Develop Reference

Avoid garbage collection in high performance JavaScript applications [duplicate]

I have a fairly complex Javascript app, which has a main loop that is called 60 times per second. There seems to be a lot of garbage collection going on (based on the 'sawtooth' output from the Memory timeline in the Chrome dev tools) - and this often impacts the performance of the application.
So, I'm trying to research best practices for reducing the amount of work that the garbage collector has to do. (Most of the information I've been able to find on the web regards avoiding memory leaks, which is a slightly different question - my memory is getting freed up, it's just that there's too much garbage collection going on.) I'm assuming that this mostly comes down to reusing objects as much as possible, but of course the devil is in the details.
The app is structured in 'classes' along the lines of John Resig's Simple JavaScript Inheritance.
I think one issue is that some functions can be called thousands of times per second (as they are used hundreds of times during each iteration of the main loop), and perhaps the local working variables in these functions (strings, arrays, etc.) might be the issue.
I'm aware of object pooling for larger/heavier objects (and we use this to a degree), but I'm looking for techniques that can be applied across the board, especially relating to functions that are called very many times in tight loops.
What techniques can I use to reduce the amount of work that the garbage collector must do?
And, perhaps also - what techniques can be employed to identify which objects are being garbage collected the most? (It's a farly large codebase, so comparing snapshots of the heap has not been very fruitful)

A lot of the things you need to do to minimize GC churn go against what is considered idiomatic JS in most other scenarios, so please keep in mind the context when judging the advice I give.
Allocation happens in modern interpreters in several places:
When you create an object via new or via literal syntax [...], or {}.
When you concatenate strings.
When you enter a scope that contains function declarations.
When you perform an action that triggers an exception.
When you evaluate a function expression: (function (...) { ... }).
When you perform an operation that coerces to Object like Object(myNumber) or Number.prototype.toString.call(42)
When you call a builtin that does any of these under the hood, like Array.prototype.slice.
When you use arguments to reflect over the parameter list.
When you split a string or match with a regular expression.
Avoid doing those, and pool and reuse objects where possible.
Specifically, look out for opportunities to:
Pull inner functions that have no or few dependencies on closed-over state out into a higher, longer-lived scope. (Some code minifiers like Closure compiler can inline inner functions and might improve your GC performance.)
Avoid using strings to represent structured data or for dynamic addressing. Especially avoid repeatedly parsing using split or regular expression matches since each requires multiple object allocations. This frequently happens with keys into lookup tables and dynamic DOM node IDs. For example, lookupTable['foo-' + x] and document.getElementById('foo-' + x) both involve an allocation since there is a string concatenation. Often you can attach keys to long-lived objects instead of re-concatenating. Depending on the browsers you need to support, you might be able to use Map to use objects as keys directly.
Avoid catching exceptions on normal code-paths. Instead of try { op(x) } catch (e) { ... }, do if (!opCouldFailOn(x)) { op(x); } else { ... }.
When you can't avoid creating strings, e.g. to pass a message to a server, use a builtin like JSON.stringify which uses an internal native buffer to accumulate content instead of allocating multiple objects.
Avoid using callbacks for high-frequency events, and where you can, pass as a callback a long-lived function (see 1) that recreates state from the message content.
Avoid using arguments since functions that use that have to create an array-like object when called.
I suggested using JSON.stringify to create outgoing network messages. Parsing input messages using JSON.parse obviously involves allocation, and lots of it for large messages. If you can represent your incoming messages as arrays of primitives, then you can save a lot of allocations. The only other builtin around which you can build a parser that does not allocate is String.prototype.charCodeAt. A parser for a complex format that only uses that is going to be hellish to read though.

The Chrome developer tools have a very nice feature for tracing memory allocation. It's called the Memory Timeline. This article describes some details. I suppose this is what you're talking about re the "sawtooth"? This is normal behavior for most GC'ed runtimes. Allocation proceeds until a usage threshold is reached triggering a collection. Normally there are different kinds of collections at different thresholds.
Garbage collections are included in the event list associated with the trace along with their duration. On my rather old notebook, ephemeral collections are occurring at about 4Mb and take 30ms. This is 2 of your 60Hz loop iterations. If this is an animation, 30ms collections are probably causing stutter. You should start here to see what's going on in your environment: where the collection threshold is and how long your collections are taking. This gives you a reference point to assess optimizations. But you probably won't do better than to decrease the frequency of the stutter by slowing the allocation rate, lengthening the interval between collections.
The next step is to use the Profiles | Record Heap Allocations feature to generate a catalog of allocations by record type. This will quickly show which object types are consuming the most memory during the trace period, which is equivalent to allocation rate. Focus on these in descending order of rate.
The techniques are not rocket science. Avoid boxed objects when you can do with an unboxed one. Use global variables to hold and reuse single boxed objects rather than allocating fresh ones in each iteration. Pool common object types in free lists rather than abandoning them. Cache string concatenation results that are likely reusable in future iterations. Avoid allocation just to return function results by setting variables in an enclosing scope instead. You will have to consider each object type in its own context to find the best strategy. If you need help with specifics, post an edit describing details of the challenge you're looking at.
I advise against perverting your normal coding style throughout an application in a shotgun attempt to produce less garbage. This is for the same reason you should not optimize for speed prematurely. Most of your effort plus much of the added complexity and obscurity of code will be meaningless.

As a general principle you'd want to cache as much as possible and do as little creating and destroying for each run of your loop.
The first thing that pops in my head is to reduce the use of anonymous functions (if you have any) inside your main loop. Also it'd be easy to fall into the trap of creating and destroying objects that are passed into other functions. I'm by no means a javascript expert, but I would imagine that this:
var options = {var1: value1, var2: value2, ChangingVariable: value3};
function loopfunc()
{
//do something
}
while(true)
{
$.each(listofthings, loopfunc);
options.ChangingVariable = newvalue;
someOtherFunction(options);
}
would run much faster than this:
while(true)
{
$.each(listofthings, function(){
//do something on the list
});
someOtherFunction({
var1: value1,
var2: value2,
ChangingVariable: newvalue
});
}
Is there ever any downtime for your program? Maybe you need it to run smoothly for a second or two (e.g. for an animation) and then it has more time to process? If this is the case I could see taking objects that would normally be garbage collected throughout the animation and keeping a reference to them in some global object. Then when the animation ends you can clear all the references and let the garbage collector do it's work.
Sorry if this is all a bit trivial compared to what you've already tried and thought of.

I'd make one or few objects in the global scope (where I'm sure garbage collector is not allowed to touch them), then I'd try to refactor my solution to use those objects to get the job done, instead of using local variables.
Of course it couldn't be done everywhere in the code, but generally that's my way to avoid garbage collector.
P.S. It might make that specific part of code a little bit less maintainable.

Is having a nested function inside a constructor memory inefficient? [duplicate]

In JavaScript, we have two ways of making a "class" and giving it public functions.
Method 1:
function MyClass() {
var privateInstanceVariable = 'foo';
this.myFunc = function() { alert(privateInstanceVariable ); }
}
Method 2:
function MyClass() { }
MyClass.prototype.myFunc = function() {
alert("I can't use private instance variables. :(");
}
I've read numerous times people saying that using Method 2 is more efficient as all instances share the same copy of the function rather than each getting their own. Defining functions via the prototype has a huge disadvantage though - it makes it impossible to have private instance variables.
Even though, in theory, using Method 1 gives each instance of an object its own copy of the function (and thus uses way more memory, not to mention the time required for allocations) - is that what actually happens in practice? It seems like an optimization web browsers could easily make is to recognize this extremely common pattern, and actually have all instances of the object reference the same copy of functions defined via these "constructor functions". Then it could only give an instance its own copy of the function if it is explicitly changed later on.
Any insight - or, even better, real world experience - about performance differences between the two, would be extremely helpful.

See http://jsperf.com/prototype-vs-this
Declaring your methods via the prototype is faster, but whether or not this is relevant is debatable.
If you have a performance bottleneck in your app it is unlikely to be this, unless you happen to be instantiating 10000+ objects on every step of some arbitrary animation, for example.
If performance is a serious concern, and you'd like to micro-optimise, then I would suggest declaring via prototype. Otherwise, just use the pattern that makes most sense to you.
I'll add that, in JavaScript, there is a convention of prefixing properties that are intended to be seen as private with an underscore (e.g. _process()). Most developers will understand and avoid these properties, unless they're willing to forgo the social contract, but in that case you might as well not cater to them. What I mean to say is that: you probably don't really need true private variables...

In the new version of Chrome, this.method is about 20% faster than prototype.method, but creating new object is still slower.
If you can reuse the object instead of always creating an new one, this can be 50% - 90% faster than creating new objects. Plus the benefit of no garbage collection, which is huge:
http://jsperf.com/prototype-vs-this/59

It only makes a difference when you're creating lots of instances. Otherwise, the performance of calling the member function is exactly the same in both cases.
I've created a test case on jsperf to demonstrate this:
http://jsperf.com/prototype-vs-this/10

You might not have considered this, but putting the method directly on the object is actually better in one way:
Method invocations are very slightly faster (jsperf) since the prototype chain does not have to be consulted to resolve the method.
However, the speed difference is almost negligible. On top of that, putting a method on a prototype is better in two more impactful ways:
Faster to create instances (jsperf)
Uses less memory
Like James said, this difference can be important if you are instantiating thousands of instances of a class.
That said, I can certainly imagine a JavaScript engine that recognizes that the function you are attaching to each object does not change across instances and thus only keeps one copy of the function in memory, with all instance methods pointing to the shared function. In fact, it seems that Firefox is doing some special optimization like this but Chrome is not.
ASIDE:
You are right that it is impossible to access private instance variables from inside methods on prototypes. So I guess the question you must ask yourself is do you value being able to make instance variables truly private over utilizing inheritance and prototyping? I personally think that making variables truly private is not that important and would just use the underscore prefix (e.g., "this._myVar") to signify that although the variable is public, it should be considered to be private. That said, in ES6, there is apparently a way to have the both of both worlds!

You may use this approach and it will allow you to use prototype and access instance variables.
var Person = (function () {
function Person(age, name) {
this.age = age;
this.name = name;
}
Person.prototype.showDetails = function () {
alert('Age: ' + this.age + ' Name: ' + this.name);
};
return Person; // This is not referencing `var Person` but the Person function
}()); // See Note1 below
Note1:
The parenthesis will call the function (self invoking function) and assign the result to the var Person.
Usage
var p1 = new Person(40, 'George');
var p2 = new Person(55, 'Jerry');
p1.showDetails();
p2.showDetails();

In short, use method 2 for creating properties/methods that all instances will share. Those will be "global" and any change to it will be reflected across all instances. Use method 1 for creating instance specific properties/methods.
I wish I had a better reference but for now take a look at this. You can see how I used both methods in the same project for different purposes.
Hope this helps. :)

This answer should be considered an expansion of the rest of the answers filling in missing points. Both personal experience and benchmarks are incorporated.
As far as my experience goes, I use constructors to literally construct my objects religiously, whether methods are private or not. The main reason being that when I started that was the easiest immediate approach to me so it's not a special preference. It might have been as simple as that I like visible encapsulation and prototypes are a bit disembodied. My private methods will be assigned as variables in the scope as well. Although this is my habit and keeps things nicely self contained, it's not always the best habit and I do sometimes hit walls. Apart from wacky scenarios with highly dynamic self assembling according to configuration objects and code layout it tends to be the weaker approach in my opinion particularly if performance is a concern. Knowing that the internals are private is useful but you can achieve that via other means with the right discipline. Unless performance is a serious consideration, use whatever works best otherwise for the task at hand.
Using prototype inheritance and a convention to mark items as private does make debugging easier as you can then traverse the object graph easily from the console or debugger. On the other hand, such a convention makes obfuscation somewhat harder and makes it easier for others to bolt on their own scripts onto your site. This is one of the reasons the private scope approach gained popularity. It's not true security but instead adds resistance. Unfortunately a lot of people do still think it's a genuinely way to program secure JavaScript. Since debuggers have gotten really good, code obfuscation takes its place. If you're looking for security flaws where too much is on the client, it's a design pattern your might want to look out for.
A convention allows you to have protected properties with little fuss. That can be a blessing and a curse. It does ease some inheritance issues as it is less restrictive. You still do have the risk of collision or increased cognitive load in considering where else a property might be accessed. Self assembling objects let you do some strange things where you can get around a number of inheritance problems but they can be unconventional. My modules tend to have a rich inner structure where things don't get pulled out until the functionality is needed elsewhere (shared) or exposed unless needed externally. The constructor pattern tends to lead to creating self contained sophisticated modules more so than simply piecemeal objects. If you want that then it's fine. Otherwise if you want a more traditional OOP structure and layout then I would probably suggest regulating access by convention. In my usage scenarios complex OOP isn't often justified and modules do the trick.
All of the tests here are minimal. In real world usage it is likely that modules will be more complex making the hit a lot greater than tests here will indicate. It's quite common to have a private variable with multiple methods working on it and each of those methods will add more overhead on initialisation that you wont get with prototype inheritance. In most cases is doesn't matter because only a few instances of such objects float around although cumulatively it might add up.
There is an assumption that prototype methods are slower to call because of prototype lookup. It's not an unfair assumption, I made the same myself until I tested it. In reality it's complex and some tests suggest that aspect is trivial. Between, prototype.m = f, this.m = f and this.m = function... the latter performs significantly better than the first two which perform around the same. If the prototype lookup alone were a significant issue then the last two functions instead would out perform the first significantly. Instead something else strange is going on at least where Canary is concerned. It's possible functions are optimised according to what they are members of. A multitude of performance considerations come into play. You also have differences for parameter access and variable access.
Memory Capacity. It's not well discussed here. An assumption you can make up front that's likely to be true is that prototype inheritance will usually be far more memory efficient and according to my tests it is in general. When you build up your object in your constructor you can assume that each object will probably have its own instance of each function rather than shared, a larger property map for its own personal properties and likely some overhead to keep the constructor scope open as well. Functions that operate on the private scope are extremely and disproportionately demanding of memory. I find that in a lot of scenarios the proportionate difference in memory will be much more significant than the proportionate difference in CPU cycles.
Memory Graph. You also can jam up the engine making GC more expensive. Profilers do tend to show time spent in GC these days. It's not only a problem when it comes to allocating and freeing more. You'll also create a larger object graph to traverse and things like that so the GC consumes more cycles. If you create a million objects and then hardly touch them, depending on the engine it might turn out to have more of an ambient performance impact than you expected. I have proven that this does at least make the gc run for longer when objects are disposed of. That is there tends to be a correlation with memory used and the time it takes to GC. However there are cases where the time is the same regardless of the memory. This indicates that the graph makeup (layers of indirection, item count, etc) has more impact. That's not something that is always easy to predict.
Not many people use chained prototypes extensively, myself included I have to admit. Prototype chains can be expensive in theory. Someone will but I've not measured the cost. If you instead build your objects entirely in the constructor and then have a chain of inheritance as each constructor calls a parent constructor upon itself, in theory method access should be much faster. On the other hand you can accomplish the equivalent if it matters (such as flatten the prototypes down the ancestor chain) and you don't mind breaking things like hasOwnProperty, perhaps instanceof, etc if you really need it. In either case things start to get complex once you down this road when it comes to performance hacks. You'll probably end up doing things you shouldn't be doing.
Many people don't directly use either approach you've presented. Instead they make their own things using anonymous objects allowing method sharing any which way (mixins for example). There are a number of frameworks as well that implement their own strategies for organising modules and objects. These are heavily convention based custom approaches. For most people and for you your first challenge should be organisation rather than performance. This is often complicated in that Javascript gives many ways of achieving things versus languages or platforms with more explicit OOP/namespace/module support. When it comes to performance I would say instead to avoid major pitfalls first and foremost.
There's a new Symbol type that's supposed to work for private variables and methods. There are a number of ways to use this and it raises a host of questions related to performance and access. In my tests the performance of Symbols wasn't great compared to everything else but I never tested them thoroughly.
Disclaimers:
There are lots of discussions about performance and there isn't always a permanently correct answer for this as usage scenarios and engines change. Always profile but also always measure in more than one way as profiles aren't always accurate or reliable. Avoid significant effort into optimisation unless there's definitely a demonstrable problem.
It's probably better instead to include performance checks for sensitive areas in automated testing and to run when browsers update.
Remember sometimes battery life matters as well as perceptible performance. The slowest solution might turn out faster after running an optimising compiler on it (IE, a compiler might have a better idea of when restricted scope variables are accessed than properties marked as private by convention). Consider backend such as node.js. This can require better latency and throughput than you would often find on the browser. Most people wont need to worry about these things with something like validation for a registration form but the number of diverse scenarios where such things might matter is growing.
You have to be careful with memory allocation tracking tools in to persist the result. In some cases where I didn't return and persist the data it was optimised out entirely or the sample rate was not sufficient between instantiated/unreferenced, leaving me scratching my head as to how an array initialised and filled to a million registered as 3.4KiB in the allocation profile.
In the real world in most cases the only way to really optimise an application is to write it in the first place so you can measure it. There are dozens to hundreds of factors that can come into play if not thousands in any given scenario. Engines also do things that can lead to asymmetric or non-linear performance characteristics. If you define functions in a constructor, they might be arrow functions or traditional, each behaves differently in certain situations and I have no idea about the other function types. Classes also don't behave the same in terms as performance for prototyped constructors that should be equivalent. You need to be really careful with benchmarks as well. Prototyped classes can have deferred initialisation in various ways, especially if your prototyped your properties as well (advice, don't). This means that you can understate initialisation cost and overstate access/property mutation cost. I have also seen indications of progressive optimisation. In these cases I have filled a large array with instances of objects that are identical and as the number of instances increase the objects appear to be incrementally optimised for memory up to a point where the remainder is the same. It is also possible that those optimisations can also impact CPU performance significantly. These things are heavily dependent not merely on the code you write but what happens in runtime such as number of objects, variance between objects, etc.

Variables vs Expressions as Javascript Arguments

Are there any performance issues with passing an argument as an expression, instead of first making it a variable?
someFunction( x+2 );
vs.
var total = x+2;
someFunction( total );
And how about functions?
someFunction( someOtherFunction() );

No. And, more important, this sort of micro-optimization is (almost certainly) meaningless.
Having said that, if you were to use the result of the expression more than once, then there might be some completely imperceptible and totally not-worth-worrying-about benefit to saving the result of the calculation.
Write it to be readable. Don't worry about this stuff.

Just the obvious: Making a variable creates a variable. This costs memory and consumes some time when executing. Afterwards, it either will need time to garbage collect it, or not free the memory if your function leaks.
However, you won't notice any differences. The performance is not measurable at that level. Rule of thumb: Use variables when you really need them or when they improve readabilty of your code.

Though the difference is minimal, the answer is really implementation-specific; JavaScript engines almost certainly differ in how they allocate things. However, I can tell you that most likely, the differences are similar to what they would be in most other languages of which I can examine the memory and processor registers in the debugger. Let's examine one scenario:
var sum = x+2;
someFunction(sum);
This allocates memory to hold sum, which hangs around as long as the function is in scope. If the function ends up being a closure, this could be forever. In a recursive function this could be significant.
someFunction(x+2);
In most languages, this will compute x+2 on the stack and pass the result to someFunction. No memory is left hanging around.
The answer would be exactly the same for a function return value.
So in summary:
The exact answer depends on the JavaScript engine's implementation.
Most likely you won't notice a performance difference.
You may want to use variables when the result is re-used, or, when you want to examine the result easily in the debugger.
It's mostly a matter of personal preference.

Creating a local variable whose scope does not extend beyond the current function does not incur any cost compared to not creating one and writing the expression directly as the argument to the function. In fact, nothing tells you that writing someFunction(x*2) won't be transformed to code that binds a variable to the result of x*2 internally by the javascript compiler - indeed, many compilers and JITs use SSA as one of their intermediate representations, in which form a variable is always bound to the result of every sub expression. See the relevant Wikipedia entry. Whether in a closure or not makes no difference.
The only two relevant questions that should concern you to make a choice between introducing a new variable and writing the expression as an argument directly are:
Readability: does naming the result of the expression make clearer what the expression is computing;
Cost of evaluating the expression: if you will be writing the expression more than once, then binding a variable to the result will you to reuse avoid recomputing the result everytime. This is only relevant if your expression is expected to take a long time to compute.
If you only need to write the expression once inside a function definition then binding a variable to the result may well make the result live in memory longer than is strictly necessary, but this is nearly always completely irrelevant: most function calls are very short lived, in many cases the result does not take up much memory and the memory allocated on the stack will be reclaimed upon function exit and the memory allocated on the heap will be reclaimed by garbage collector soon thereafter.

What are the benefits to using anonymous functions instead of named functions for callbacks and parameters in JavaScript event code?

I'm new-ish to JavaScript. I understand many of the concepts of the language, I've been reading up on the prototype inheritance model, and I'm whetting my whistle with more and more interactive front-end stuff. It's an interesting language, but I'm always a bit turned off by the callback spaghetti that is typical of many non-trivial interaction models.
Something that has always seemed strange to me is that in spite of the readability nightmare that is a nest of JavaScript nested callbacks, the one thing that I very rarely see in many examples and tutorials is the use of predefined named functions as callback arguments. I'm a Java programmer by day, and discarding the stereotypical jabs about Enterprise-y names for units of code one of the things I've come to enjoy about working in a language with a strong selection of featureful IDE's is that using meaningful, if long, names can make the intent and meaning of code much clearer without making it more difficult to actually be productive. So why not use the same approach when writing JavaScript code?
Giving it thought, I can come up with arguments that are both for and against this idea, but my naivety and newness to the language impairs me from reaching any conclusions as to why this would be good at a technical level.
Pros:
Flexibility. An asynchronous function with a callback parameter could be reached by one of many different code paths and it could be harried to have to write a named function to account for every single possible edge case.
Speed. It plays heavily in to the hacker mentality. Bolt things on to it until it works.
Everyone else is doing it
Smaller file sizes, even if trivially so, but every bit counts on the web.
Simpler AST? I would assume that anonymous functions are generated at runtime and so the JIT won't muck about with mapping the name to instructions, but I'm just guessing at this point.
Quicker dispatching? Not sure about this one either. Guessing again.
Cons:
It's hideous and unreadable
It adds to the confusion when you're nested nuts deep in a swamp of callbacks (which, to be fair, probably means you're writing poorly constructed code to begin with, but it's quite common).
For someone without a functional background it can be a bizarre concept to grok
With so many modern browsers showing the ability to execute JavaScript code much faster than before, I'm failing to see how any trivial sort of performance gain one might get out using anonymous callbacks would be a necessity. It seems that, if you are in a situation where using a named function is feasible (predictable behavior and path of execution) then there would be no reason not to.
So are there any technical reasons or gotchas that I'm not aware of that makes this practice so commonplace for a reason?

I use anonymous functions for three reasons:
If no name is needed because the function is only ever called in one place, then why add a name to whatever namespace you're in.
Anonymous functions are declared inline and inline functions have advantages in that they can access variables in the parent scopes. Yes, you can put a name on an anonymous function, but that's usually pointless if it's declared inline. So inline has a significant advantage and if you're doing inline, there's little reason to put a name on it.
The code seems more self-contained and readable when handlers are defined right inside the code that's calling them. You can read the code in almost sequential fashion rather than having to go find the function with that name.
I do try to avoid deep nesting of anonymous functions because that can be hairy to understand and read. Usually when that happens, there's a better way to structure the code (sometimes with a loop, sometimes with a data table, etc...) and named functions isn't usually the solution there either.
I guess I'd add that if a callback starts to get more than about 15-20 lines long and it doesn't need direct access to variables in the parent scope, I would be tempted to give it a name and break it out into it's own named function declared elsewhere. There is definitely a readability point here where a non-trivial function that gets long is just more maintainable if it's put in its own named unit. But, most callbacks I end up with are not that long and I find it more readable to keep them inline.

I prefer named functions myself, but for me it comes down to one question:
Will I use this function anywhere else?
If the answer is yes, I name/define it. If not, pass it as an anonymous function.
If you only use it once, it doesn't make sense to crowd the global namespace with it. In today's complex front-ends, the number of named functions that could have been anonymous grows quickly (easily over 1000 on really intricate designs), resulting in (relatively) large performance gains by preferring anonymous functions.
However, code maintainability is also extremely important. Each situation is different. If you're not writing a lot of these functions to begin with, there's no harm in doing it either way. It's really up to your preference.
Another note about names. Getting in the habit of defining long names will really hurt your file size. Take the following example.
Assume both of these functions do the same thing:
function addTimes(time1, time2)
{
// return time1 + time2;
}
function addTwoTimesIn24HourFormat(time1, time2)
{
// return time1 + time2;
}
The second tells you exactly what it does in the name. The first is more ambiguous. However, there are 17 characters of difference in the name. Say the function is called 8 times throughout the code, that's 153 extra bytes your code didn't need to have. Not colossal, but if it's a habit, extrapolating that to 10s or even 100s of functions will easily mean a few KB of difference in the download.
Again however, maintainability needs to be weighed against the benefits of performance. This is the pain of dealing with a scripted language.

A bit late to the party, but some not yet mentioned aspects to functions, anonymous or otherwise...
Anon funcs are not easily referred to in humanoid conversations about code, amongst a team. E.g., "Joe, could you explain what the algorithm does, within that function. ... Which one? The 17th anonymous function within the fooApp function. ... No, not that one! The 17th one!"
Anon funcs are anonymous to the debugger as well. (duh!) Therefore, the debugger stack trace will generally just show a question mark or similar, making it less useful when you have set multiple breakpoints. You hit the breakpoint, but find yourself scrolling the debug window up/down to figure out where the hell you are in your program, because hey, question mark function just doesn't do it!
Concerns about polluting the global namespace are valid, but easily remedied by naming your functions as nodes within your own root object, like "myFooApp.happyFunc = function ( ... ) { ... }; ".
Functions that are available in the global namespace, or as nodes in your root object like above, can be invoked from the debugger directly, during development and debug. E.g., at the console command line, do "myFooApp.happyFunc(42)". This is an extremely powerful ability that does not exist (natively) in compiled programming languages. Try that with an anon func.
Anon funcs can be made more readable by assigning them to a var, and then passing the var as the callback (instead of inlining). E.g.:
var funky = function ( ... ) { ... };
jQuery('#otis').click(funky);
Using the above approach, you could potentially group several anon funcs at the top of the parental func, then below that, the meat of sequential statements becomes much tighter grouped, and easier to read.

Anonymous functions are useful because they help you control which functions are exposed.
More Detail: If there is no name, you can't reassign it or tamper with it anywhere but the exact place it was created. A good rule of thumb is, if you don't need to re-use this function anywhere, it's a good idea to consider if an anonymous function would be better to prevent getting tampered with anywhere.
Example:
If you're working on a big project with a lot of people, what if you have a function inside of a bigger function and you name it something? That means anyone working with you and also editing code in the bigger function can do stuff to that smaller function at any time. What if you named it "add" for instance, and someone reassigned "add" to a number instead inside the same scope? Then the whole thing breaks!
PS -I know this is a very old post, but there is a much simpler answer to this question and I wish someone had put it this way when I was looking for the answer myself as a beginner- I hope you're ok with reviving an old thread!

Its more readable using named functions and they are also capable of self-referencing as in the example below.
(function recursion(iteration){
if (iteration > 0) {
console.log(iteration);
recursion(--iteration);
} else {
console.log('done');
}
})(20);
console.log('recursion defined? ' + (typeof recursion === 'function'));
http://jsfiddle.net/Yq2WD/
This is nice when you want to have an immediately invoked function that references itself but does not add to the global namespace. It's still readable but not polluting. Have your cake and eat it to.
Hi, my name is Jason OR hi, my name is ???? you pick.

Well, just to be clear for the sake of my arguments, the following are all anonymous functions/function expressions in my book:
var x = function(){ alert('hi'); },
indexOfHandyMethods = {
hi: function(){ alert('hi'); },
high: function(){
buyPotatoChips();
playBobMarley();
}
};
someObject.someEventListenerHandlerAssigner( function(e){
if(e.doIt === true){ doStuff(e.someId); }
} );
(function namedButAnon(){ alert('name visible internally only'); })()
Pros:
It can reduce a bit of cruft, particularly in recursive functions (where you could (should actually since arguments.callee is deprecated) still use a named reference per the last example internally), and makes it clear the function only ever fires in this one place.
Code legibility win: in the example of the object literal with anon funcs assigned as methods, it would be silly to add more places to hunt and peck for logic in your code when the whole point of that object literal is to plop some related functionality in the same conveniently referenced spot. When declaring public methods in a constructor, however, I do tend to define labeled functions inline and then assign as references of this.sameFuncName. It lets me use the same methods internally without the 'this.' cruft and makes order of definition a non-concern when they call each other.
Useful for avoiding needless global namespace pollution - internal namespaces, however, shouldn't ever be that broadly filled or handled by multiple teams simultaneously so that argument seems a bit silly to me.
I agree with the inline callbacks when setting short event handlers. It's silly to have to hunt for a 1-5 line function, especially since with JS and function hoisting, the definitions could end up anywhere, not even within the same file. This could happen by accident without breaking anything and no, you don't always have control of that stuff. Events always result in a callback function being fired. There's no reason to add more links to the chain of names you need to scan through just to reverse engineer simple event-handlers in a large codebase and the stack trace concern can be addressed by abstracting event triggers themselves into methods that log useful info when debug mode is on and fire the triggers. I'm actually starting to build entire interfaces this way.
Useful when you WANT the order of function definition to matter. Sometimes you want to be certain a default function is what you think it is until a certain point in the code where it's okay to redefine it. Or you want breakage to be more obvious when dependencies get shuffled.
Cons:
Anon functions can't take advantage of function hoisting. This is a major difference. I tend to take heavy advantage of hoisting to define my own explicitly named funcs and object constructors towards the bottom and get to the object definition and main-loop type stuff right up at the top. I find it makes the code easier to read when you name your vars well and get a broad view of what's going on before ctrl-Fing for details only when they matter to you. Hoisting can also be a huge benefit in heavily event-driven interfaces where imposing a strict order of what's available when can bite you in the butt. Hoisting has its own caveats (like circular reference potential) but it is a very useful tool for organizing and making code legible when used right.
Legibility/Debug. Absolutely they get used way too heavily at times and it can make debug and code legibility a hassle. Codebases that rely heavily on JQ, for instance, can be a serious PITA to read and debug if you don't encapsulate the near-inevitable anon-heavy and massively overloaded args of the $ soup in a sensible way. JQuery's hover method for instance, is a classic example of over-use of anon funcs when you drop two anon funcs into it, since it's easy for a first-timer to assume it's a standard event listener assignment method rather than one method overloaded to assign handlers for one or two events. $(this).hover(onMouseOver, onMouseOut) is a lot more clear than two anon funcs.

performance difference between creating a closure and some_func.bind(this)

I'm writing a lot of code in javascript lately and I'm using Prototype.js to help with a lot of the boilerplate, just the bind method is more than worth it since I like using closures instead of objects to do the heavy lifting. It's hard to give up habits picked up from using ruby blocks. So here's my question: Is there any particular performance difference between
var func = some_func.bind(this);
...
func();
and
var that = this;
...
some_func(); // we just rename 'this' everywhere inside some_func to 'that'
These tricks are required because inner functions default to the global context instead of the context of the outer function. In particular which version keeps things more 'flat'. If there is recursion involved then the bind version will come to a crawl because bind will keep folding functions inside functions until unrolling things is impossible, at least I think that's what happens. Does the second version have the same problem.

The performance difference is really extremely negligible unless you are running the operation thousands of times in rapid fire. So I'd usually go with the bind because it makes for cleaner, less bug-prone code.