# -*- coding: utf-8 -*-
from pyfr.backends.base.generator import BaseKernelGenerator
[docs]class OpenMPKernelGenerator(BaseKernelGenerator):
[docs] def render(self):
if self.ndim == 1:
core = f'''
for (int _xi = 0; _xi < BLK_SZ; _xi += SOA_SZ)
{{
#pragma omp simd
for (int _xj = 0; _xj < SOA_SZ; _xj++)
{{
{self.body}
}}
}}'''
clean = f'''
for (int _xi = 0; _xi < _remi; _xi += SOA_SZ)
{{
#pragma omp simd
for (int _xj = 0; _xj < SOA_SZ; _xj++)
{{
{self.body}
}}
}}
for (int _xi = _remi, _xj = 0; _xj < _remj; _xj++)
{{
{self.body}
}}'''
else:
core = f'''
for (int _y = 0; _y < _ny; _y++)
{{
for (int _xi = 0; _xi < BLK_SZ; _xi += SOA_SZ)
{{
#pragma omp simd
for (int _xj = 0; _xj < SOA_SZ; _xj++)
{{
{self.body}
}}
}}
}}'''
clean = f'''
for (int _y = 0; _y < _ny; _y++)
{{
for (int _xi = 0; _xi < _remi; _xi += SOA_SZ)
{{
#pragma omp simd
for (int _xj = 0; _xj < SOA_SZ; _xj++)
{{
{self.body}
}}
}}
}}
for (int _y = 0; _y < _ny; _y++)
{{
for (int _xi = _remi, _xj = 0; _xj < _remj; _xj++)
{{
{self.body}
}}
}}'''
return f'''{self._render_spec()}
{{
int nci = _nx / BLK_SZ;
int _remi = ((_nx % BLK_SZ) / SOA_SZ)*SOA_SZ;
int _remj = (_nx % BLK_SZ) % SOA_SZ;
#define X_IDX (_xi + _xj)
#define X_IDX_AOSOA(v, nv)\
((_xi/SOA_SZ*(nv) + (v))*SOA_SZ + _xj)
#define BLK_IDX ib*BLK_SZ
#define BCAST_BLK(i, ld)\
((i) % (ld) + ((i) / (ld))*(ld)*_ny)
#pragma omp parallel for
for (int ib = 0; ib < nci; ib++)
{{
{core}
}}
int ib = nci;
{clean}
#undef X_IDX
#undef X_IDX_AOSOA
#undef BLK_IDX
#undef BCAST_BLK
}}'''
[docs] def ldim_size(self, name, *factor):
return '*'.join(['BLK_SZ'] + [str(f) for f in factor])
[docs] def needs_ldim(self, arg):
return False
[docs] def _render_spec(self):
# We first need the argument list; starting with the dimensions
kargs = ['int ' + d for d in self._dims]
# Now add any scalar arguments
kargs.extend(f'{sa.dtype} {sa.name}' for sa in self.scalargs)
# Finally, add the vector arguments
for va in self.vectargs:
# Views
if va.isview:
kargs.append(f'{va.dtype}* __restrict__ {va.name}_v')
kargs.append(f'const int* __restrict__ {va.name}_vix')
if va.ncdim == 2:
kargs.append(f'const int* __restrict__ {va.name}_vrstri')
# Arrays
else:
# Intent in arguments should be marked constant
const = 'const' if va.intent == 'in' else ''
kargs.append(f'{const} {va.dtype}* __restrict__ {va.name}_v'
.strip())
return 'void {0}({1})'.format(self.name, ', '.join(kargs))