The ultimate goal of the QMCkl library is to provide a high-performance -implementation of the main kernels of QMC. In this particular -implementation of the library, we focus on the definition of the API -and the tests, and on a pedagogical presentation of the -algorithms. We expect the HPC experts to use this repository as a -reference for re-writing optimized libraries. +implementation of the main kernels of Quantum Monte Carlo +methods. +In this particular implementation of the library, we focus on the +definition of the API and the tests, and on a pedagogical +presentation of the algorithms. We expect the HPC experts to use +this repository as a reference for re-writing optimized libraries.
@@ -357,7 +359,7 @@ and bug reports should be submitted at
diff --git a/qmckl.html b/qmckl.html index 7edcebb..27901dc 100644 --- a/qmckl.html +++ b/qmckl.html @@ -3,7 +3,7 @@ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> - +
The qmckl.h header file installed in the ${prefix}/include directory
@@ -385,12 +385,12 @@ Both files are located in the include/ directory.
In a traditional source code, most of the lines of source files of a program @@ -435,8 +435,8 @@ interactively, in the same spirit as Jupyter notebooks.
For a tutorial on literate programming with org-mode, follow this link. @@ -467,8 +467,8 @@ org-mode.
Most of the codes of the TREX CoE are written in Fortran with some scripts in @@ -516,8 +516,8 @@ For more guidelines on using Fortran to generate a C interface, see
The authors should follow the recommendations of the C99
@@ -535,8 +535,8 @@ Compliance can be checked with cppcheck as:
The proposed API should allow the library to: deal with memory transfers @@ -547,8 +547,8 @@ functions (see below).
To avoid namespace collisions, we use qmckl_ as a prefix for all exported
@@ -573,8 +573,8 @@ form is allowed.
In the C language, the number of bits used by the integer types can change @@ -606,15 +606,15 @@ bindings in other languages in other repositories.
Global variables should be avoided in the library, because it is
possible that one single program needs to use multiple instances
of the library. To solve this problem we propose to use a pointer
to a context variable, built by the library with the
-qmckl_context_create function. The =context= contains the global
+qmckl_context_create function. The =context= contains the global
state of the library, and is used as the first argument of many
QMCkl functions.
A single qmckl.h header to be distributed by the library
@@ -717,8 +717,8 @@ and the types definitions should be written in the *fh_type.f90 fil
Low-level functions are very simple functions which are leaves of @@ -727,14 +727,14 @@ the function call tree (they don't call any other QMCkl function).
These functions are pure, and unaware of the QMCkl
-context. They are not allowed to allocate/deallocate memory, and
+context. They are not allowed to allocate/deallocate memory, and
if they need temporary memory it should be provided in input.
High-level functions are at the top of the function call tree. @@ -747,20 +747,20 @@ temporary storage, to simplify the use of accelerators.
The high-level functions should be pure, unless the introduction
of non-purity is justified. All the side effects should be made in
-the context variable.
+the context variable.
The number of bits of precision required for a function should be
given as an input of low-level computational functions. This input
will be used to define the values of the different thresholds that
might be used to avoid computing unnecessary noise. High-level
-functions will use the precision specified in the context
+functions will use the precision specified in the context
variable.
Reducing the scaling of an algorithm usually implies also reducing @@ -847,7 +847,7 @@ implemented adapted to different problem sizes.
The following arrays are stored in the context: @@ -679,8 +679,8 @@ prim_factor = [ 1.0006253235944540e+01, 2.4169531573445120e+00, 7.96109248497664
typedef struct qmckl_ao_basis_struct { @@ -756,8 +756,8 @@ this mechanism.
In the following functions, when an array is passed as an argument
@@ -777,8 +777,8 @@ function returns true.
To set the basis set, all the following functions need to be @@ -816,20 +816,20 @@ the context.
qmckl_ao_gaussian_vgl computes the values, gradients and
@@ -1061,18 +1061,18 @@ Requirements
qmckl_exit_code qmckl_get_ao_basis_primitive_vgl(qmckl_context context, double* const primitive_vgl); @@ -1081,14 +1081,14 @@ Requirements
context is not QMCKL_NULL_CONTEXTqmckl_exit_code qmckl_ao_power ( @@ -1631,8 +1631,8 @@ the \(n\) points:
integer function qmckl_ao_power_f(context, n, X, LMAX, P, ldp) result(info) @@ -1683,15 +1683,15 @@ the \(n\) points:
integer(c_int32_t) function test_qmckl_ao_power(context) bind(C) @@ -1742,8 +1742,8 @@ the \(n\) points:
A polynomial is centered on a nucleus \(\mathbf{R}_i\)
@@ -1788,7 +1788,7 @@ Laplacians at a given point in space, of all polynomials with an
angular momentum up to lmax.
context is not QMCKL_NULL_CONTEXTqmckl_exit_code qmckl_ao_polynomial_vgl ( @@ -1912,8 +1912,8 @@ For example, with a=0, b=2 and c=1 the string is "yyz"
integer function qmckl_ao_polynomial_vgl_f(context, X, R, lmax, n, L, ldl, VGL, ldv) result(info) @@ -2047,16 +2047,16 @@ For example, with a=0, b=2 and c=1 the string is "yyz"
integer(c_int32_t) function test_qmckl_ao_polynomial_vgl(context) bind(C) @@ -2161,12 +2161,12 @@ assert(0 == test_qmckl_ao_polynomial_vgl(context));
qmckl_exit_code qmckl_get_ao_vgl(qmckl_context context, double* const ao_vgl); @@ -2175,14 +2175,14 @@ assert(0 == test_qmckl_ao_polynomial_vgl(context));
context is not QMCKL_NULL_CONTEXTqmckl_exit_code qmckl_dgemm ( @@ -517,8 +520,8 @@ TODO: Add description about the external library dependence.
integer function qmckl_dgemm_f(context, TransA, TransB, m, n, k, alpha, A, LDA, B, LDB, beta, C, LDC) & @@ -601,11 +604,23 @@ TODO: Add description about the external library dependence. endif if (TransA) then - C = matmul(AT,B) + if (alpha == 1.d0 && beta == 0.d0) then + C = matmul(AT,B) + else + C = beta*C + alpha*matmul(AT,B) + endif else if (TransB) then - C = matmul(A,BT) + if (alpha == 1.d0 && beta == 0.d0) then + C = matmul(A,BT) + else + C = beta*C + alpha*matmul(A,BT) + endif else - C = matmul(A,B) + if (alpha == 1.d0 && beta == 0.d0) then + C = matmul(A,B) + else + C = beta*C + alpha*matmul(A,B) + endif endif end function qmckl_dgemm_f@@ -617,7 +632,7 @@ TODO: Add description about the external library dependence.
The context variable is a handle for the state of the library,
@@ -338,7 +338,7 @@ A value of QMCKL_NULL_CONTEXT for the context is equivalent to a
typedef int64_t qmckl_context ; +typedef int64_t qmckl_context ; #define QMCKL_NULL_CONTEXT (qmckl_context) 0
ctx is a qmckl_context_struct* pointer.
The context keeps a ``date'' that allows to check which data needs @@ -367,7 +367,7 @@ coordinates are updated.
When a new element is added to the context, the functions -qmcklcontextcreate, qmcklcontextdestroy and qmcklcontextcopy +qmcklcontextcreate, qmcklcontextdestroy and qmcklcontextcopy should be updated inorder to make deep copies.
@@ -416,8 +416,8 @@ if the context is valid,QMCKL_NULL_CONTEXT otherwise.
To create a new context, qmckl_context_create() should be used.
@@ -502,8 +502,8 @@ To create a new context, qmckl_context_create() should be used.
For thread safety, the context may be locked/unlocked. The lock is @@ -548,8 +548,8 @@ number of times the thread has locked it is saved in the
qmckl_context_copy makes a deep copy of a context. It returns
@@ -597,8 +597,8 @@ number of times the thread has locked it is saved in the
The context is destroyed with qmckl_context_destroy, leaving the ancestors untouched.
@@ -652,7 +652,7 @@ It frees the context, and returns the previous context.
qmckl_distance_sqqmckl_distance_sq
qmckl_distance_sq computes the matrix of the squared distances
@@ -403,7 +403,7 @@ between all pairs of points in two sets, one point within each set:
\]
context is not QMCKL_NULL_CONTEXTqmckl_exit_code qmckl_distance_sq ( @@ -536,8 +536,8 @@ between all pairs of points in two sets, one point within each set:
integer function qmckl_distance_sq_f(context, transa, transb, m, n, & @@ -672,8 +672,8 @@ between all pairs of points in two sets, one point within each set:
This function is more efficient when A and B are
@@ -683,12 +683,12 @@ transposed.
qmckl_distanceqmckl_distance
qmckl_distance computes the matrix of the distances between all
@@ -706,7 +706,7 @@ If the input array is normal ('N'), the xyz coordinates are in
the leading dimension: [n][3] in C and (3,n) in Fortran.
context is not QMCKL_NULL_CONTEXT[n][3] in C and (3,n) in Fortra
qmckl_exit_code qmckl_distance ( @@ -839,8 +839,8 @@ the leading dimension:[n][3]in C and(3,n)in Fortra
integer function qmckl_distance_f(context, transa, transb, m, n, & @@ -1007,8 +1007,8 @@ the leading dimension:[n][3]in C and(3,n)in Fortra
This function is more efficient when A and B are transposed.
@@ -1018,12 +1018,12 @@ This function is more efficient when A and B are trans
qmckl_distance_rescaledqmckl_distance_rescaled
qmckl_distance_rescaled computes the matrix of the rescaled distances between all
@@ -1041,7 +1041,7 @@ If the input array is normal ('N'), the xyz coordinates are in
the leading dimension: [n][3] in C and (3,n) in Fortran.
context is not QMCKL_NULL_CONTEXT[n][3] in C and (3,n) in Fortra
qmckl_exit_code qmckl_distance_rescaled ( @@ -1182,8 +1182,8 @@ the leading dimension:[n][3]in C and(3,n)in Fortra
integer function qmckl_distance_rescaled_f(context, transa, transb, m, n, & @@ -1353,8 +1353,8 @@ the leading dimension:[n][3]in C and(3,n)in Fortra
This function is more efficient when A and B are transposed.
@@ -1363,12 +1363,12 @@ This function is more efficient when A and B are trans
qmckl_distance_rescaled_deriv_eqmckl_distance_rescaled_deriv_e
qmckl_distance_rescaled_deriv_e computes the matrix of the gradient and laplacian of the
@@ -1435,7 +1435,7 @@ If the input array is normal ('N'), the xyz coordinates are in
the leading dimension: [n][3] in C and (3,n) in Fortran.
context is not QMCKL_NULL_CONTEXT[n][3] in C and (3,n) in Fortra
qmckl_exit_code qmckl_distance_rescaled_deriv_e ( @@ -1576,8 +1576,8 @@ the leading dimension:[n][3]in C and(3,n)in Fortra
integer function qmckl_distance_rescaled_deriv_e_f(context, transa, transb, m, n, & @@ -1768,8 +1768,8 @@ the leading dimension:[n][3]in C and(3,n)in Fortra
This function is more efficient when A and B are transposed.
@@ -1781,7 +1781,7 @@ This function is more efficient when A and B are trans
The following data stored in the context: @@ -557,8 +557,8 @@ The following data stored in the context:
typedef struct qmckl_electron_struct { @@ -633,8 +633,8 @@ this mechanism.
Access functions return QMCKL_SUCCESS when the data has been
@@ -646,12 +646,12 @@ contains the requested data. Otherwise, this variable is untouched.
A walker is a set of electron coordinates that are arguments of
@@ -660,12 +660,12 @@ the wave function. walk_num is the number of walkers.
Returns the current electron coordinates. The pointer is assumed @@ -709,8 +709,8 @@ The order of the indices is:
To set the data relative to the electrons in the context, the @@ -797,8 +797,8 @@ in the context.
/* Reference input data */ @@ -904,8 +904,8 @@ rc = qmckl_get_electron_coord (context, 'N'
The computed data is stored in the context so that it can be reused @@ -918,12 +918,12 @@ current date is stored.
qmckl_exit_code qmckl_get_electron_ee_distance(qmckl_context context, double* const distance); @@ -932,10 +932,10 @@ current date is stored.