelem;
U **a, **b;
//---------------------------------------------------------------------
//
// create a new tensor for the result
//
//---------------------------------------------------------------------
nelem = p1->u.tensor->nelem;
ndim = p1->u.tensor->ndim;
p2 = alloc_tensor(nelem);
p2->u.tensor->ndim = ndim;
for (i = 0; i < ndim; i++)
p2->u.tensor->dim[i] = p1->u.tensor->dim[i];
//---------------------------------------------------------------------
//
// b = eval(a)
//
//---------------------------------------------------------------------
a = p1->u.tensor->elem;
b = p2->u.tensor->elem;
for (i = 0; i < nelem; i++) {
push(a[i]);
eval();
b[i] = pop();
}
//---------------------------------------------------------------------
//
// push the result
//
//---------------------------------------------------------------------
push(p2);
promote_tensor();
}
//-----------------------------------------------------------------------------
//
// Add tensors
//
// Input: Operands on stack
//
// Output: Result on stack
//
//-----------------------------------------------------------------------------
void
tensor_plus_tensor(void)
{
int i, ndim, nelem;
U **a, **b, **c;
save();
p2 = pop();
p1 = pop();
// are the dimension lists equal?
ndim = p1->u.tensor->ndim;
if (ndim != p2->u.tensor->ndim) {
push(symbol(NIL));
restore();
return;
}
for (i = 0; i < ndim; i++)
if (p1->u.tensor->dim[i] != p2->u.tensor->dim[i]) {
push(symbol(NIL));
restore();
return;
}
// create a new tensor for the result
nelem = p1->u.tensor->nelem;
p3 = alloc_tensor(nelem);
p3->u.tensor->ndim = ndim;
for (i = 0; i < ndim; i++)
p3->u.tensor->dim[i] = p1->u.tensor->dim[i];
// c = a + b
a = p1->u.tensor->elem;
b = p2->u.tensor->elem;
c = p3->u.tensor->elem;
for (i = 0; i < nelem; i++) {
push(a[i]);
push(b[i]);
add();
c[i] = pop();
}
// push the result
push(p3);
restore();
}
//-----------------------------------------------------------------------------
//
// careful not to reorder factors
//
//-----------------------------------------------------------------------------
void
tensor_times_scalar(void)
{
int i, ndim, nelem;
U **a, **b;
save();
p2 = pop();
p1 = pop();
ndim = p1->u.tensor->ndim;
nelem = p1->u.tensor->nelem;
p3 = alloc_tensor(nelem);
p3->u.tensor->ndim = ndim;
for (i = 0; i < ndim; i++)
p3->u.tensor->dim[i] = p1->u.tensor->dim[i];
a = p1->u.tensor->elem;
b = p3->u.tensor->elem;
for (i = 0; i < nelem; i++) {
push(a[i]);
push(p2);
multiply();
b[i] = pop();
}
push(p3);
restore();
}
void
scalar_times_tensor(void)
{
int i, ndim, nelem;
U **a, **b;
save();
p2 = pop();
p1 = pop();
ndim = p2->u.tensor->ndim;
nelem = p2->u.tensor->nelem;
p3 = alloc_tensor(nelem);
p3->u.tensor->ndim = ndim;
for (i = 0; i < ndim; i++)
p3->u.tensor->dim[i] = p2->u.tensor->dim[i];
a = p2->u.tensor->elem;
b = p3->u.tensor->elem;
for (i = 0; i < nelem; i++) {
push(p1);
push(a[i]);
multiply();
b[i] = pop();
}
push(p3);
restore();
}
int
is_square_matrix(U *p)
{
if (istensor(p) && p->u.tensor->ndim == 2 && p->u.tensor->dim[0] == p->u.tensor->dim[1])
return 1;
else
return 0;
}
//-----------------------------------------------------------------------------
//
// gradient of tensor
//
//-----------------------------------------------------------------------------
void
d_tensor_tensor(void)
{
int i, j, ndim, nelem;
U **a, **b, **c;
ndim = p1->u.tensor->ndim;
nelem = p1->u.tensor->nelem;
if (ndim + 1 >= MAXDIM)
goto dont_evaluate;
p3 = alloc_tensor(nelem * p2->u.tensor->nelem);
p3->u.tensor->ndim = ndim + 1;
for (i = 0; i < ndim; i++)
p3->u.tensor->dim[i] = p1->u.tensor->dim[i];
p3->u.tensor->dim[ndim] = p2->u.tensor->dim[0];
a = p1->u.tensor->elem;
b = p2->u.tensor->elem;
c = p3->u.tensor->elem;
for (i = 0; i < nelem; i++) {
for (j = 0; j < p2->u.tensor->nelem; j++) {
push(a[i]);
push(b[j]);
derivative();
c[i * p2->u.tensor->nelem + j] = pop();
}
}
push(p3);
return;
dont_evaluate:
push_symbol(DERIVATIVE);
push(p1);
push(p2);
list(3);
}
//-----------------------------------------------------------------------------
//
// gradient of scalar
//
//-----------------------------------------------------------------------------
void
d_scalar_tensor(void)
{
int i;
U **a, **b;
p3 = alloc_tensor(p2->u.tensor->nelem);
p3->u.tensor->ndim = 1;
p3->u.tensor->dim[0] = p2->u.tensor->dim[0];
a = p2->u.tensor->elem;
b = p3->u.tensor->elem;
for (i = 0; i < p2->u.tensor->nelem; i++) {
push(p1);
push(a[i]);
derivative();
b[i] = pop();
}
push(p3);
}
//-----------------------------------------------------------------------------
//
// Derivative of tensor
//
//-----------------------------------------------------------------------------
void
d_tensor_scalar(void)
{
int i;
U **a, **b;
p3 = alloc_tensor(p1->u.tensor->nelem);
p3->u.tensor->ndim = p1->u.tensor->ndim;
for (i = 0; i < p1->u.tensor->ndim; i++)
p3->u.tensor->dim[i] = p1->u.tensor->dim[i];
a = p1->u.tensor->elem;
b = p3->u.tensor->elem;
for (i = 0; i < p1->u.tensor->nelem; i++) {
push(a[i]);
push(p2);
derivative();
b[i] = pop();
}
push(p3);
}
int
compare_tensors(U *p1, U *p2)
{
int i;
if (p1->u.tensor->ndim < p2->u.tensor->ndim)
return -1;
if (p1->u.tensor->ndim > p2->u.tensor->ndim)
return 1;
for (i = 0; i < p1->u.tensor->ndim; i++) {
if (p1->u.tensor->dim[i] < p2->u.tensor->dim[i])
return -1;
if (p1->u.tensor->dim[i] > p2->u.tensor->dim[i])
return 1;
}
for (i = 0; i < p1->u.tensor->nelem; i++) {
if (equal(p1->u.tensor->elem[i], p2->u.tensor->elem[i]))
continue;
if (lessp(p1->u.tensor->elem[i], p2->u.tensor->elem[i]))
return -1;
else
return 1;
}
return 0;
}
//-----------------------------------------------------------------------------
//
// Raise a tensor to a power
//
// Input: p1 tensor
//
// p2 exponent
//
// Output: Result on stack
//
//-----------------------------------------------------------------------------
void
power_tensor(void)
{
int i, k, n;
// first and last dims must be equal
k = p1->u.tensor->ndim - 1;
if (p1->u.tensor->dim[0] != p1->u.tensor->dim[k]) {
push_symbol(POWER);
push(p1);
push(p2);
list(3);
return;
}
push(p2);
n = pop_integer();
if (n == (int) 0x80000000) {
push_symbol(POWER);
push(p1);
push(p2);
list(3);
return;
}
if (n == 0) {
if (p1->u.tensor->ndim != 2)
stop("power(tensor,0) with tensor rank not equal to 2");
n = p1->u.tensor->dim[0];
p1 = alloc_tensor(n * n);
p1->u.tensor->ndim = 2;
p1->u.tensor->dim[0] = n;
p1->u.tensor->dim[1] = n;
for (i = 0; i < n; i++)
p1->u.tensor->elem[n * i + i] = one;
push(p1);
return;
}
if (n < 0) {
n = -n;
push(p1);
inv();
p1 = pop();
}
push(p1);
for (i = 1; i < n; i++) {
push(p1);
inner();
if (iszero(stack[tos - 1]))
break;
}
}
void
copy_tensor(void)
{
int i;
save();
p1 = pop();
p2 = alloc_tensor(p1->u.tensor->nelem);
p2->u.tensor->ndim = p1->u.tensor->ndim;
for (i = 0; i < p1->u.tensor->ndim; i++)
p2->u.tensor->dim[i] = p1->u.tensor->dim[i];
for (i = 0; i < p1->u.tensor->nelem; i++)
p2->u.tensor->elem[i] = p1->u.tensor->elem[i];
push(p2);
restore();
}
// Tensors with elements that are also tensors get promoted to a higher rank.
static void
promote_tensor(void)
{
int i, j, k, nelem, ndim;
save();
p1 = pop();
if (!istensor(p1)) {
push(p1);
restore();
return;
}
p2 = p1->u.tensor->elem[0];
for (i = 1; i < p1->u.tensor->nelem; i++)
if (!compatible(p2, p1->u.tensor->elem[i]))
stop("Cannot promote tensor due to inconsistent tensor components.");
if (!istensor(p2)) {
push(p1);
restore();
return;
}
ndim = p1->u.tensor->ndim + p2->u.tensor->ndim;
if (ndim > MAXDIM)
stop("tensor rank > 24");
nelem = p1->u.tensor->nelem * p2->u.tensor->nelem;
p3 = alloc_tensor(nelem);
p3->u.tensor->ndim = ndim;
for (i = 0; i < p1->u.tensor->ndim; i++)
p3->u.tensor->dim[i] = p1->u.tensor->dim[i];
for (j = 0; j < p2->u.tensor->ndim; j++)
p3->u.tensor->dim[i + j] = p2->u.tensor->dim[j];
k = 0;
for (i = 0; i < p1->u.tensor->nelem; i++) {
p2 = p1->u.tensor->elem[i];
for (j = 0; j < p2->u.tensor->nelem; j++)
p3->u.tensor->elem[k++] = p2->u.tensor->elem[j];
}
push(p3);
restore();
}
static int
compatible(U *p, U *q)
{
int i;
if (!istensor(p) && !istensor(q))
return 1;
if (!istensor(p) || !istensor(q))
return 0;
if (p->u.tensor->ndim != q->u.tensor->ndim)
return 0;
for (i = 0; i < p->u.tensor->ndim; i++)
if (p->u.tensor->dim[i] != q->u.tensor->dim[i])
return 0;
return 1;
}
#if SELFTEST
static const char *s[] = {
"#test_tensor",
"a=(1,2,3)",
"",
"b=(4,5,6)",
"",
"c=(7,8,9)",
"",
"rank((a,b,c))",
"2",
"(a,b,c)",
"((1,2,3),(4,5,6),(7,8,9))",
// check tensor promotion
"((1,0),(0,0))",
"((1,0),(0,0))",
"a=quote(a)",
"",
"b=quote(b)",
"",
"c=quote(c)",
"",
};
void
test_tensor(void)
{
test(__FILE__, s, sizeof s / sizeof (char *));
}
#endif
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。