EIC code displayed by LXR master/​include/​Geant4/​G4Cache.hh	Source navigation Diff markup Identifier search General search
	Version: master test Revert   Change

File indexing completed on 2026-09-21 09:09:58

 //
 // ********************************************************************
 // * License and Disclaimer                                           *
 // *                                                                  *
 // * The  Geant4 software  is  copyright of the Copyright Holders  of *
 // * the Geant4 Collaboration.  It is provided  under  the terms  and *
 // * conditions of the Geant4 Software License,  included in the file *
 // * LICENSE and available at  http://cern.ch/geant4/license .  These *
 // * include a list of copyright holders.                             *
 // *                                                                  *
 // * Neither the authors of this software system, nor their employing *
 // * institutes,nor the agencies providing financial support for this *
 // * work  make  any representation or  warranty, express or implied, *
 // * regarding  this  software system or assume any liability for its *
 // * use.  Please see the license in the file  LICENSE  and URL above *
 // * for the full disclaimer and the limitation of liability.         *
 // *                                                                  *
 // * This  code  implementation is the result of  the  scientific and *
 // * technical work of the GEANT4 collaboration.                      *
 // * By using,  copying,  modifying or  distributing the software (or *
 // * any work based  on the software)  you  agree  to acknowledge its *
 // * use  in  resulting  scientific  publications,  and indicate your *
 // * acceptance of all terms of the Geant4 Software license.          *
 // ********************************************************************
 //
 // G4Cache
 //
 // Class Description:
 //
 //      Helper classes for Geant4 Multi-Threaded.
 //      The classes defined in this header file provide a thread-private
 //      cache to store a thread-local variable V in a class instance
 //      shared among threads.
 //      These are templated classes on the to-be-stored object.
 //
 // Example:
 //      Let's assume an instance myObject of class G4Shared is sharead between
 //      threads. Still a data member of this class needs to be thread-private.
 //      A typical example of this being a "cache" for a local calculation.
 //      The helper defined here can be used to guarantee thread-safe operations
 //      on the thread-private object.
 //      Example:
 //         class G4Shared
 //         {
 //           G4double sharedData;
 //           G4Cache<G4double> threadPrivate;
 //           void foo()
 //           {
 //             G4double priv = threadPrivate.Get();
 //             if ( priv < 10 ) priv += sharedData;
 //             threadPrivate.Put( priv );
 //           }
 //         }
 //
 //      Two variants of the base G4Cache exist. The first one being
 //      G4VectorCache similar to std::vector.
 //      Example:
 //         G4VectorCache<G4double> aVect;
 //         aVect.Push_back( 3.2 );
 //         aVect.Push_back( 4.1 );
 //         std::cout << aVect[0] << std::endl;
 //      The second one being:
 //      G4MapCache, similar to std::map.
 //      Example:
 //         G4MapCache<G4int, G4double> aMap;
 //         aMap[320]=1.234;
 //
 //      See classes definition for details.
 
 // Author: A.Dotti, 21 October 2013 - First implementation
 // --------------------------------------------------------------------
 #ifndef G4CACHE_HH
 #define G4CACHE_HH
 
 // Debug this code
 // #define g4cdebug 1
 
 #include <atomic>
 #include <map>
 #include <system_error>
 
 #include "G4AutoLock.hh"
 #include "G4CacheDetails.hh"  // Thread Local storage details are here
 
 // A templated cache to store a thread-private data of type VALTYPE.
 //
 template <class VALTYPE>
 class G4Cache
 {
  public:
   using value_type = VALTYPE;
   // The stored type
 
   G4Cache();
   // Default constructor
 
   G4Cache(const value_type& v);
   // Construct cache object with initial value
 
   virtual ~G4Cache();
   // Default destructor
 
   inline value_type& Get() const;
   // Gets reference to cached value of this threads
 
   inline void Put(const value_type& val) const;
   // Sets this thread cached value to val
 
   inline value_type Pop();
   // Gets copy of cached value
 
   G4Cache(const G4Cache& rhs);
   G4Cache& operator=(const G4Cache& rhs);
 
  protected:
   const G4int& GetId() const { return id; }
 
  private:
   G4int id;
   mutable G4CacheReference<value_type> theCache;
   static std::atomic<unsigned int> instancesctr;
   static std::atomic<unsigned int> dstrctr;
 
   inline value_type& GetCache() const
   {
     theCache.Initialize(id);
     return theCache.GetCache(id);
   }
 };
 
 // A vector version of the cache. Implements vector interface.
 // Can be used directly as a std::vector would be used.
 //
 template <class VALTYPE>
 class G4VectorCache : public G4Cache<std::vector<VALTYPE>>
 {
  public:
   // Some useful definitions
   //
   using value_type = VALTYPE;
   using vector_type = typename std::vector<value_type>;
   using size_type = typename vector_type::size_type;
   using iterator = typename vector_type::iterator;
   using const_iterator = typename vector_type::const_iterator;
 
   G4VectorCache();
   // Default constructor
 
   G4VectorCache(G4int nElems);
   // Creates a vector cache of nElems elements
 
   G4VectorCache(G4int nElems, value_type* vals);
   // Creates a vector cache with elements from an array
 
   virtual ~G4VectorCache();
   // Default destructor
 
   // Interface with functionalities of similar name of std::vector
   //
   inline void Push_back(const value_type& val);
   inline value_type Pop_back();
   inline value_type& operator[](const G4int& idx);
   inline iterator Begin();
   inline iterator End();
   inline void Clear();
   inline size_type Size() { return G4Cache<vector_type>::Get().size(); }
   //  Needs to be here for a VC9 compilation problem
 };
 
 // a Map version of the cache. Implements std::map interface.
 // Can be used directly as a std::map would be used.
 // KEYTYPE being the key type and VALTYPE the value type.
 //
 template <class KEYTYPE, class VALTYPE>
 class G4MapCache : public G4Cache<std::map<KEYTYPE, VALTYPE>>
 {
  public:
   // Some useful definitions
   //
   using key_type = KEYTYPE;
   using value_type = VALTYPE;
   using map_type = typename std::map<key_type, value_type>;
   using size_type = typename map_type::size_type;
   using iterator = typename map_type::iterator;
   using const_iterator = typename map_type::const_iterator;
 
   virtual ~G4MapCache();
   // Default destructor
 
   inline G4bool Has(const key_type& k);
   // Returns true if map contains element corresponding to key k
 
   // Interface with functionalities of similar name of std::map
   //
   inline std::pair<iterator, G4bool> Insert(const key_type& k,
                                             const value_type& v);
   inline iterator Begin();
   inline iterator End();
   inline iterator Find(const key_type& k);
   inline value_type& Get(const key_type& k);
   inline size_type Erase(const key_type& k);
   inline value_type& operator[](const key_type& k);
   inline size_type Size() { return G4Cache<map_type>::Get().size(); }
   //  Needs to be here for a VC9 compilation problem
 };
 
 //========= Implementation: G4Cache<V> ====================================
 
 template <class V>
 G4Cache<V>::G4Cache()
 {
   G4AutoLock l(G4TypeMutex<G4Cache<V>>());
   id = instancesctr++;
   theCache.Initialize(id);
 #ifdef g4cdebug
   std::cout << "G4Cache id: " << id << std::endl;
 #endif
 }
 
 template <class V>
 G4Cache<V>::G4Cache(const G4Cache<V>& rhs)
 {
   // Copy is special, we need to copy the content
   // of the cache, not the cache object
 
   if(this == &rhs)
     return;
   G4AutoLock l(G4TypeMutex<G4Cache<V>>());
   id = instancesctr++;
 
   // Force copy of cached data
   //
   V aCopy = rhs.GetCache();
   Put(aCopy);
 
 #ifdef g4cdebug
   std::cout << "Copy constructor with id: " << id << std::endl;
 #endif
 }
 
 template <class V>
 G4Cache<V>& G4Cache<V>::operator=(const G4Cache<V>& rhs)
 {
   if(this == &rhs)
     return *this;
 
   // Force copy of cached data
   //
   V aCopy = rhs.GetCache();
   Put(aCopy);
 
 #ifdef g4cdebug
   std::cout << "Assignement operator with id: " << id << std::endl;
 #endif
   return *this;
 }
 
 template <class V>
 G4Cache<V>::G4Cache(const V& v)
 {
   G4AutoLock l(G4TypeMutex<G4Cache<V>>());
   id = instancesctr++;
   Put(v);
 
 #ifdef g4cdebug
   std::cout << "G4Cache id: " << id << std::endl;
 #endif
 }
 
 template <class V>
 G4Cache<V>::~G4Cache()
 {
 #ifdef g4cdebug
   std::cout << "~G4Cache id: " << id << std::endl;
 #endif
   // don't automatically lock --> wait until we can catch an error
   // without scoping the G4AutoLock
   //
   G4AutoLock l(G4TypeMutex<G4Cache<V>>(), std::defer_lock);
 
   // sometimes the mutex is unavailable in destructors so
   // try to lock the associated mutex, but catch if it fails
   try
   {
     // a system_error in lock means that the mutex is unavailable
     // we want to throw the error that comes from locking an unavailable
     // mutex so that we know there is a memory leak
     // if the mutex is valid, this will hold until the other thread finishes
     //
     l.lock();
   } catch(std::system_error& e)
   {
     // the error that comes from locking an unavailable mutex
 #ifdef g4cdebug
     std::cout << "Non-critical error: mutex lock failure in ~G4Cache<"
            << typeid(V).name() << ">. " << std::endl
            << "If the RunManagerKernel has been deleted, it failed to "
            << "delete an allocated resource" << std::endl
            << "and this destructor is being called after the statics "
            << "were destroyed." << std::endl;
     std::cout << "Exception: [code: " << e.code() << "] caught: " << e.what()
            << std::endl;
 #endif
   }
   ++dstrctr;
   G4bool last = (dstrctr == instancesctr);
   theCache.Destroy(id, last);
   if(last)
   {
     instancesctr.store(0);
     dstrctr.store(0);
   }
 }
 
 template <class V>
 V& G4Cache<V>::Get() const
 {
   return GetCache();
 }
 
 template <class V>
 void G4Cache<V>::Put(const V& val) const
 {
   GetCache() = val;
 }
 
 // Should here remove from cache element?
 template <class V>
 V G4Cache<V>::Pop()
 {
   return GetCache();
 }
 
 template <class V>
 std::atomic<unsigned int> G4Cache<V>::instancesctr(0);
 
 template <class V>
 std::atomic<unsigned int> G4Cache<V>::dstrctr(0);
 
 //========== Implementation: G4VectorCache<V> ===========================
 
 template <class V>
 G4VectorCache<V>::G4VectorCache() = default;
 
 template <class V>
 G4VectorCache<V>::~G4VectorCache()
 {
 #ifdef g4cdebug
   std::cout << "~G4VectorCache "
             << G4Cache<G4VectorCache<V>::vector_type>::GetId()
             << " with size: " << Size() << "->";
   for(size_type i = 0; i < Size(); ++i)
     std::cout << operator[](i) << ",";
   std::cout << "<-" << std::endl;
 #endif
 }
 
 template <class V>
 G4VectorCache<V>::G4VectorCache(G4int nElems)
 {
   vector_type& cc = G4Cache<vector_type>::Get();
   cc.resize(nElems);
 }
 
 template <class V>
 G4VectorCache<V>::G4VectorCache(G4int nElems, V* vals)
 {
   vector_type& cc = G4Cache<vector_type>::Get();
   cc.resize(nElems);
   for(G4int idx = 0; idx < nElems; ++idx)
     cc[idx] = vals[idx];
 }
 
 template <class V>
 void G4VectorCache<V>::Push_back(const V& val)
 {
   G4Cache<vector_type>::Get().push_back(val);
 }
 
 template <class V>
 V G4VectorCache<V>::Pop_back()
 {
   vector_type& cc = G4Cache<vector_type>::Get();
   value_type val  = cc[cc.size() - 1];
   cc.pop_back();
   return val;
 }
 
 template <class V>
 V& G4VectorCache<V>::operator[](const G4int& idx)
 {
   vector_type& cc = G4Cache<vector_type>::Get();
   return cc[idx];
 }
 
 template <class V>
 typename G4VectorCache<V>::iterator G4VectorCache<V>::Begin()
 {
   return G4Cache<vector_type>::Get().begin();
 }
 
 template <class V>
 typename G4VectorCache<V>::iterator G4VectorCache<V>::End()
 {
   return G4Cache<vector_type>::Get().end();
 }
 
 template <class V>
 void G4VectorCache<V>::Clear()
 {
   G4Cache<vector_type>::Get().clear();
 }
 
 // template<class V>
 // typename G4VectorCache<V>::size_type G4VectorCache<V>::Size()
 //{
 //    return G4Cache<vector_type>::Get().size();
 //}
 
 //======== Implementation: G4MapType<K,V> ===========================
 
 template <class K, class V>
 G4MapCache<K, V>::~G4MapCache()
 {
 #ifdef g4cdebug
   std::cout << "~G4MacCache " << G4Cache<map_type>::GetId()
             << " with size: " << Size() << "->";
   for(iterator it = Begin(); it != End(); ++it)
     std::cout << it->first << ":" << it->second << ",";
   std::cout << "<-" << std::endl;
 #endif
 }
 
 template <class K, class V>
 std::pair<typename G4MapCache<K, V>::iterator, G4bool> G4MapCache<K, V>::Insert(
   const K& k, const V& v)
 {
   return G4Cache<map_type>::Get().insert(std::pair<key_type, value_type>(k, v));
 }
 
 // template<class K, class V>
 // typename G4MapCache<K,V>::size_type G4MapCache<K,V>::Size()
 //{
 //    return G4Cache<map_type>::Get().size();
 //}
 
 template <class K, class V>
 typename G4MapCache<K, V>::iterator G4MapCache<K, V>::Begin()
 {
   return G4Cache<map_type>::Get().begin();
 }
 template <class K, class V>
 typename G4MapCache<K, V>::iterator G4MapCache<K, V>::End()
 {
   return G4Cache<map_type>::Get().end();
 }
 
 template <class K, class V>
 typename G4MapCache<K, V>::iterator G4MapCache<K, V>::Find(const K& k)
 {
   return G4Cache<map_type>::Get().find(k);
 }
 
 template <class K, class V>
 G4bool G4MapCache<K, V>::Has(const K& k)
 {
   return (Find(k) != End());
 }
 
 template <class K, class V>
 V& G4MapCache<K, V>::Get(const K& k)
 {
   return Find(k)->second;
 }
 
 template <class K, class V>
 typename G4MapCache<K, V>::size_type G4MapCache<K, V>::Erase(const K& k)
 {
   return G4Cache<map_type>::Get().erase(k);
 }
 
 template <class K, class V>
 V& G4MapCache<K, V>::operator[](const K& k)
 {
   return (G4Cache<map_type>::Get())[k];
 }
 
 #endif

This page was automatically generated by the 2.3.7 LXR engine. The LXR team