llvm-project/polly/lib/ScheduleOptimizer.cpp

252 lines
8.5 KiB
C++
Raw Normal View History

//===- Schedule.cpp - Calculate an optimized schedule ---------------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This pass the isl to calculate a schedule that is optimized for parallelism
// and tileablility. The algorithm used in isl is an optimized version of the
// algorithm described in following paper:
//
// U. Bondhugula, A. Hartono, J. Ramanujam, and P. Sadayappan.
// A Practical Automatic Polyhedral Parallelizer and Locality Optimizer.
// In Proceedings of the 2008 ACM SIGPLAN Conference On Programming Language
// Design and Implementation, PLDI 08, pages 101113. ACM, 2008.
//===----------------------------------------------------------------------===//
#include "polly/Cloog.h"
#include "polly/LinkAllPasses.h"
#include "polly/Dependences.h"
#include "polly/ScopInfo.h"
#include "isl/dim.h"
#include "isl/map.h"
#include "isl/constraint.h"
#include "isl/schedule.h"
#define DEBUG_TYPE "polly-optimize-isl"
#include "llvm/Support/Debug.h"
using namespace llvm;
using namespace polly;
namespace {
class ScheduleOptimizer : public ScopPass {
public:
static char ID;
explicit ScheduleOptimizer() : ScopPass(ID) {}
virtual bool runOnScop(Scop &S);
void printScop(llvm::raw_ostream &OS) const;
void getAnalysisUsage(AnalysisUsage &AU) const;
};
}
char ScheduleOptimizer::ID = 0;
static int getSingleMap(__isl_take isl_map *map, void *user) {
isl_map **singleMap = (isl_map **) user;
*singleMap = map;
return 0;
}
static void extendScattering(Scop &S, unsigned scatDimensions) {
for (Scop::iterator SI = S.begin(), SE = S.end(); SI != SE; ++SI) {
ScopStmt *stmt = *SI;
if (stmt->isFinalRead())
continue;
isl_map *scattering = stmt->getScattering();
isl_dim *dim = isl_dim_alloc(isl_map_get_ctx(scattering),
isl_map_n_param(scattering),
isl_map_n_out(scattering),
scatDimensions);
isl_basic_map *changeScattering = isl_basic_map_universe(isl_dim_copy(dim));
for (unsigned i = 0; i < isl_map_n_out(scattering); i++) {
isl_constraint *c = isl_equality_alloc(isl_dim_copy(dim));
isl_constraint_set_coefficient_si(c, isl_dim_in, i, 1);
isl_constraint_set_coefficient_si(c, isl_dim_out, i, -1);
changeScattering = isl_basic_map_add_constraint(changeScattering, c);
}
for (unsigned i = isl_map_n_out(scattering); i < scatDimensions; i++) {
isl_constraint *c = isl_equality_alloc(isl_dim_copy(dim));
isl_constraint_set_coefficient_si(c, isl_dim_out, i, 1);
changeScattering = isl_basic_map_add_constraint(changeScattering, c);
}
isl_map *changeScatteringMap = isl_map_from_basic_map(changeScattering);
stmt->setScattering(isl_map_apply_range(scattering, changeScatteringMap));
}
}
// @brief Tile a band.
//
// This function recieves a map that assigns to the instances of a statement
// an execution time.
//
// [i_0, i_1, i_2] -> [o_0, o_1, o_2, i_0, i_1, i_2]:
// o_0 % 32 = 0 and o_1 % 32 = 0 and o_2 % 32 = 0
// and o0 <= i0 <= o0 + 32 and o1 <= i1 <= o1 + 32 and o2 <= i2 <= o2 + 32
static isl_map *tileBand(isl_map *band) {
int dimensions = isl_map_n_out(band);
int tileSize = 32;
isl_dim *dim = isl_dim_alloc(isl_map_get_ctx(band), isl_map_n_param(band),
dimensions, dimensions * 3);
isl_basic_map *tiledBand = isl_basic_map_universe(isl_dim_copy(dim));
for (int i = 0; i < dimensions; i++) {
isl_constraint *c = isl_equality_alloc(isl_dim_copy(dim));
isl_constraint_set_coefficient_si(c, isl_dim_out, i, 1);
isl_constraint_set_coefficient_si(c, isl_dim_out, 2 * dimensions + i,
-tileSize);
tiledBand = isl_basic_map_add_constraint(tiledBand, c);
c = isl_equality_alloc(isl_dim_copy(dim));
isl_constraint_set_coefficient_si(c, isl_dim_in, i, -1);
isl_constraint_set_coefficient_si(c, isl_dim_out, dimensions + i, 1);
tiledBand = isl_basic_map_add_constraint(tiledBand, c);
c = isl_inequality_alloc(isl_dim_copy(dim));
isl_constraint_set_coefficient_si(c, isl_dim_out, i, -1);
isl_constraint_set_coefficient_si(c, isl_dim_out, dimensions + i, 1);
tiledBand = isl_basic_map_add_constraint(tiledBand, c);
c = isl_inequality_alloc(isl_dim_copy(dim));
isl_constraint_set_coefficient_si(c, isl_dim_out, i, 1);
isl_constraint_set_coefficient_si(c, isl_dim_out, dimensions + i, -1);
isl_constraint_set_constant_si(c, tileSize - 1);
tiledBand = isl_basic_map_add_constraint(tiledBand, c);
}
// Project out auxilary dimensions (introduced to ensure 'ii % tileSize = 0')
//
// The real dimensions are transformed into existentially quantified ones.
// This reduces the number of visible scattering dimensions. Also, Cloog
// produces better code, if auxilary dimensions are existentially quantified.
tiledBand = isl_basic_map_project_out(tiledBand, isl_dim_out, 2 * dimensions,
dimensions);
return isl_map_apply_range(band, isl_map_from_basic_map(tiledBand));
}
bool ScheduleOptimizer::runOnScop(Scop &S) {
Dependences *D = &getAnalysis<Dependences>();
// Build input data.
int dependencyKinds = Dependences::TYPE_RAW
| Dependences::TYPE_WAR
| Dependences::TYPE_WAW;
isl_union_map *validity = D->getDependences(dependencyKinds);
isl_union_map *proximity = D->getDependences(dependencyKinds);
isl_union_set *domain = NULL;
for (Scop::iterator SI = S.begin(), SE = S.end(); SI != SE; ++SI)
if ((*SI)->isFinalRead())
continue;
else if (!domain)
domain = isl_union_set_from_set((*SI)->getDomain());
else
domain = isl_union_set_union(domain,
isl_union_set_from_set((*SI)->getDomain()));
if (!domain)
return false;
DEBUG(dbgs() << "\n\nCompute schedule from: ");
DEBUG(dbgs() << "Domain := "; isl_union_set_dump(domain); dbgs() << ";\n");
DEBUG(dbgs() << "Proximity := "; isl_union_map_dump(proximity);
dbgs() << ";\n");
DEBUG(dbgs() << "Validity := "; isl_union_map_dump(validity);
dbgs() << ";\n");
isl_schedule *schedule;
schedule = isl_union_set_compute_schedule(domain, validity, proximity);
// Get the complete schedule.
isl_union_map *scheduleMap = isl_schedule_get_map(schedule);
DEBUG(dbgs() << "Computed schedule: ");
DEBUG(isl_union_map_dump(scheduleMap));
DEBUG(dbgs() << "Individual bands: ");
// Get individual tileable bands.
for (int i = 0; i < isl_schedule_n_band(schedule); i++) {
isl_union_map *band = isl_schedule_get_band(schedule, i);
DEBUG(dbgs() << "Band " << i << ": ");
DEBUG(isl_union_map_dump(band));
for (Scop::iterator SI = S.begin(), SE = S.end(); SI != SE; ++SI) {
ScopStmt *stmt = *SI;
if (stmt->isFinalRead())
continue;
isl_set *domain = stmt->getDomain();
isl_union_map *stmtBand;
stmtBand = isl_union_map_intersect_domain(isl_union_map_copy(band),
isl_union_set_from_set(domain));
isl_map *sband;
isl_union_map_foreach_map(stmtBand, getSingleMap, &sband);
sband = tileBand(sband);
DEBUG(dbgs() << "tiled band: ");
DEBUG(isl_map_dump(sband));
if (i == 0)
stmt->setScattering(sband);
else {
isl_map *scattering = stmt->getScattering();
scattering = isl_map_range_product(scattering, sband);
scattering = isl_map_flatten(scattering);
stmt->setScattering(scattering);
}
}
}
unsigned maxScatDims = 0;
for (Scop::iterator SI = S.begin(), SE = S.end(); SI != SE; ++SI)
maxScatDims = std::max(isl_map_n_out((*SI)->getScattering()), maxScatDims);
extendScattering(S, maxScatDims);
isl_schedule_free(schedule);
return false;
}
void ScheduleOptimizer::printScop(raw_ostream &OS) const {
}
void ScheduleOptimizer::getAnalysisUsage(AnalysisUsage &AU) const {
ScopPass::getAnalysisUsage(AU);
AU.addRequired<Dependences>();
}
static RegisterPass<ScheduleOptimizer> A("polly-optimize-isl",
"Polly - Calculate optimized "
"schedules using the isl schedule "
"calculator");
Pass* polly::createScheduleOptimizerPass() {
return new ScheduleOptimizer();
}