Related
I am working on some (embedded) device, recently I just started thinking maybe to use less memory, in case stack size isn't that big.
I have long functions (unfortunately).
And inside I was thinking to save space in this way.
Imagine there is code
1. void f()
2. {
3. ...
4. char someArray[300];
5. char someOtherArray[300];
6. someFunc(someArray, someOtherArray);
7. ...
8. }
Now, imagine, someArray and someOtherArray are never used in f function beyond line: 6.
Would following save some stack space??
1. void f()
2. {
3. ...
4. {//added
5. char someArray[300];
6. char someOtherArray[300];
7. someFunc(someArray, someOtherArray);
8. }//added
9. ...
8. }
nb: removed second part of the question
For the compiler proper both are exactly the same and thus makes no difference. The preprocessor would replace all instances of TEXT1 with the string constant.
#define TEXT1 "SomeLongStringLiteral"
someFunc(TEXT1)
someOtherFunc(TEXT1)
After the preprocessor's job is done, the above snippet becomes
someFunc("SomeLongStringLiteral");
someOtherFunc("SomeLongStringLiteral");
Thus it makes no difference performance or memory-wise.
Aside: The reason #define TEXT1 "SomeLongStringLiteral" is done is to have a single place to change all instances of TEXT1s usage; but that's a convinience only for the programmer and has no effect on the produced output.
recently I just started thinking maybe to use less memory, in case stack size isn't that big.
Never micro optimise or prematurely optimise. In case the stack size isn't that big, you'll get to know it when you benchmark/measure it. Don't make any assumptions when you optimise; 99% of the times it'd be wrong.
I am working on some device
Really? Are you? I wouldn't have thought that.
Now, imagine, someArray and someOtherArray are never used in f function beyond line 6. Would following save some stack space?
On a good compiler, it wouldn't make a difference. By the standard, it isn't specified if it saves or not, it isn't even specified if there is a stack or not.
But on a not so good compiler, the one with the additional {} may be better. It is worth a test: compile it and look at the generated assembler code.
it seems my compiler doesn't allow me to do this (this is C), so never mind...
But it should so. What happens then? Maybe you are just confusing levels of {} ...
I'll ask another one here.
Would better be a separate question...
someFunc("SomeLongStringLiteral");
someOtherFunc("SomeLongStringLiteral");
vs.
someFunc(TEXT1)
someOtherFunc(TEXT1)
A #define is processed before any compilation step, so it makes absolutely no difference.
If it happens within the same compilation unit, the compiler will tie them together anyway. (At least, in this case. On an ATXmega, if you use PSTR("whatever") for having them in flash space only, each occurrence of them will be put into flash separately. But that's a completely different thing...)
Modern compilers should push stack variables before they are used, and pop them when they are no longer needed. The old thinking with { ... } marking the start and end of a stack push/pop should be rather obsolete by now.
Since 1999, C allows stack variables to be allocated anywhere and not just immediately after a {. C++ allowed this far earlier. Today, where the local variable is declared inside the scope has little to do with when it actually starts to exist in the machine code. And similarly, the } has little to do with when is ceases to exist.
So regarding adding extra { }, don't bother. It is premature optimization and only adds pointless clutter.
Regarding the #define it absolutely makes no difference in terms of efficiency. Macros are just text replacement.
Furthermore, from the generic point-of-view, data must always be allocated somewhere. Data used by a program cannot be allocated in thin air! That's a very common misunderstanding. For example, many people incorrectly believe that
int x = func();
if(x == something)
consumes more memory than
if(func() == something)
But both examples compile into identical machine code. The result of func must be stored somewhere, it cannot be stored in thin air. In the first example, the result is stored in a memory segment that the programmer may refer to as x.
In the second example, it is stored in the very same memory segment, taking up the same amount of space, for the same duration of program execution. The only difference is that the memory segment is anonymous and the programmer has no name for it. As far as the machine code is concerned, that doesn't matter, since no variable names exist in machine code.
And this would be why every professional C programmer needs to understand a certain amount of assembler. You cannot hope to ever do any kind of manual code optimization if you don't.
(Please don't ask two questions in one, this is really annoying since you get two types of answer for your two different questions.)
For your first question. Probably putting {} around the use of a variable will not help. The lifetime of automatic variables that are not VLA (see below) is not bound to the scope in which it is declared. So compilers may have a hard time in figuring out how the use of the stack may be optimized, and maybe don't do such an optimization at all. In your case this is most likely the case, since you are exporting pointers to your data to a function that is perhaps not visible, here. The compiler has no way to figure out if there is a valid use of the arrays later on in the code.
I see two ways to "force" the compiler into optimizing that space, functions or VLA. The first, functions is simple: instead of putting the block around the code, put it in a static function. Function calls are quite optimized on modern platforms, and here the compiler knows exactly how he may clear the stack at the end.
The second alternative in your case is a VLA, variable length array, if you compiler supports that c99 feature. Arrays that have a size that doesn't depend on a compile time constant have a special rule for their lifetime. That lifetime exactly ends at the end of the scope where they are defined. Even a const-qualified variable could be used for that:
{
size_t const len = 300;
char someArray[len];
char someOtherArray[len];
someFunc(someArray, someOtherArray);
}
At the end, on a given platform, you'd really have to inspect what assembler your compiler produces.
In a codebase I reviewed, I found the following idiom.
void notify(struct actor_t act) {
write(act.pipe, "M", 1);
}
// thread A sending data to thread B
void send(byte *data) {
global.data = data;
notify(threadB);
}
// in thread B event loop
read(this.sock, &cmd, 1);
switch (cmd) {
case 'M': use_data(global.data);break;
...
}
"Hold it", I said to the author, a senior member of my team, "there's no memory barrier here! You don't guarantee that global.data will be flushed from the cache to main memory. If thread A and thread B will run in two different processors - this scheme might fail".
The senior programmer grinned, and explained slowly, as if explaining his five years old boy how to tie his shoelaces: "Listen young boy, we've seen here many thread related bugs, in high load testing, and in real clients", he paused to scratch his longish beard, "but we've never had a bug with this idiom".
"But, it says in the book..."
"Quiet!", he hushed me promptly, "Maybe theoretically, it's not guaranteed, but in practice, the fact you used a function call is effectively a memory barrier. The compiler will not reorder the instruction global.data = data, since it can't know if anyone using it in the function call, and the x86 architecture will ensure that the other CPUs will see this piece of global data by the time thread B reads the command from the pipe. Rest assured, we have ample real world problems to worry about. We don't need to invest extra effort in bogus theoretical problems.
"Rest assured my boy, in time you'll understand to separate the real problem from the I-need-to-get-a-PhD non-problems."
Is he correct? Is that really a non-issue in practice (say x86, x64 and ARM)?
It's against everything I learned, but he does have a long beard and a really smart looks!
Extra points if you can show me a piece of code proving him wrong!
Memory barriers aren't just to prevent instruction reordering. Even if instructions aren't reordered it can still cause problems with cache coherence. As for the reordering - it depends on your compiler and settings. ICC is particularly agressive with reordering. MSVC w/ whole program optimization can be, too.
If your shared data variable is declared as volatile, even though it's not in the spec most compilers will generate a memory variable around reads and writes from the variable and prevent reordering. This is not the correct way of using volatile, nor what it was meant for.
(If I had any votes left, I'd +1 your question for the narration.)
In practice, a function call is a compiler barrier, meaning that the compiler will not move global memory accesses past the call. A caveat to this is functions which the compiler knows something about, e.g. builtins, inlined functions (keep in mind IPO!) etc.
So a processor memory barrier (in addition to a compiler barrier) is in theory needed to make this work. However, since you're calling read and write which are syscalls that change the global state, I'm quite sure that the kernel issues memory barriers somewhere in the implementation of those. There is no such guarantee though, so in theory you need the barriers.
The basic rule is: the compiler must make the global state appear to be exactly as you coded it, but if it can prove that a given function doesn't use global variables then it can implement the algorithm any way it chooses.
The upshot is that traditional compilers always treated functions in another compilation unit as a memory barrier because they couldn't see inside those functions. Increasingly, modern compilers are growing "whole program" or "link time" optimization strategies which break down these barriers and will cause poorly written code to fail, even though it's been working fine for years.
If the function in question is in a shared library then it won't be able to see inside it, but if the function is one defined by the C standard then it doesn't need to -- it already knows what the function does -- so you have to be careful of those also. Note that a compiler will not recognise a kernel call for what it is, but the very act of inserting something that the compiler can't recognise (inline assembler, or a function call to an assembler file) will create a memory barrier in itself.
In your case, notify will either be a black box the compiler can't see inside (a library function) or else it will contain a recognisable memory barrier, so you are most likely safe.
In practice, you have to write very bad code to fall over this.
In practice, he's correct and a memory barrier is implied in this specific case.
But the point is that if its presence is "debatable", the code is already too complex and unclear.
Really guys, use a mutex or other proper constructs. It's the only safe way to deal with threads and to write maintainable code.
And maybe you'll see other errors, like that the code is unpredictable if send() is called more than one time.
I'm currently developing a very fast algorithm, with one part of it being an extremely fast scanner and statistics function.
In this quest, i'm after any performance benefit.
Therefore, I'm also interested in keeping the code "multi-thread" friendly.
Now for the question :
i've noticed that putting some very frequently accessed variables and arrays into "Global", or "static local" (which does the same), there is a measurable performance benefit (in the range of +10%).
I'm trying to understand why, and to find a solution about it, since i would prefer to avoid using these types of allocation.
Note that i don't think the difference comes from "allocation", since allocating a few variables and small array on the stack is almost instantaneous. I believe the difference comes from "accessing" and "modifying" data.
In this search, i've found this old post from stackoverflow :
C++ performance of global variables
But i'm very disappointed by the answers there. Very little explanation, mostly ranting about "you should not do that" (hey, that's not the question !) and very rough statements like 'it doesn't affect performance', which is obviously incorrect, since i'm measuring it with precise benchmark tools.
As said above, i'm looking for an explanation, and, if it exists, a solution to this issue. So far, i've got the feeling that calculating the memory address of a local (dynamic) variable costs a bit more than a global (or local static). Maybe something like an ADD operation difference. But that doesn't help finding a solution...
It really depends on your compiler, platform, and other details. However, I can describe one scenario where global variables are faster.
In many cases, a global variable is at a fixed offset. This allows the generated instructions to simply use that address directly. (Something along the lines of MOV AX,[MyVar].)
However, if you have a variable that's relative to the current stack pointer or a member of a class or array, some math is required to take the address of the array and determine the address of the actual variable.
Obviously, if you need to place some sort of mutex on your global variable in order to keep it thread-safe, then you'll almost certainly more than lose any performance gain.
Creating local variables can be literally free if they are POD types. You likely are overflowing a cache line with too many stack variables or other similar alignment-based causes which are very specific to your piece of code. I usually find that non-local variables significantly decrease performance.
It's hard to beat static allocation for speed, and while the 10% is a pretty small difference, it could be due to address calculation.
But if you're looking for speed,
your example in a comment while(p<end)stats[*p++]++; is an obvious candidate for unrolling, such as:
static int stats[M];
static int index_array[N];
int *p = index_array, *pend = p+N;
// ... initialize the arrays ...
while (p < pend-8){
stats[p[0]]++;
stats[p[1]]++;
stats[p[2]]++;
stats[p[3]]++;
stats[p[4]]++;
stats[p[5]]++;
stats[p[6]]++;
stats[p[7]]++;
p += 8;
}
while(p<pend) stats[*p++]++;
Don't count on the compiler to do it for you. It might or might not be able to figure it out.
Other possible optimizations come to mind, but they depend on what you're actually trying to do.
If you have something like
int stats[256]; while (p<end) stats[*p++]++;
static int stats[256]; while (p<end) stats[*p++]++;
you are not really comparing the same thing because for the first instance you are not doing an initialization of your array. Written explicitly the second line is equivalent to
static int stats[256] = { 0 }; while (p<end) stats[*p++]++;
So to be a fair comparison you should have the first read
int stats[256] = { 0 }; while (p<end) stats[*p++]++;
Your compiler might deduce much more things if he has the variables in a known state.
Now then, there could be runtime advantage of the static case, since the initialization is done at compile time (or program startup).
To test if this makes up for your difference you should run the same function with the static declaration and the loop several times, to see if the difference vanishes if your number of invocations grows.
But as other said already, best is to inspect the assembler that your compiler produces to see what effective difference there are in the code that is produced.
I think this is wrong, it should start as NULL and not with a random value. In the case that you have a pointer with a random memory address as its default value it could be a very dangerous thing, no?
The variables start out uninitialized because that's the fastest way - why waste the CPU cycles on initialization if you're going to write another value there anyway?
If you want a variable to be initialized after creation, just initialize it. :)
About it being a dangerous thing: Every good compiler will warn you if you try to use a variable without initialization.
No. C is a very efficient language, one that has traditionally been faster that a lot of other languages. One of the reasons for this is that it doesn't do too much on it's own. The programmer controls this.
In the case of initialization, C variables are not initialized to a random value. Rather, they are not initialized and so they contain whatever was at the memory location before.
If you wanted to initialize a variable to, say, 1 in your program, then it would be inefficient if the variable had already been initialized to zero or null. That would mean it was initialized twice.
Execution speed and overhead (or lack thereof) are the main reasons why. C is notorious for letting you walk off the proverbial cliff because it always assumes that the user knows better than it does.
Note that if you declared the variable as static it actually is guaranteed to be initialized to 0.
Variables start out with a random value because you are just handed a block of memory and told to deal with it yourself. It has whatever value that block of memory had before hand. Why should the program waste time setting the value to some arbitrary default when you are likely going to set it yourself later?
The design choice is performance, and it is one of the many reasons why C isn't the preferred language for most projects.
This has nothing to do with "if C were being designed today" or with efficiency of one initialization. Instead think of something like
void foo()
{
struct bar *ptrs[10000];
/* do something where only a few indices end up actually getting used */
}
Any language that forces useless initialization on you is doomed to be slow as hell for algorithms that can make use of sparse arrays where you don't care about the majority of the values, and have an easy way of knowing which values you care about.
If you don't like my example with such a large object on the stack, substitute malloc instead. It has the same semantics with regard to initialization.
In either case, if you want zero-initialization, you can get it with {0} or calloc.
It was a design choice made many ears ago, probably for efficiency reasons.
Statically allocated variables (globals and statics) are initialized to 0 if there's no explicit initialization - this could be justified even taking efficiency into account becuase it only occurs once. I'd guess the thinking was that for automatic variables (locals) that are allocated each time a scope is entered, implicit initialization was considered something that might cost too much and therefore should be left to the programmer's responsibility.
If C were being designed today, I wouldn't be surprised if that design decision were changed - especially since compilers are intelligent enough today to be able to optimize away an initialization that gets overwritten before any other use (or potential use).
However, there are so many C compiler toolchains that follow the spec of not initializing automatically, it would be foolish for a compiler to perform implicit initialization to a 'useful' value (like 0 or NULL). That would just encourage people targeting that tool chain to write code that didn't work correctly on other tool chains.
However, compilers can initialize local variables, and they often do. It's just that they initialize the locals to a values that's not generally useful (especially, that doesn't set a pointer to the null pointer). That kind of initialization isn't useful in writing your programming logic against, and it's not intended for that. It's intended to cause deterministic and reproducible errors so that if you erroneously use values that have been set by implicit initialization, you'll be able to find it easily in test/debug.
Usually this compiler behavior is turned on only for debug builds; I could see an argument being made for turning it on in release builds as well - particular if the release build can still optimize it away when the compiler can prove that the implicit initialized value is never used.
My function will be called thousands of times. If i want to make it faster, will changing the local function variables to static be of any use? My logic behind this is that, because static variables are persistent between function calls, they are allocated only the first time, and thus, every subsequent call will not allocate memory for them and will become faster, because the memory allocation step is not done.
Also, if the above is true, then would using global variables instead of parameters be faster to pass information to the function every time it is called? i think space for parameters is also allocated on every function call, to allow for recursion (that's why recursion uses up more memory), but since my function is not recursive, and if my reasoning is correct, then taking off parameters will in theory make it faster.
I know these things I want to do are horrible programming habits, but please, tell me if it is wise. I am going to try it anyway but please give me your opinion.
The overhead of local variables is zero. Each time you call a function, you are already setting up the stack for the parameters, return values, etc. Adding local variables means that you're adding a slightly bigger number to the stack pointer (a number which is computed at compile time).
Also, local variables are probably faster due to cache locality.
If you are only calling your function "thousands" of times (not millions or billions), then you should be looking at your algorithm for optimization opportunities after you have run a profiler.
Re: cache locality (read more here):
Frequently accessed global variables probably have temporal locality. They also may be copied to a register during function execution, but will be written back into memory (cache) after a function returns (otherwise they wouldn't be accessible to anything else; registers don't have addresses).
Local variables will generally have both temporal and spatial locality (they get that by virtue of being created on the stack). Additionally, they may be "allocated" directly to registers and never be written to memory.
The best way to find out is to actually run a profiler. This can be as simple as executing several timed tests using both methods and then averaging out the results and comparing, or you may consider a full-blown profiling tool which attaches itself to a process and graphs out memory use over time and execution speed.
Do not perform random micro code-tuning because you have a gut feeling it will be faster. Compilers all have slightly different implementations of things and what is true on one compiler on one environment may be false on another configuration.
To tackle that comment about fewer parameters: the process of "inlining" functions essentially removes the overhead related to calling a function. Chances are a small function will be automatically in-lined by the compiler, but you can suggest a function be inlined as well.
In a different language, C++, the new standard coming out supports perfect forwarding, and perfect move semantics with rvalue references which removes the need for temporaries in certain cases which can reduce the cost of calling a function.
I suspect you're prematurely optimizing, however, you should not be this concerned with performance until you've discovered your real bottlenecks.
Absolutly not! The only "performance" difference is when variables are initialised
int anint = 42;
vs
static int anint = 42;
In the first case the integer will be set to 42 every time the function is called in the second case ot will be set to 42 when the program is loaded.
However the difference is so trivial as to be barely noticable. Its a common misconception that storage has to be allocated for "automatic" variables on every call. This is not so C uses the already allocated space in the stack for these variables.
Static variables may actually slow you down as its some aggresive optimisations are not possible on static variables. Also as locals are in a contiguous area of the stack they are easier to cache efficiently.
There is no one answer to this. It will vary with the CPU, the compiler, the compiler flags, the number of local variables you have, what the CPU's been doing before you call the function, and quite possibly the phase of the moon.
Consider two extremes; if you have only one or a few local variables, it/they might easily be stored in registers rather than be allocated memory locations at all. If register "pressure" is sufficiently low that this may happen without executing any instructions at all.
At the opposite extreme there are a few machines (e.g., IBM mainframes) that don't have stacks at all. In this case, what we'd normally think of as stack frames are actually allocated as a linked list on the heap. As you'd probably guess, this can be quite slow.
When it comes to accessing the variables, the situation's somewhat similar -- access to a machine register is pretty well guaranteed to be faster than anything allocated in memory can possible hope for. OTOH, it's possible for access to variables on the stack to be pretty slow -- it normally requires something like an indexed indirect access, which (especially with older CPUs) tends to be fairly slow. OTOH, access to a global (which a static is, even though its name isn't globally visible) typically requires forming an absolute address, which some CPUs penalize to some degree as well.
Bottom line: even the advice to profile your code may be misplaced -- the difference may easily be so tiny that even a profiler won't detect it dependably, and the only way to be sure is to examine the assembly language that's produced (and spend a few years learning assembly language well enough to know say anything when you do look at it). The other side of this is that when you're dealing with a difference you can't even measure dependably, the chances that it'll have a material effect on the speed of real code is so remote that it's probably not worth the trouble.
It looks like the static vs non-static has been completely covered but on the topic of global variables. Often these will slow down a programs execution rather than speed it up.
The reason is that tightly scoped variables make it easy for the compiler to heavily optimise, if the compiler has to look all over your application for instances of where a global might be used then its optimising won't be as good.
This is compounded when you introduce pointers, say you have the following code:
int myFunction()
{
SomeStruct *A, *B;
FillOutSomeStruct(B);
memcpy(A, B, sizeof(A);
return A.result;
}
the compiler knows that the pointer A and B can never overlap and so it can optimise the copy. If A and B are global then they could possibly point to overlapping or identical memory, this means the compiler must 'play it safe' which is slower. The problem is generally called 'pointer aliasing' and can occur in lots of situations not just memory copies.
http://en.wikipedia.org/wiki/Pointer_alias
Using static variables may make a function a tiny bit faster. However, this will cause problems if you ever want to make your program multi-threaded. Since static variables are shared between function invocations, invoking the function simultaneously in different threads will result in undefined behaviour. Multi-threading is the type of thing you may want to do in the future to really speed up your code.
Most of the things you mentioned are referred to as micro-optimizations. Generally, worrying about these kind of things is a bad idea. It makes your code harder to read, and harder to maintain. It's also highly likely to introduce bugs. You'll likely get more bang for your buck doing optimizations at a higher level.
As M2tM suggests, running a profiler is also a good idea. Check out gprof for one which is quite easy to use.
You can always time your application to truly determine what is fastest. Here is what I understand: (all of this depends on the architecture of your processor, btw)
C functions create a stack frame, which is where passed parameters are put, and local variables are put, as well as the return pointer back to where the caller called the function. There is no memory management allocation here. It usually a simple pointer movement and thats it. Accessing data off the stack is also pretty quick. Penalties usually come into play when you're dealing with pointers.
As for global or static variables, they're the same...from the standpoint that they're going to be allocated in the same region of memory. Accessing these may use a different method of access than local variables, depends on the compiler.
The major difference between your scenarios is memory footprint, not so much speed.
Using static variables can actually make your code significantly slower. Static variables must exist in a 'data' region of memory. In order to use that variable, the function must execute a load instruction to read from main memory, or a store instruction to write to it. If that region is not in the cache, you lose many cycles. A local variable that lives on the stack will most surely have an address that is in the cache, and might even be in a cpu register, never appearing in memory at all.
I agree with the others comments about profiling to find out stuff like that, but generally speaking, function static variables should be slower. If you want them, what you are really after is a global. Function statics insert code/data to check if the thing has been initialized already that gets run every time your function is called.
Profiling may not see the difference, disassembling and knowing what to look for might.
I suspect you are only going to get a variation as much as a few clock cycles per loop (on average depending on the compiler, etc). Sometimes the change will be dramatic improvement or dramatically slower, and that wont necessarily be because the variables home has moved to/from the stack. Lets say you save four clock cycles per function call for 10000 calls on a 2ghz processor. Very rough calculation: 20 microseconds saved. Is 20 microseconds a lot or a little compared to your current execution time?
You will likely get more a performance improvement by making all of your char and short variables into ints, among other things. Micro-optimization is a good thing to know but takes lots of time experimenting, disassembling, timing the execution of your code, understanding that fewer instructions does not necessarily mean faster for example.
Take your specific program, disassemble both the function in question and the code that calls it. With and without the static. If you gain only one or two instructions and this is the only optimization you are going to do, it is probably not worth it. You may not be able to see the difference while profiling. Changes in where the cache lines hit could show up in profiling before changes in the code for example.