portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-09-20 01:05:38 +01:00
Code Releases Activity

Fixing nocompile

Browse Source
This commit is contained in:
Peter Boyle 2015-04-10 05:24:01 +02:00
parent 993419d9fb
commit 6e90038bf6
6 changed files with 31 additions and 39 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
Grid_config.h
View File

@ -11,10 +11,10 @@
/* #undef AVX512 */
/* GRID_COMMS_MPI */
#define GRID_COMMS_MPI 1
/* #undef GRID_COMMS_MPI */
/* GRID_COMMS_NONE */
/* #undef GRID_COMMS_NONE */
#define GRID_COMMS_NONE 1
/* Define to 1 if you have the `gettimeofday' function. */
#define HAVE_GETTIMEOFDAY 1

4
Grid_main.cc
View File

@ -407,7 +407,5 @@ int main (int argc, char ** argv)
} // loop for lat
} // loop for omp
MPI_Finalize();
Grid_finalize();
}

7
Grid_stencil.h
View File

@ -1,3 +1,5 @@
#ifndef GRID_STENCIL_H
#define GRID_STENCIL_H
//////////////////////////////////////////////////////////////////////////////////////////
// Must not lose sight that goal is to be able to construct really efficient
// gather to a point stencil code. CSHIFT is not the best way, so probably need
@ -45,14 +47,12 @@ namespace Grid {
void Stencil_local (int dimension,int shift,int cbmask);
void Stencil_comms (int dimension,int shift,int cbmask);
void Stencil_comms_simd(int dimension,int shift,int cbmask);
// Will need to implement actions for
//
Copy_plane;
Copy_plane_permute;
Gather_plane;
// The offsets to all neibours in stencil in each direction
int _checkerboard;
int _npoints; // Move to template param?
@ -360,3 +360,4 @@ namespace Grid {
};
#endif

9
Grid_vComplexD.h
View File

@ -186,15 +186,6 @@ namespace Grid {
Gextract<vComplexD,ComplexD>(y,extracted);
}
////////////////////////////////////////////////////////////////////////
// FIXME: gonna remove these load/store, get, set, prefetch
////////////////////////////////////////////////////////////////////////
void vload(zvec& a){
this->v = a;
}
zvec vget(){
return this->v ;
}
///////////////////////
// Splat
///////////////////////

9
Grid_vRealD.h
View File

@ -128,15 +128,6 @@ namespace Grid {
Gextract<vRealD,RealD>(y,extracted);
}
////////////////////////////////////////////////////////////////////////
// FIXME: gonna remove these load/store, get, set, prefetch
////////////////////////////////////////////////////////////////////////
void vload(dvec& a){
this->v = a;
}
dvec vget(){
return this->v ;
}
friend inline void vsplat(vRealD &ret,double a){
#if defined (AVX1)|| defined (AVX2)

37
TODO
View File

@ -1,12 +1,12 @@
* FIXME audit
* Remove vload/store etc..
* Replace vset with a call to merge.
* Replace vset with a call to merge.
* Replace vset with a call to merge.;
* care in Gmerge,Gextract over set.
* Const audit
* extract / merge extra implementation removal
* Conditional execution, where etc... -----DONE, simple test
* Conditional execution, where etc... -----DONE, simple test
* Integer relational support -----DONE
* Coordinate information, integers etc... -----DONE
* Integer type padding/union to vector. -----DONE
@ -22,18 +22,32 @@
-----Suggests generalised cartesian subblocking
sums, returning modified grid.
Two classes of subset;
i) red black parit subsetting.
(pick checkerboard).
i) Two classes of subset; red black parity subsetting (pick checkerboard).
cartesian sub-block subsetting
ii) Need to be able to project one Grid to another Grid.
Interface: (?)
Lattice<vobj> coarse_data SubBlockSum (GridBase *CoarseGrid, Lattice<vobj> &fine_data)
Operation ensure either:
rd[dim] divide rd[dim] fine_data
This will give a distributed array over mpi ranks in a given dim IF coarse gd != 1 and _processors[d]>1
Dimension can be *replicated* on all ranks in dimension. Need a "replicated" option on GridCartesian etc..
This will give "slice" summation and fourier projection assistance.
Generic concept is to subdivide (based on RD so applies to red/black or full).
Return a type on SUB-grid from CellSum TOP-grid
SUB-grid need not distribute but be replicated in some dims if that is how the
cartesian communicator works.
iii) No general permutation map.
Instead of subsetting
iii) No general permutation map.
* Consider switch std::vector to boost arrays.
boost::multi_array<type, 3> A()... to replace multi1d, multi2d etc..
@ -41,15 +55,12 @@ iii) No general permutation map.
*? Cell definition <-> sliceSum.
? Replicated arrays.
* Check for missing functionality - partially audited against QDP++ layout
* Optimise the extract/merge SIMD routines; Azusa??
-- I have collated into single location at least.
-- Need to use _mm_*insert/extract routines.
- I have collated into single location at least.
- Need to use _mm_*insert/extract routines.
* Conformable test in Cshift routines.