3
0
mirror of https://github.com/triqs/dft_tools synced 2024-12-25 13:53:40 +01:00
dft_tools/test/speed/array_short1.cpp
Olivier Parcollet b534936589 Fix matrix * alias issue and adapt det_manip
- The previous version of the * operator for matrix was too clever.
It was giving a lazy object and then rewriting C = A *B into gemm (a,A,B,0,C).
The pb was in case of aliasing : when e.g. C = A, or is a part of A.
gemm is not correct that case, and as a result generic code like
a = a *b
may not be correct in matrix case, which is unacceptable.

- So we revert to a simple * operator for matrix
that does immediate computation.
Same thing for matrix* vector

- we also suppress a_x_ty class.

-> for M = a * b,
when M is a matrix, there is no overhead due to move assignment
-> however, when M is a view, there is an additionnal copy.

-Correctness comes first, hence the fix.
However, if one wants more speed and one can guarantee that
there is no aliasing possible, then one has to write a direct gemm call.

-> det_manip class was adapted, since in that case, we can show there
no alias, and we want the speed gain, so the * ops where replaced
by direct blas call (using the array blas interface).

-> also gemm, gemv, ger were overloaded in the case the return
matrix/vector (i.e. last parameter of the function) is not an lvalue,
but a temporary view created on the fly.
2013-09-11 18:49:22 +02:00

168 lines
4.9 KiB
C++

/*******************************************************************************
*
* TRIQS: a Toolbox for Research in Interacting Quantum Systems
*
* Copyright (C) 2011 by O. Parcollet
*
* TRIQS is free software: you can redistribute it and/or modify it under the
* terms of the GNU General Public License as published by the Free Software
* Foundation, either version 3 of the License, or (at your option) any later
* version.
*
* TRIQS is distributed in the hope that it will be useful, but WITHOUT ANY
* WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
* FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
* details.
*
* You should have received a copy of the GNU General Public License along with
* TRIQS. If not, see <http://www.gnu.org/licenses/>.
*
******************************************************************************/
#include <triqs/arrays.hpp>
#include <triqs/arrays/linalg/det_and_inverse.hpp>
using namespace triqs::arrays;
using namespace triqs;
const int N1= 10, N2 = 10;
struct plain_for_no_ptr_const {
void operator()() {
triqs::arrays::array<double,2> A (N1,N2,FORTRAN_LAYOUT);
for (int u =0; u<2500000; ++u)
for (int j=0; j<N2; ++j)
for (int i =0; i<N1; ++i)
A(i,j) = 1876;
}
};
struct assign_to_const {
void operator()() {
triqs::arrays::array<double,2,TRAVERSAL_ORDER_FORTRAN> A (N1,N2,FORTRAN_LAYOUT);
auto V = make_view(A);
for (int u =0; u<2500000; ++u)
//make_view(A) = 1867;
V= 1867;
}
};
struct plain_for_no_ptr {
void operator()() {
triqs::arrays::matrix<double > A (N1,N2, FORTRAN_LAYOUT);
for (int u =0; u<2500000; ++u)
for (int i =0; i<N1; ++i)
for (int j=0; j<N2; ++j)
A(i,j) = 10*i+ j;
}
};
//typedef double value_type;
//typedef triqs::arrays::matrix<double>::indexmap_type::domain_type::index_value_type index_value_type;
struct F {
triqs::arrays::matrix<double,TRAVERSAL_ORDER_FORTRAN > & a;
F(triqs::arrays::matrix<double,TRAVERSAL_ORDER_FORTRAN > & a_): a(a_){}
//void operator()(value_type & p, index_value_type const & key) const { p=key[0]*10 + key[1] ;}
//void operator()(value_type & p, value_type const & x0, value_type const & x1) const { p=x0*10 + x1 ;}
template<typename X0, typename X1> void operator()( X0 const & x0, X1 const & x1) const { a(x0,x1) =x0*10 + x1 ;}
};
struct with_foreach {
void operator()() {
triqs::arrays::matrix<double,TRAVERSAL_ORDER_FORTRAN > A (N1,N2,FORTRAN_LAYOUT);
// triqs::arrays::indexmaps::
for (int u =0; u<2500000; ++u) foreach(A,F(A));
// for (int u =0; u<5000; ++u) make_view(A) = 1867;
}
};
struct iterators {
void operator()() {
typedef triqs::arrays::matrix<double > MM;
MM A (N1,N2,FORTRAN_LAYOUT);
for (int u =0; u<3000; ++u)
{
for (auto it = A.begin(); it; ++it) {
*it =10*it.indices()[0] + it.indices()[1] ;
}
}
}
};
template<class int_type>
struct pointer_generic_form {
void operator()() {
triqs::arrays::matrix<double > A (N1,N2, FORTRAN_LAYOUT);
const int_type s0 = A.indexmap().strides()[0];
const int_type s1 = A.indexmap().strides()[1];
const int_type l0 = A.indexmap().lengths()[0];
const int_type l1 = A.indexmap().lengths()[1];
for (int u =0; u<1000; ++u) {
double * restrict p= &(A(0,0));
//typedef int int_type;
//assert(s0==1); assert(s1==N1);
//assert(l0==N1); assert(l1==N2);
for (int_type i0 =0; i0<l0; ++i0)
for (int_type i1 =0; i1<l1; ++i1)
p[i0*s0 + i1*s1] = 10*i0+ i1;
}
}
};
struct const_with_iterators {
void operator()() {
typedef triqs::arrays::matrix<double> MM;
MM A (N1,N2,FORTRAN_LAYOUT);
for (int u =0; u<5000; ++u)
{
for (MM::iterator it = A.begin(); it; ++it) {
*it = 1876;
}
}
}
};
struct foreach_manual_simple {
void operator()() {
triqs::arrays::array<double,2> A (N1,N2,FORTRAN_LAYOUT);
mini_vector<int,2> t;
const int u[2] = {1,0}; //=1, u1=0;
//int u[2]; for (int i=0; i<2; ++i) u[i] = 1-i;
for (int k =0; k<5000; ++k)
for (t[u[0]]=0; t[u[0]]<A.shape(u[0]); ++t[u[0]])
for (t[u[1]] =0; t[u[1]]<A.shape(u[1]); ++t[u[1]])
A(t[u[1]],t[u[0]]) = 1876;
//A[t] = 1876;
}
};
#include "./speed_tester.hpp"
int main() {
const int l = 100;
std::cout << "----------- Comparing const assign : A() =const to a manual loop --------------------"<< std::endl ;
speed_tester<plain_for_no_ptr_const> (l);
speed_tester<assign_to_const> (l);
std::cout << "----------- Comparing manual loop and foreach for A(i,j) = 10 *i + j--------------------"<< std::endl ;
speed_tester<plain_for_no_ptr> (l);
speed_tester<with_foreach> (l);
//speed_tester<pointer_generic_form <std::ptrdiff_t> > (l);
//speed_tester<pointer_generic_form < size_t> > (l);
//speed_tester<foreach_manual_simple>(l);
std::cout << "----------- Comparing manual loop and foreach for A(i,j) = 10 *i + j--------------------"<< std::endl ;
speed_tester<iterators> (l);
speed_tester<const_with_iterators> (l);
return 0;
}