Related
I read that there is minor performance hit when calling virtual functions from derived classes if called repeatedly. One thing that I am not clear about is whether this affects function calls from the base class itself. If I call a method in the base class with the virtual keyword, does this performance hit still occur?
If I call a method in the base class with the virtual keyword, does this performance hit still occur?
That the virtual function is being called from the base class will not prevent virtual lookup.
Consider this trivial example:
class Base
{
public:
virtual get_int() { return 1; }
void print_int()
{
// Calling a virtual function from the base class
std::cout << get_int();
}
};
class Derived : public Base
{
public:
virtual get_int() { return 2; }
};
int main()
{
Base().print_int();
Derived().print_int();
}
Is print_int() guaranteed to print 1? It is not.
That the function print_int() exists in the base class does not prove that the object its called from is not derived.
Yes, there will be a performance overhead.
This is due to the fact that virtual functions in an inheritance hierarchy may or may not be overloaded by any derived class.
This requires a lookup in a v-table, as the base class doesn't know any better as to which class is dynamically implementing the function.
Edit: As mentioned, there may be some optimization, but this shouldn't be relied on
Virtual functions are implemented by a virtual function table. Each class has a table of the virtual functions' addresses. An instance of a class with a virtual function table has a pointer to the table, which is set by the constructor.
When the code calls a regular function, its address is hard coded into it. When it calls a virtual function it has to calculate *(*vtbl + 8 * function-offset), which requires two memory accesses. That's the overhead, which can be avoided in cases mentioned by others above.
Point is, if the same function is called repeatedly, much of the overhead might be avoided. The first call would bring the virtual function table from RAM to the CPU's cache, meaning it would be as cheap as 1-2 CPU cycles to fetch again. Doing an integer shift and an addition is rather cheap. If the compiler knows its the same function of the same class, it could calculate it once and reuse the value.
Since the question is "Is there a performance hit?" and not "Can there be a performance hit?", it is surprisingly tricky to answer accurately. This is because compilers are given a lot of leeway when it comes time to optimize your code, and they often make use of it. The process of eliminating that specific overhead has a particular name: devirtualization.
Because of this, whether a cost is incurred or not will depend on:
Which compiler is being used
The compiler version
The compiler settings
The linker settings
How the class is used
Whether there are any subclasses that override the method.
So what should you do with so much uncertainty? Since the overhead is minor in the first place, the first thing is to not worry about it unless you need to improve performance in that specific area. Writing structurally sound and maintainable code trumps premature optimisation every single time.
But if you have to worry about it, then a good rule of thumb is that if you are calling a non-final virtual function from a pointer or reference (which includes this) to a non-final class, then you should write code with the assumption that the tiny overhead associated with an indirect lookup through a vtable will be
paid.
That doesn't mean that the overhead will necessarily occur, just that there is a non-zero chance that it will be there.
So in your scenario, given:
class Base {
public:
virtual void foo() { std::cout << "foo is called\n"; }
void bar() { foo(); }
};
int main() {
Base b;
b.bar();
}
Base is not final
Base::foo is not final.
this is a pointer.
So you should operate under the assumption that the overhead might be present, regardless of whether or not it ends up being there in the final built application.
If several threads share a pointer to the same object, is it safe to call its virtual method? Of course we should assume that the method itself is thread-safe.
class Base {
public:
virtual int test(int arg) = 0;
};
class Derived1: public Base {
public:
virtual int test(int arg) { return arg * 2; }
};
class Derived2: public Base {
public:
virtual int test(int arg) { return arg * 3; }
};
//call in threads:
Base* base = (pointer to the same Derived1);
int value = base->test(1);
Yes, that's fine, provided the lifetime of the object *base exceeds that of the function call.
Calling virtual functions is normally implemented by looking up a vtable entry in the class. This vtable doesn't change, so for this implementation, it should be thread safe.
I don't think there is a general guarantee.
I've started this answer about three times, based on "Kerrek SB's comment on a previous answer. As I see it, without reading specs 'til my eyes bleed, I'm uncertain as to what shouldn't be threadsafe of a virtual call - the implementation of the "virtual function selection" (vtable or whatever it may be) should definitely not cause any problem, since it should just be a pointer to the function, which gets called based on the index [or similar] of the function chosen.
I'm sorry, this isn't an answer, but I have a very hard time seeing any scenario, on any processor that I know how it works (x86, 68K, 29k, ARM are ones that I've worked enough with to understand how a virtual function is implemented in assembler) where this would go wrong because of threads - assuming the other code is safe - if in the above example we were to run a second thread that modifies which element base points at, you could potentially have some sort of race-condition, where the value of base is pointing at the wrong set of virtual functions, or some such. But that's not the call itself, but the code modifying base that is 'not threadsafe'.
Of course, there may some "amateur" compiler around that solves virtual functions in some other way.
Of course, there wouldn't be a sensible workaround, should there be a problem - unless you block other threads for the entire duration of the virtual call - and if you have a class that implements threads by having a virtual function run(void *args) as the "this is what to run inside the thread", which is something I've seen a several times, that would pretty much kill that functionality completely!
So, basically, although I'm not able to refer to a spec section that says this is safe, I can't see any solution other than "it has to be".
If the method is threadsafe, then it is fine.
In short:
If not, C++ would not be useable at all for multithreaded programming.
In long:
After compiling the programm is constant. So its thread safe.
While loading (of modules) the runtime system changes data structures for
RTTI (dynamic_cast, ...). This is out of your scope but it should be implemented in a thread safe way.
After construction the type of your object does not change (you can with very dirty tricks). So all virtual functions of all your objects should not change. So its thread safe.
But you should consider that member of classes which can be seen as an replacment for virtual functions (boost function, loki functor, ...) may have value semantics and can change while called. In this case it should be documented or better interface for using them should be implemented.
In my opinion you can safely call a virtual method. Even in the case somebody use tries to mimic virtual functions you can expect normal (safe) behavior.
I have some trouble. While planning my program, I consider 2 versions:
Use static methods in classes with additional parameter (pointer to copy of a class).
Use virtual methods (vtable)
What is faster? Why?
Edit: I need to make next algorithm: array should be storing pointers to method of different classes (they meet diff. game objects), for example, method Draw().
The main task is storing and calling methods of different classes.
At this point, you probably shouldn't be thinking about micro-optimisations at all - focus on choosing efficient algorithms, and making your code clear and correct; then identify any bottlenecks that prevent it from performing as required. Having said that, here are some thoughts in the unlikely event that you do find that virtual dispatch becomes an issue.
The two are not equivalent - the first (if I understand what you're saying) is trying to emulate a non-virtual function by explicitly passing a this-like pointer to a static function, and is likely to be exactly as fast as the non-static equivalent. It will behave differently to a virtual function, and so can't be used if you need virtual dispatch.
A non-virtual function will (almost certainly) be a bit faster than a virtual function - it's more likely to be inlined, and if not inlined it can be called directly rather than looked up at runtime. Therefore, only declare functions virtual when you need virtual dispatch.
In extreme circumstances, you might be able to save a level of indirection over virtual functions, by storing function pointers in the object rather than using the compiler-generated virtual dispatch. You should only do this as a last resort, if you find that virtual dispatch is a serious bottle-neck, and you can't sensibly redesign your algorithm to avoid it.
First, virtual functions and what you are proposing have different
semantics. If you need different behavior because you have different
types of objects, then it's highly unlikely that you can do better than
the compilers implementation of virtual functions. If you don't need
it, then just don't declare the function virtual.
And don't worry about the performance issue until you know it is one.
If, after getting the code to work, you do find performance issues due
to virtual function calls (usually because the compiler can't inline the
function, so you loose all of the optimizations which would follow
inlining), you can, punctually, avoid the virtual function cost if you
design the class correctly. Supposing the function in question is
f():
class Base
{
virtual void doF() = 0;
public:
void f() { doF(); }
};
class Derived : public Base
{
virtual void doF() { f(); }
public:
void f() { /* ... */ }
};
If you do this, and you have, for example, a tight loop where you're
constantly calling f() on the same object, you can do something like:
void
tightLoop( Base& object )
{
Derived& o = dynamic_cast<Derived&>( object );
for ( /* ... */ ) {
o.f();
}
}
If you do this, of course, tightLoop can only be called with an object
which is actually Derived.
Why does C++ not have a virtual constructor?
Hear it from the horse's mouth. :)
From Bjarne Stroustrup's C++ Style and Technique FAQ Why don't we have virtual constructors?
A virtual call is a mechanism to get work done given partial
information. In particular, "virtual" allows us to call a function
knowing only any interfaces and not the exact type of the object. To
create an object you need complete information. In particular, you
need to know the exact type of what you want to create. Consequently,
a "call to a constructor" cannot be virtual.
The FAQ entry goes on to give the code for a way to achieve this end without a virtual constructor.
Virtual functions basically provide polymorphic behavior. That is, when you work with an object whose dynamic type is different than the static (compile time) type with which it is referred to, it provides behavior that is appropriate for the actual type of object instead of the static type of the object.
Now try to apply that sort of behavior to a constructor. When you construct an object the static type is always the same as the actual object type since:
To construct an object, a constructor needs the exact type of the object it is to create [...] Furthermore [...]you cannot have a pointer to a constructor
(Bjarne Stroustup (P424 The C++ Programming Language SE))
Unlike object oriented languages such as Smalltalk or Python, where the constructor is a virtual method of the object representing the class (which means you don't need the GoF abstract factory pattern, as you can pass the object representing the class around instead of making your own), C++ is a class based language, and does not have objects representing any of the language's constructs. The class does not exist as an object at runtime, so you can't call a virtual method on it.
This fits with the 'you don't pay for what you don't use' philosophy, though every large C++ project I've seen has ended up implementing some form of abstract factory or reflection.
two reasons I can think of:
Technical reason
The object exists only after the constructor ends.In order for the constructor to be dispatched using the virtual table , there has to be an existing object with a pointer to the virtual table , but how can a pointer to the virtual table exist if the object still doesn't exist? :)
Logic reason
You use the virtual keyword when you want to declare a somewhat polymorphic behaviour. But there is nothing polymorphic with constructors , constructors job in C++ is to simply put an object data on the memory . Since virtual tables (and polymorphism in general) are all about polymorphic behaviour rather on polymorphic data , There is no sense with declaring a virtual constructor.
Summary: the C++ Standard could specify a notation and behaviour for "virtual constructor"s that's reasonably intuitive and not too hard for compilers to support, but why make a Standard change for this specifically when the functionality can already be cleanly implemented using create() / clone() (see below)? It's not nearly as useful as many other language proposal in the pipeline.
Discussion
Let's postulate a "virtual constructor" mechanism:
Base* p = new Derived(...);
Base* p2 = new p->Base(); // possible syntax???
In the above, the first line constructs a Derived object, so *p's virtual dispatch table can reasonably supply a "virtual constructor" for use in the second line. (Dozens of answers on this page stating "the object doesn't yet exist so virtual construction is impossible" are unnecessarily myopically focused on the to-be-constructed object.)
The second line postulates the notation new p->Base() to request dynamic allocation and default construction of another Derived object.
Notes:
the compiler must orchestrate memory allocation before calling the constructor - constructors normally support automatic (informally "stack") allocation, static (for global/namespace scope and class-/function-static objects), and dynamic (informally "heap") when new is used
the size of object to be constructed by p->Base() can't generally be known at compile-time, so dynamic allocation is the only approach that makes sense
it is possible to allocate runtime-specified amounts of memory on the stack - e.g. GCC's variable-length array extension, alloca() - but leads to significant inefficiencies and complexities (e.g. here and here respectively)
for dynamic allocation it must return a pointer so memory can be deleted later.
the postulated notation explicitly lists new to emphasise dynamic allocation and the pointer result type.
The compiler would need to:
find out how much memory Derived needed, either by calling an implicit virtual sizeof function or having such information available via RTTI
call operator new(size_t) to allocate memory
invoke Derived() with placement new.
OR
create an extra vtable entry for a function that combines dynamic allocation and construction
So - it doesn't seem insurmountable to specify and implement virtual constructors, but the million-dollar question is: how would it be better than what's possible using existing C++ language features...? Personally, I see no benefit over the solution below.
`clone()` and `create()`
The C++ FAQ documents a "virtual constructor" idiom, containing virtual create() and clone() methods to default-construct or copy-construct a new dynamically-allocated object:
class Shape {
public:
virtual ~Shape() { } // A virtual destructor
virtual void draw() = 0; // A pure virtual function
virtual void move() = 0;
// ...
virtual Shape* clone() const = 0; // Uses the copy constructor
virtual Shape* create() const = 0; // Uses the default constructor
};
class Circle : public Shape {
public:
Circle* clone() const; // Covariant Return Types; see below
Circle* create() const; // Covariant Return Types; see below
// ...
};
Circle* Circle::clone() const { return new Circle(*this); }
Circle* Circle::create() const { return new Circle(); }
It's also possible to change or overload create() to accept arguments, though to match the base class / interface's virtual function signature, arguments to overrides must exactly match one of the base class overloads. With these explicit user-provided facilities, it's easy to add logging, instrumentation, alter memory allocation etc..
We do, it's just not a constructor :-)
struct A {
virtual ~A() {}
virtual A * Clone() { return new A; }
};
struct B : public A {
virtual A * Clone() { return new B; }
};
int main() {
A * a1 = new B;
A * a2 = a1->Clone(); // virtual construction
delete a2;
delete a1;
}
Semantic reasons aside, there is no vtable until after the object is constructed, thus making a virtual designation useless.
Virtual functions in C++ are an implementation of run-time polymorphism, and they will do function overriding. Generally the virtual keyword is used in C++ when you need dynamic behavior. It will work only when object exists. Whereas constructors are used to create the objects. Constructors will be called at the time of object creation.
So if you create the constructor as virtual, as per the virtual keyword definition, it should have existing object to use, but constructor is used to to create the object, so this case will never exist. So you should not use the constructor as virtual.
So, if we try to declare virtual constructor compiler throw an Error:
Constructors cannot be declared virtual
You can find an example and the technical reason to why it is not allowed in #stefan 's answer. Now a logical answer to this question according to me is:
The major use of virtual keyword is to enable polymorphic behaviour when we don't know what type of the object the base class pointer will point to.
But think of this is more primitive way, for using virtual functionality you will require a pointer. And what does a pointer require? An object to point to! (considering case for correct execution of the program)
So, we basically require an object that already exists somewhere in the memory (we are not concerned with how the memory was allocated, it may be at compile time or either runtime) so that our pointer can correctly point to that object.
Now, think of the situation about the moment when the object of the class to be pointed is being assigned some memory -> Its constructor will be called automatically at that instance itself!
So we can see that we don't actually need to worry about the constructor being virtual, because in any of the cases you wish to use a polymorphic behaviour our constructor would have already been executed making our object ready for usage!
When people ask a question like this, I like to think to myself "what would happen if this were actually possible?" I don't really know what this would mean, but I guess it would have something to do with being able to override the constructor implementation based on the dynamic type of the object being created.
I see a number of potential problems with this. For one thing, the derived class will not be fully constructed at the time the virtual constructor is called, so there are potential issues with the implementation.
Secondly, what would happen in the case of multiple inheritance? Your virtual constructor would be called multiple times presumably, you would then need to have some way of know which one was being called.
Thirdly, generally speaking at the time of construction, the object does not have the virtual table fully constructed, this means it would require a large change to the language specification to allow for the fact that the dynamic type of the object would be known at construction time. This would then allow the base class constructor to maybe call other virtual functions at construction time, with a not fully constructed dynamic class type.
Finally, as someone else has pointed out you can implement a kind of virtual constructor using static "create" or "init" type functions that basically do the same thing as a virtual constructor would do.
Although the concept of virtual constructors does not fit in well since object type is pre-requisite for object creation, its not completly over-ruled.
GOF's 'factory method' design pattern makes use of the 'concept' of virtual constructor, which is handly in certain design situations.
Virtual functions are used in order to invoke functions based on the type of object pointed to by the pointer, and not the type of pointer itself. But a constructor is not "invoked". It is called only once when an object is declared. So, a constructor cannot be made virtual in C++.
Interview answer is : virtual ptr and table are related to objects but not the class.hence constructor builds the virtual table
hence we cant have virtual constructor as there is no Vtable before obj creation.
You shouldn't call virtual function within your constructor either. See : http://www.artima.com/cppsource/nevercall.html
In addition I'm not sure that you really need a virtual constructor. You can achieve polymorphic construction without it: you can write a function that will construct your object according to the needed parameters.
A virtual-table(vtable) is made for each Class having one or more 'virtual-functions'. Whenever an Object is created of such class, it contains a 'virtual-pointer' which points to the base of corresponding vtable. Whenever there is a virtual function call, the vtable is used to resolve to the function address.
Constructor can not be virtual, because when constructor of a class is executed there is no vtable in the memory, means no virtual pointer defined yet. Hence the constructor should always be non-virtual.
Cant we simply say it like.. We cannot inherit constructors. So there is no point declaring them virtual because the virtual provides polymorphism .
The virtual mechanism only works when you have a based class pointer to a derived class object. Construction has it's own rules for the calling of base class constructors, basically base class to derived. How could a virtual constructor be useful or called? I don't know what other languages do, but I can't see how a virtual constructor could be useful or even implemented. Construction needs to have taken place for the virtual mechanism to make any sense and construction also needs to have taken place for the vtable structures to have been created which provides the mechanics of the polymorphic behaviour.
C++ virtual constructor is not possible.For example you can not mark a constructor as virtual.Try this code
#include<iostream.h>
using namespace std;
class aClass
{
public:
virtual aClass()
{
}
};
int main()
{
aClass a;
}
It causes an error.This code is trying to declare a constructor as virtual.
Now let us try to understand why we use virtual keyword. Virtual keyword is used to provide run time polymorphism. For example try this code.
#include<iostream.h>
using namespace std;
class aClass
{
public:
aClass()
{
cout<<"aClass contructor\n";
}
~aClass()
{
cout<<"aClass destructor\n";
}
};
class anotherClass:public aClass
{
public:
anotherClass()
{
cout<<"anotherClass Constructor\n";
}
~anotherClass()
{
cout<<"anotherClass destructor\n";
}
};
int main()
{
aClass* a;
a=new anotherClass;
delete a;
getchar();
}
In main a=new anotherClass; allocates a memory for anotherClass in a pointer a declared as type of aClass.This causes both the constructor (In aClass and anotherClass) to call automatically.So we do not need to mark constructor as virtual.Because when an object is created it must follow the chain of creation (i.e first the base and then the derived classes).
But when we try to delete a delete a; it causes to call only the base destructor.So we have to handle the destructor using virtual keyword. So virtual constructor is not possible but virtual destructor is.Thanks
There's a very basic reason: Constructors are effectively static functions, and in C++ no static function can be virtual.
If you have much experience with C++, you know all about the difference between static & member functions. Static functions are associated with the CLASS, not the objects (instances), so they don't see a "this" pointer. Only member functions can be virtual, because the vtable- the hidden table of function pointers that makes 'virtual' work- is really a data member of each object.
Now, what is the constructor's job? It is in the name- a "T" constructor initializes T objects as they're allocated. This automatically precludes it being a member function! An object has to EXIST before it has a "this" pointer and thus a vtable. That means that even if the language treated constructors as ordinary functions (it doesn't, for related reasons I won't get into), they'd have to be static member functions.
A great way to see this is to look at the "Factory" pattern, especially factory functions. They do what you're after, and you'll notice that if class T has a factory method, it is ALWAYS STATIC. It has to be.
If you think logically about how constructors work and what the meaning/usage of a virtual function is in C++ then you will realise that a virtual constructor would be meaningless in C++. Declaring something virtual in C++ means that it can be overridden by a sub-class of the current class, however the constructor is called when the objected is created, at that time you cannot be creating a sub-class of the class, you must be creating the class so there would never be any need to declare a constructor virtual.
And another reason is, the constructors have the same name as its class name and if we declare constructor as virtual, then it should be redefined in its derived class with the same name, but you can not have the same name of two classes. So it is not possible to have a virtual constructor.
When a constructor is invoked, although there is no object created till that point, we still know the kind of object that is gonna be created because the specific constructor of the class to which the object belongs to has already been called.
Virtual keyword associated with a function means the function of a particular object type is gonna be called.
So, my thinking says that there is no need to make the virtual constructor because already the desired constructor whose object is gonna be created has been invoked and making constructor virtual is just a redundant thing to do because the object-specific constructor has already been invoked and this is same as calling class-specific function which is achieved through the virtual keyword.
Although the inner implementation won’t allow virtual constructor for vptr and vtable related reasons.
Another reason is that C++ is a statically typed language and we need to know the type of a variable at compile-time.
The compiler must be aware of the class type to create the object. The type of object to be created is a compile-time decision.
If we make the constructor virtual then it means that we don’t need to know the type of the object at compile-time(that’s what virtual function provide. We don’t need to know the actual object and just need the base pointer to point an actual object call the pointed object’s virtual functions without knowing the type of the object) and if we don’t know the type of the object at compile time then it is against the statically typed languages. And hence, run-time polymorphism cannot be achieved.
Hence, Constructor won’t be called without knowing the type of the object at compile-time. And so the idea of making a virtual constructor fails.
"A constructor can not be virtual"
there are some valid reasons that justify this statement.
to create an object the constructor of the object class must be of the same type as the class. But, this is not possible with a virtually implemented constructor.
at the time of calling the constructor, the virtual table would not have been created to resolve any virtual function calls. Thus, a virtual constructor itself would not have anywhere to look up to.
As a result, it is not possible to declare a constructor to be virtual.
The Vpointer is created at the time of object creation. vpointer wont exists before object creation. so there is no point of making the constructor as virtual.
Is there ever a good reason to not declare a virtual destructor for a class? When should you specifically avoid writing one?
There is no need to use a virtual destructor when any of the below is true:
No intention to derive classes from it
No instantiation on the heap
No intention to store with access via a pointer to a superclass
No specific reason to avoid it unless you are really so pressed for memory.
To answer the question explicitly, i.e. when should you not declare a virtual destructor.
C++ '98/'03
Adding a virtual destructor might change your class from being POD (plain old data)* or aggregate to non-POD. This can stop your project from compiling if your class type is aggregate initialized somewhere.
struct A {
// virtual ~A ();
int i;
int j;
};
void foo () {
A a = { 0, 1 }; // Will fail if virtual dtor declared
}
In an extreme case, such a change can also cause undefined behaviour where the class is being used in a way that requires a POD, e.g. passing it via an ellipsis parameter, or using it with memcpy.
void bar (...);
void foo (A & a) {
bar (a); // Undefined behavior if virtual dtor declared
}
[* A POD type is a type that has specific guarantees about its memory layout. The standard really only says that if you were to copy from an object with POD type into an array of chars (or unsigned chars) and back again, then the result will be the same as the original object.]
Modern C++
In recent versions of C++, the concept of POD was split between the class layout and its construction, copying and destruction.
For the ellipsis case, it is no longer undefined behavior it is now conditionally-supported with implementation-defined semantics (N3937 - ~C++ '14 - 5.2.2/7):
...Passing a potentially-evaluated argument of class type (Clause 9) having a non-trivial copy constructor, a non-trivial move constructor, or a on-trivial destructor, with no corresponding parameter, is conditionally-supported with implementation-defined semantics.
Declaring a destructor other than =default will mean it's not trivial (12.4/5)
... A destructor is trivial if it is not user-provided ...
Other changes to Modern C++ reduce the impact of the aggregate initialization problem as a constructor can be added:
struct A {
A(int i, int j);
virtual ~A ();
int i;
int j;
};
void foo () {
A a = { 0, 1 }; // OK
}
I declare a virtual destructor if and only if I have virtual methods. Once I have virtual methods, I don't trust myself to avoid instantiating it on the heap or storing a pointer to the base class. Both of these are extremely common operations and will often leak resources silently if the destructor is not declared virtual.
A virtual destructor is needed whenever there is any chance that delete might be called on a pointer to an object of a subclass with the type of your class. This makes sure the correct destructor gets called at run time without the compiler having to know the class of an object on the heap at compile time. For example, assume B is a subclass of A:
A *x = new B;
delete x; // ~B() called, even though x has type A*
If your code is not performance critical, it would be reasonable to add a virtual destructor to every base class you write, just for safety.
However, if you found yourself deleteing a lot of objects in a tight loop, the performance overhead of calling a virtual function (even one that's empty) might be noticeable. The compiler cannot usually inline these calls, and the processor might have a difficult time predicting where to go. It is unlikely this would have a significant impact on performance, but it's worth mentioning.
Virtual functions mean every allocated object increases in memory cost by a virtual function table pointer.
So if your program involves allocating a very large number of some object, it would be worth avoiding all virtual functions in order to save the additional 32 bits per object.
In all other cases, you will save yourself debug misery to make the dtor virtual.
Not all C++ classes are suitable for use as a base class with dynamic polymorphism.
If you want your class to be suitable for dynamic polymorphism, then its destructor must be virtual. In addition, any methods which a subclass could conceivably want to override (which might mean all public methods, plus potentially some protected ones used internally) must be virtual.
If your class is not suitable for dynamic polymorphism, then the destructor should not be marked virtual, because to do so is misleading. It just encourages people to use your class incorrectly.
Here's an example of a class which would not be suitable for dynamic polymorphism, even if its destructor were virtual:
class MutexLock {
mutex *mtx_;
public:
explicit MutexLock(mutex *mtx) : mtx_(mtx) { mtx_->lock(); }
~MutexLock() { mtx_->unlock(); }
private:
MutexLock(const MutexLock &rhs);
MutexLock &operator=(const MutexLock &rhs);
};
The whole point of this class is to sit on the stack for RAII. If you're passing around pointers to objects of this class, let alone subclasses of it, then you're Doing It Wrong.
A good reason for not declaring a destructor as virtual is when this saves your class from having a virtual function table added, and you should avoid that whenever possible.
I know that many people prefer to just always declare destructors as virtual, just to be on the safe side. But if your class does not have any other virtual functions then there is really, really no point in having a virtual destructor. Even if you give your class to other people who then derive other classes from it then they would have no reason to ever call delete on a pointer that was upcast to your class - and if they do then I would consider this a bug.
Okay, there is one single exception, namely if your class is (mis-)used to perform polymorphic deletion of derived objects, but then you - or the other guys - hopefully know that this requires a virtual destructor.
Put another way, if your class has a non-virtual destructor then this is a very clear statement: "Don't use me for deleting derived objects!"
If you have a very small class with a huge number of instances, the overhead of a vtable pointer can make a difference in your program's memory usage. As long as your class doesn't have any other virtual methods, making the destructor non-virtual will save that overhead.
I usually declare the destructor virtual, but if you have performance critical code that is used in an inner loop, you might want to avoid the virtual table lookup. That can be important in some cases, like collision checking. But be careful about how you destroy those objects if you use inheritance, or you will destroy only half of the object.
Note that the virtual table lookup happens for an object if any method on that object is virtual. So no point in removing the virtual specification on a destructor if you have other virtual methods in the class.
If you absolutely positively must ensure that your class does not have a vtable then you must not have a virtual destructor as well.
This is a rare case, but it does happen.
The most familiar example of a pattern that does this are the DirectX D3DVECTOR and D3DMATRIX classes. These are class methods instead of functions for the syntactic sugar, but the classes intentionally do not have a vtable in order to avoid the function overhead because these classes are specifically used in the inner loop of many high-performance applications.
On operation that will be performed on the base class, and that should behave virtually, should be virtual. If deletion can be performed polymorphically through the base class interface, then it must behave virtually and be virtual.
The destructor has no need to be virtual if you don't intend to derive from the class. And even if you do, a protected non-virtual destructor is just as good if deletion of base class pointers isn't required.
The performance answer is the only one I know of which stands a chance of being true. If you've measured and found that de-virtualizing your destructors really speeds things up, then you've probably got other things in that class that need speeding up too, but at this point there are more important considerations. Some day someone is going to discover that your code would provide a nice base class for them and save them a week's work. You'd better make sure they do that week's work, copying and pasting your code, instead of using your code as a base. You'd better make sure you make some of your important methods private so that no one can ever inherit from you.