portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2024-11-10 07:55:35 +00:00
Code Releases Activity

Hadrons: DiskVector save-on-eviction and faster CRC32 for Eigen matrices

Browse Source
This commit is contained in:
Antonin Portelli 2018-10-18 17:48:25 +01:00
parent f709329d96
commit f333f3e575
2 changed files with 63 additions and 43 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
Grid/util/Sha.h
View File

@ -34,10 +34,11 @@ extern "C" {
class GridChecksum
{
public:
static inline uint32_t crc32(void *data,size_t bytes)
static inline uint32_t crc32(const void *data,size_t bytes)
{
return ::crc32(0L,(unsigned char *)data,bytes);
}
template <typename T>
static inline std::string sha256_string(const std::vector<T> &hash)
{

103
Hadrons/DiskVector.hpp
View File

@ -59,14 +59,18 @@ public:
: master_(master), cmaster_(master), i_(i) {}
// operator=: somebody is trying to store a vector element
// write to disk and cache
// write to cache and tag as modified
T &operator=(const T &obj) const
{
auto &cache = *master_.cachePtr_;
auto &modified = *master_.modifiedPtr_;
auto &index = *master_.indexPtr_;
DV_DEBUG_MSG(&master_, "writing to " << i_);
master_.cacheInsert(i_, obj);
master_.save(master_.filename(i_), obj);
modified[index.at(i_)] = true;
return master_.cachePtr_->at(i_);
return cache[index.at(i_)];
}
// implicit cast to const object reference and redirection
@ -98,13 +102,14 @@ private:
void cacheInsert(const unsigned int i, const T &obj) const;
void clean(void);
private:
std::string dirname_;
unsigned int size_, cacheSize_;
double access_{0.}, hit_{0.};
bool clean_;
std::string dirname_;
unsigned int size_, cacheSize_;
double access_{0.}, hit_{0.};
bool clean_;
// using pointers to allow modifications when class is const
// semantic: const means data unmodified, but cache modification allowed
std::unique_ptr<std::vector<T>> cachePtr_;
std::unique_ptr<std::vector<bool>> modifiedPtr_;
std::unique_ptr<std::map<unsigned int, unsigned int>> indexPtr_;
std::unique_ptr<std::stack<unsigned int>> freePtr_;
std::unique_ptr<std::deque<unsigned int>> loadsPtr_;
@ -149,24 +154,24 @@ public:
private:
virtual void load(EigenDiskVectorMat<T> &obj, const std::string filename) const
{
std::ifstream f(filename, std::ios::binary);
std::vector<unsigned char> hash(SHA256_DIGEST_LENGTH);
Eigen::Index nRow, nCol;
size_t matSize;
double t;
std::ifstream f(filename, std::ios::binary);
uint32_t crc, check;
Eigen::Index nRow, nCol;
size_t matSize;
double t;
f.read(reinterpret_cast<char *>(hash.data()), hash.size()*sizeof(unsigned char));
f.read(reinterpret_cast<char *>(&nRow), sizeof(Eigen::Index));
f.read(reinterpret_cast<char *>(&nCol), sizeof(Eigen::Index));
f.read(reinterpret_cast<char *>(&crc), sizeof(crc));
f.read(reinterpret_cast<char *>(&nRow), sizeof(nRow));
f.read(reinterpret_cast<char *>(&nCol), sizeof(nCol));
obj.resize(nRow, nCol);
matSize = nRow*nCol*sizeof(T);
t = -usecond();
f.read(reinterpret_cast<char *>(obj.data()), matSize);
t += usecond();
DV_DEBUG_MSG(this, "Eigen read " << matSize/t*1.0e6/1024/1024 << " MB/s");
auto check = GridChecksum::sha256(obj.data(), matSize);
DV_DEBUG_MSG(this, "Eigen sha256 " << GridChecksum::sha256_string(check));
if (hash != check)
check = GridChecksum::crc32(obj.data(), matSize);
DV_DEBUG_MSG(this, "Eigen crc32 " << std::hex << check << std::dec);
if (crc != check)
{
HADRONS_ERROR(Io, "checksum failed")
}
@ -174,24 +179,24 @@ private:
virtual void save(const std::string filename, const EigenDiskVectorMat<T> &obj) const
{
std::ofstream f(filename, std::ios::binary);
std::vector<unsigned char> hash(SHA256_DIGEST_LENGTH);
Eigen::Index nRow, nCol;
size_t matSize;
double t;
std::ofstream f(filename, std::ios::binary);
uint32_t crc;
Eigen::Index nRow, nCol;
size_t matSize;
double t;
nRow = obj.rows();
nCol = obj.cols();
matSize = nRow*nCol*sizeof(T);
hash = GridChecksum::sha256(obj.data(), matSize);
DV_DEBUG_MSG(this, "Eigen sha256 " << GridChecksum::sha256_string(hash));
f.write(reinterpret_cast<char *>(hash.data()), hash.size()*sizeof(unsigned char));
f.write(reinterpret_cast<char *>(&nRow), sizeof(Eigen::Index));
f.write(reinterpret_cast<char *>(&nCol), sizeof(Eigen::Index));
crc = GridChecksum::crc32(obj.data(), matSize);
f.write(reinterpret_cast<char *>(&crc), sizeof(crc));
f.write(reinterpret_cast<char *>(&nRow), sizeof(nRow));
f.write(reinterpret_cast<char *>(&nCol), sizeof(nCol));
t = -usecond();
f.write(reinterpret_cast<const char *>(obj.data()), matSize);
t += usecond();
DV_DEBUG_MSG(this, "Eigen write " << matSize/t*1.0e6/1024/1024 << " MB/s");
DV_DEBUG_MSG(this, "Eigen crc32 " << std::hex << crc << std::dec);
}
};
@ -205,6 +210,7 @@ DiskVectorBase<T>::DiskVectorBase(const std::string dirname,
const bool clean)
: dirname_(dirname), size_(size), cacheSize_(cacheSize), clean_(clean)
, cachePtr_(new std::vector<T>(size))
, modifiedPtr_(new std::vector<bool>(size, false))
, indexPtr_(new std::map<unsigned int, unsigned int>())
, freePtr_(new std::stack<unsigned int>)
, loadsPtr_(new std::deque<unsigned int>())
@ -311,15 +317,24 @@ std::string DiskVectorBase<T>::filename(const unsigned int i) const
template <typename T>
void DiskVectorBase<T>::evict(void) const
{
auto &index = *indexPtr_;
auto &freeInd = *freePtr_;
auto &loads = *loadsPtr_;
auto &cache = *cachePtr_;
auto &modified = *modifiedPtr_;
auto &index = *indexPtr_;
auto &freeInd = *freePtr_;
auto &loads = *loadsPtr_;
if (index.size() >= cacheSize_)
{
DV_DEBUG_MSG(this, "evicting " << loads.front());
freeInd.push(index.at(loads.front()));
index.erase(loads.front());
unsigned int i = loads.front();
DV_DEBUG_MSG(this, "evicting " << i);
if (modified[index.at(i)])
{
DV_DEBUG_MSG(this, "element " << i << " modified, saving to disk");
save(filename(i), cache[index.at(i)]);
}
freeInd.push(index.at(i));
index.erase(i);
loads.pop_front();
}
}
@ -327,10 +342,11 @@ void DiskVectorBase<T>::evict(void) const
template <typename T>
void DiskVectorBase<T>::fetch(const unsigned int i) const
{
auto &cache = *cachePtr_;
auto &index = *indexPtr_;
auto &freeInd = *freePtr_;
auto &loads = *loadsPtr_;
auto &cache = *cachePtr_;
auto &modified = *modifiedPtr_;
auto &index = *indexPtr_;
auto &freeInd = *freePtr_;
auto &loads = *loadsPtr_;
struct stat s;
@ -346,21 +362,24 @@ void DiskVectorBase<T>::fetch(const unsigned int i) const
freeInd.pop();
load(cache[index.at(i)], filename(i));
loads.push_back(i);
modified[index.at(i)] = false;
}
template <typename T>
void DiskVectorBase<T>::cacheInsert(const unsigned int i, const T &obj) const
{
auto &cache = *cachePtr_;
auto &index = *indexPtr_;
auto &freeInd = *freePtr_;
auto &loads = *loadsPtr_;
auto &cache = *cachePtr_;
auto &modified = *modifiedPtr_;
auto &index = *indexPtr_;
auto &freeInd = *freePtr_;
auto &loads = *loadsPtr_;
evict();
index[i] = freeInd.top();
freeInd.pop();
cache[index.at(i)] = obj;
loads.push_back(i);
modified[index.at(i)] = false;
#ifdef DV_DEBUG
std::string msg;