How to avoid run-time checks for running parts of code that become unreachable after compilation? The Next CEO of Stack OverflowHow can I profile C++ code running on Linux?Counting the total of same running processes in C++Usage of this next_combination codeIdentify objects in boost::shared_ptr<boost::thread>How C++ reference worksNamespaces and the Pre-Processorstd::atomic_is_lock_free(shared_ptr<T>*) didn't compileCleaning data after exception on class constructorTwo-level bind fails on GCC/libstdc++ and Clang/libc++How do i read bool value using std::istream
Why the difference in type-inference over the as-pattern in two similar function definitions?
Where do students learn to solve polynomial equations these days?
Running a General Election and the European Elections together
Do I need to write [sic] when a number is less than 10 but isn't written out?
How to write a definition with variants?
What did we know about the Kessel run before the prequels?
Grabbing quick drinks
Is it professional to write unrelated content in an almost-empty email?
How many extra stops do monopods offer for tele photographs?
Domestic-to-international connection at Orlando (MCO)
Would this house-rule that treats advantage as a +1 to the roll instead (and disadvantage as -1) and allows them to stack be balanced?
Prepend last line of stdin to entire stdin
Is it possible to replace duplicates of a character with one character using tr
Example of a Mathematician/Physicist whose Other Publications during their PhD eclipsed their PhD Thesis
Axiom Schema vs Axiom
0-rank tensor vs vector in 1D
Is the D&D universe the same as the Forgotten Realms universe?
How to edit “Name” property in GCI output?
Necessary condition on homology group for a set to be contractible
What connection does MS Office have to Netscape Navigator?
Why, when going from special to general relativity, do we just replace partial derivatives with covariant derivatives?
Can MTA send mail via a relay without being told so?
Why do remote US companies require working in the US?
Are police here, aren't itthey?
How to avoid run-time checks for running parts of code that become unreachable after compilation?
The Next CEO of Stack OverflowHow can I profile C++ code running on Linux?Counting the total of same running processes in C++Usage of this next_combination codeIdentify objects in boost::shared_ptr<boost::thread>How C++ reference worksNamespaces and the Pre-Processorstd::atomic_is_lock_free(shared_ptr<T>*) didn't compileCleaning data after exception on class constructorTwo-level bind fails on GCC/libstdc++ and Clang/libc++How do i read bool value using std::istream
My program gets a couple of Boolean variables from the user, and their values won't change afterwards. Each Boolean variable enables a part of code. Something like this:
#include <iostream>
void callback_function(bool task_1, bool task_2, bool task_3)
if (task_1)
std::cout << "Running task 1" << std::endl;
if (task_2)
std::cout << "Running task 2" << std::endl;
if (task_3)
std::cout << "Running task 3" << std::endl;
int main()
bool task_1 = true;
bool task_2 = false;
bool task_3 = true;
while (true)
callback_function(task_1, task_2, task_3);
return 0;
Now my question is, since the Boolean variables are fixed every time the program calls callback_function(), is there a way to avoid the if statements inside the callback function?
This is one way to avoid the run-time checks (implement a callback function for all permutations of the Boolean variables --- only two cases are shown below):
#include <functional>
#include <iostream>
void callback_function_for_tasks_1_2_3()
std::cout << "Running task 1" << std::endl;
std::cout << "Running task 2" << std::endl;
std::cout << "Running task 3" << std::endl;
void callback_function_for_tasks_1_3()
std::cout << "Running task 1" << std::endl;
std::cout << "Running task 3" << std::endl;
int main()
bool task_1 = true;
bool task_2 = false;
bool task_3 = true;
std::function<void()> callback_function;
if (task_1 && task_2 && task_3)
callback_function = callback_function_for_tasks_1_2_3;
else if (task_1 && !task_2 && task_3)
callback_function = callback_function_for_tasks_1_3;
while (true)
callback_function();
return 0;
The problem is I have to implement 2^n different callback functions, if there are n Boolean variables. Is there a better way to accomplish this?
c++
New contributor
Alireza Shafaei is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
|
show 4 more comments
My program gets a couple of Boolean variables from the user, and their values won't change afterwards. Each Boolean variable enables a part of code. Something like this:
#include <iostream>
void callback_function(bool task_1, bool task_2, bool task_3)
if (task_1)
std::cout << "Running task 1" << std::endl;
if (task_2)
std::cout << "Running task 2" << std::endl;
if (task_3)
std::cout << "Running task 3" << std::endl;
int main()
bool task_1 = true;
bool task_2 = false;
bool task_3 = true;
while (true)
callback_function(task_1, task_2, task_3);
return 0;
Now my question is, since the Boolean variables are fixed every time the program calls callback_function(), is there a way to avoid the if statements inside the callback function?
This is one way to avoid the run-time checks (implement a callback function for all permutations of the Boolean variables --- only two cases are shown below):
#include <functional>
#include <iostream>
void callback_function_for_tasks_1_2_3()
std::cout << "Running task 1" << std::endl;
std::cout << "Running task 2" << std::endl;
std::cout << "Running task 3" << std::endl;
void callback_function_for_tasks_1_3()
std::cout << "Running task 1" << std::endl;
std::cout << "Running task 3" << std::endl;
int main()
bool task_1 = true;
bool task_2 = false;
bool task_3 = true;
std::function<void()> callback_function;
if (task_1 && task_2 && task_3)
callback_function = callback_function_for_tasks_1_2_3;
else if (task_1 && !task_2 && task_3)
callback_function = callback_function_for_tasks_1_3;
while (true)
callback_function();
return 0;
The problem is I have to implement 2^n different callback functions, if there are n Boolean variables. Is there a better way to accomplish this?
c++
New contributor
Alireza Shafaei is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
5
If you’re interested in performance, don’t usestd::functionwhen a function pointer will do.
– Davis Herring
4 hours ago
5
Have you actually measured whether these conditional statements make a difference? This looks like a pretty pointless optimization effort to me. Or, if you're trying to solve an actual problem with this, it may be the wrong approach, a so-called "XY problem". Please, as a new user, also take the tour and read How to Ask.
– Ulrich Eckhardt
4 hours ago
If you go with with the 2nd approach, I believe, you will end up doing more checks then the 1st approch. Because each compound checks you are doing in the 2nd one will computationally cost you more than the 1st. I am not sure what are you trying to accomplish here, but if your concern is that for a false flag, the statements inside the block will take time to execute, then you don't have to worry about that. Because if the flag is false, the block will not take any execution time. And checking 1 by 1 will be cheaper than the combinations.
– ABM Ruman
4 hours ago
1
@ABMRuman He'd be doing more checks only once, not every time in the loop. If this is a long running application... One could safe quite a lot of checks if one combined the conditions inside an unsigned int/uint32_t/uint64_t (depending on number of checks) and select the function via aswitchstatement. The functions might be generated via a template function usingif constexprinside, so one wouldn't need to write all the functions explicitly.
– Aconcagua
4 hours ago
Joining @UlrichEckhardt: You should first run a profiler to find the hottest spots to optimise. Optimising the called functions at the right places will most likely bring you much more performance gain than avoiding these viewifs...
– Aconcagua
3 hours ago
|
show 4 more comments
My program gets a couple of Boolean variables from the user, and their values won't change afterwards. Each Boolean variable enables a part of code. Something like this:
#include <iostream>
void callback_function(bool task_1, bool task_2, bool task_3)
if (task_1)
std::cout << "Running task 1" << std::endl;
if (task_2)
std::cout << "Running task 2" << std::endl;
if (task_3)
std::cout << "Running task 3" << std::endl;
int main()
bool task_1 = true;
bool task_2 = false;
bool task_3 = true;
while (true)
callback_function(task_1, task_2, task_3);
return 0;
Now my question is, since the Boolean variables are fixed every time the program calls callback_function(), is there a way to avoid the if statements inside the callback function?
This is one way to avoid the run-time checks (implement a callback function for all permutations of the Boolean variables --- only two cases are shown below):
#include <functional>
#include <iostream>
void callback_function_for_tasks_1_2_3()
std::cout << "Running task 1" << std::endl;
std::cout << "Running task 2" << std::endl;
std::cout << "Running task 3" << std::endl;
void callback_function_for_tasks_1_3()
std::cout << "Running task 1" << std::endl;
std::cout << "Running task 3" << std::endl;
int main()
bool task_1 = true;
bool task_2 = false;
bool task_3 = true;
std::function<void()> callback_function;
if (task_1 && task_2 && task_3)
callback_function = callback_function_for_tasks_1_2_3;
else if (task_1 && !task_2 && task_3)
callback_function = callback_function_for_tasks_1_3;
while (true)
callback_function();
return 0;
The problem is I have to implement 2^n different callback functions, if there are n Boolean variables. Is there a better way to accomplish this?
c++
New contributor
Alireza Shafaei is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
My program gets a couple of Boolean variables from the user, and their values won't change afterwards. Each Boolean variable enables a part of code. Something like this:
#include <iostream>
void callback_function(bool task_1, bool task_2, bool task_3)
if (task_1)
std::cout << "Running task 1" << std::endl;
if (task_2)
std::cout << "Running task 2" << std::endl;
if (task_3)
std::cout << "Running task 3" << std::endl;
int main()
bool task_1 = true;
bool task_2 = false;
bool task_3 = true;
while (true)
callback_function(task_1, task_2, task_3);
return 0;
Now my question is, since the Boolean variables are fixed every time the program calls callback_function(), is there a way to avoid the if statements inside the callback function?
This is one way to avoid the run-time checks (implement a callback function for all permutations of the Boolean variables --- only two cases are shown below):
#include <functional>
#include <iostream>
void callback_function_for_tasks_1_2_3()
std::cout << "Running task 1" << std::endl;
std::cout << "Running task 2" << std::endl;
std::cout << "Running task 3" << std::endl;
void callback_function_for_tasks_1_3()
std::cout << "Running task 1" << std::endl;
std::cout << "Running task 3" << std::endl;
int main()
bool task_1 = true;
bool task_2 = false;
bool task_3 = true;
std::function<void()> callback_function;
if (task_1 && task_2 && task_3)
callback_function = callback_function_for_tasks_1_2_3;
else if (task_1 && !task_2 && task_3)
callback_function = callback_function_for_tasks_1_3;
while (true)
callback_function();
return 0;
The problem is I have to implement 2^n different callback functions, if there are n Boolean variables. Is there a better way to accomplish this?
c++
c++
New contributor
Alireza Shafaei is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
New contributor
Alireza Shafaei is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
New contributor
Alireza Shafaei is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
asked 4 hours ago
Alireza ShafaeiAlireza Shafaei
361
361
New contributor
Alireza Shafaei is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
New contributor
Alireza Shafaei is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
Alireza Shafaei is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
5
If you’re interested in performance, don’t usestd::functionwhen a function pointer will do.
– Davis Herring
4 hours ago
5
Have you actually measured whether these conditional statements make a difference? This looks like a pretty pointless optimization effort to me. Or, if you're trying to solve an actual problem with this, it may be the wrong approach, a so-called "XY problem". Please, as a new user, also take the tour and read How to Ask.
– Ulrich Eckhardt
4 hours ago
If you go with with the 2nd approach, I believe, you will end up doing more checks then the 1st approch. Because each compound checks you are doing in the 2nd one will computationally cost you more than the 1st. I am not sure what are you trying to accomplish here, but if your concern is that for a false flag, the statements inside the block will take time to execute, then you don't have to worry about that. Because if the flag is false, the block will not take any execution time. And checking 1 by 1 will be cheaper than the combinations.
– ABM Ruman
4 hours ago
1
@ABMRuman He'd be doing more checks only once, not every time in the loop. If this is a long running application... One could safe quite a lot of checks if one combined the conditions inside an unsigned int/uint32_t/uint64_t (depending on number of checks) and select the function via aswitchstatement. The functions might be generated via a template function usingif constexprinside, so one wouldn't need to write all the functions explicitly.
– Aconcagua
4 hours ago
Joining @UlrichEckhardt: You should first run a profiler to find the hottest spots to optimise. Optimising the called functions at the right places will most likely bring you much more performance gain than avoiding these viewifs...
– Aconcagua
3 hours ago
|
show 4 more comments
5
If you’re interested in performance, don’t usestd::functionwhen a function pointer will do.
– Davis Herring
4 hours ago
5
Have you actually measured whether these conditional statements make a difference? This looks like a pretty pointless optimization effort to me. Or, if you're trying to solve an actual problem with this, it may be the wrong approach, a so-called "XY problem". Please, as a new user, also take the tour and read How to Ask.
– Ulrich Eckhardt
4 hours ago
If you go with with the 2nd approach, I believe, you will end up doing more checks then the 1st approch. Because each compound checks you are doing in the 2nd one will computationally cost you more than the 1st. I am not sure what are you trying to accomplish here, but if your concern is that for a false flag, the statements inside the block will take time to execute, then you don't have to worry about that. Because if the flag is false, the block will not take any execution time. And checking 1 by 1 will be cheaper than the combinations.
– ABM Ruman
4 hours ago
1
@ABMRuman He'd be doing more checks only once, not every time in the loop. If this is a long running application... One could safe quite a lot of checks if one combined the conditions inside an unsigned int/uint32_t/uint64_t (depending on number of checks) and select the function via aswitchstatement. The functions might be generated via a template function usingif constexprinside, so one wouldn't need to write all the functions explicitly.
– Aconcagua
4 hours ago
Joining @UlrichEckhardt: You should first run a profiler to find the hottest spots to optimise. Optimising the called functions at the right places will most likely bring you much more performance gain than avoiding these viewifs...
– Aconcagua
3 hours ago
5
5
If you’re interested in performance, don’t use
std::function when a function pointer will do.– Davis Herring
4 hours ago
If you’re interested in performance, don’t use
std::function when a function pointer will do.– Davis Herring
4 hours ago
5
5
Have you actually measured whether these conditional statements make a difference? This looks like a pretty pointless optimization effort to me. Or, if you're trying to solve an actual problem with this, it may be the wrong approach, a so-called "XY problem". Please, as a new user, also take the tour and read How to Ask.
– Ulrich Eckhardt
4 hours ago
Have you actually measured whether these conditional statements make a difference? This looks like a pretty pointless optimization effort to me. Or, if you're trying to solve an actual problem with this, it may be the wrong approach, a so-called "XY problem". Please, as a new user, also take the tour and read How to Ask.
– Ulrich Eckhardt
4 hours ago
If you go with with the 2nd approach, I believe, you will end up doing more checks then the 1st approch. Because each compound checks you are doing in the 2nd one will computationally cost you more than the 1st. I am not sure what are you trying to accomplish here, but if your concern is that for a false flag, the statements inside the block will take time to execute, then you don't have to worry about that. Because if the flag is false, the block will not take any execution time. And checking 1 by 1 will be cheaper than the combinations.
– ABM Ruman
4 hours ago
If you go with with the 2nd approach, I believe, you will end up doing more checks then the 1st approch. Because each compound checks you are doing in the 2nd one will computationally cost you more than the 1st. I am not sure what are you trying to accomplish here, but if your concern is that for a false flag, the statements inside the block will take time to execute, then you don't have to worry about that. Because if the flag is false, the block will not take any execution time. And checking 1 by 1 will be cheaper than the combinations.
– ABM Ruman
4 hours ago
1
1
@ABMRuman He'd be doing more checks only once, not every time in the loop. If this is a long running application... One could safe quite a lot of checks if one combined the conditions inside an unsigned int/uint32_t/uint64_t (depending on number of checks) and select the function via a
switch statement. The functions might be generated via a template function using if constexpr inside, so one wouldn't need to write all the functions explicitly.– Aconcagua
4 hours ago
@ABMRuman He'd be doing more checks only once, not every time in the loop. If this is a long running application... One could safe quite a lot of checks if one combined the conditions inside an unsigned int/uint32_t/uint64_t (depending on number of checks) and select the function via a
switch statement. The functions might be generated via a template function using if constexpr inside, so one wouldn't need to write all the functions explicitly.– Aconcagua
4 hours ago
Joining @UlrichEckhardt: You should first run a profiler to find the hottest spots to optimise. Optimising the called functions at the right places will most likely bring you much more performance gain than avoiding these view
ifs...– Aconcagua
3 hours ago
Joining @UlrichEckhardt: You should first run a profiler to find the hottest spots to optimise. Optimising the called functions at the right places will most likely bring you much more performance gain than avoiding these view
ifs...– Aconcagua
3 hours ago
|
show 4 more comments
2 Answers
2
active
oldest
votes
Ensuring that if statements are evaluated at compile time
C++17 introduces if constexpr, which does exactly this:
template<bool task_1, bool task_2, bool task_3>
void callback_function()
if constexpr (task_1)
std::cout << "Running task 1" << std::endl;
if constexpr (task_2)
std::cout << "Running task 2" << std::endl;
if constexpr (task_3)
std::cout << "Running task 3" << std::endl;
If you have optimizations enabled, if constexpr isn't necessary. Even if you use a regular if instead of if constexpr, because the bools are now templated, the compiler will be able to eliminate the if statements entirely, and just run the tasks. If you look at the assembly produced here, you'll see that even at -O1, there are no if statements in any of the callback functions.
We can now use callback_function directly as a function pointer, avoiding function<void()>:
int main()
using callback_t = void(*)();
callback_t func = callback_function<true, false, true>;
// Do stuff with func
We can also name the bools by assigning them to constexpr variables:
int main()
using callback_t = void(*)();
constexpr bool do_task1 = true;
constexpr bool do_task2 = false;
constexpr bool do_task3 = true;
callback_t func = callback_function<do_task1, do_task2, do_task3>;
// Do stuff with func
Automatically creating a lookup table of all possible callback functions
You mentioned choosing between different callback functions at runtime. We can do this pretty easily with a lookup table, and we can use templates to automatically create a lookup table of all possible callback functions.
The first step is to get a callback function from a particular index:
// void(*)() is ugly to type, so I alias it
using callback_t = void(*)();
// Unpacks the bits
template<size_t index>
constexpr auto getCallbackFromIndex() -> callback_t
constexpr bool do_task1 = (index & 4) != 0;
constexpr bool do_task2 = (index & 2) != 0;
constexpr bool do_task3 = (index & 1) != 0;
return callback_function<do_task1, do_task2, do_task3>;
Once we can do that, we can write a function to create a lookup table from a bunch of indexes. Our lookup table will just be a std::array.
// Create a std::array based on a list of flags
// See https://en.cppreference.com/w/cpp/utility/integer_sequence
// For more information
template<size_t... Indexes>
constexpr auto getVersionLookup(std::index_sequence<Indexes...>)
-> std::array<callback_t, sizeof...(Indexes)>
return getCallbackFromIndex<Indexes>()...;
// Makes a lookup table containing all 8 possible callback functions
constexpr auto callbackLookupTable =
getVersionLookup(std::make_index_sequence<8>());
Here, callbackLookupTable contains all 8 possible callback functions, where callbackLookupTable[i] expands the bits of i to get the callback. For example, if i == 6, then i's bits are 110 in binary, so
callbackLookupTable[6] is callback_function<true, true, false>
Using the lookup table at runtime
Using the lookup table is really simple. We can get an index from a bunch of bools by bitshifting:
callback_t getCallbackBasedOnTasks(bool task1, bool task2, bool task3)
// Get the index based on bit shifting
int index = ((int)task1 << 2) + ((int)task2 << 1) + ((int)task3);
// return the correct callback
return callbackLookupTable[index];
Example demonstrating how to read in tasks
We can get the bools at runtime now, and just call getCallbackBasedOnTasks to get the correct callback
int main()
bool t1, t2, t3;
// Read in bools
std::cin >> t1 >> t2 >> t3;
// Get the callback
callback_t func = getCallbackBasedOnTasks(t1, t2, t3);
// Invoke the callback
func();
Theconstexpris IMHO a red herring, the important part is the use of template parameters which are evaluated at compile time. Any halfway-decent compiler is able to figure this out even without theseconsexprhints.
– Ulrich Eckhardt
3 hours ago
I updated the answer explaining how to automatically generate a list of all possible callback functions. You can find the right one just by packing the bools into the bits of the index
– Jorge Perez
3 hours ago
The lookup table is a runtime solution. You can get the right function just by indexing into it
– Jorge Perez
3 hours ago
I added code showing exactly how to do that
– Jorge Perez
3 hours ago
1
(@$#*!) – can't find anything to criticise any more... Just kidding. But the automatically created lookup table is great, far better than my originally proposedswitchstatement...
– Aconcagua
3 hours ago
|
show 1 more comment
Short of JIT compilation, you can’t do better than your 2^n functions, although you can of course use a template to avoid writing them all out. You still need a (massive) if-else tree to dispatch to the correct implementation, so the source (not just the binary) still scales exponentially with the number of variables.
If you know something at compiletime, it's easy to template it. Especially in the OP's case.
– Jorge Perez
4 hours ago
add a comment |
StackExchange.ifUsing("editor", function ()
StackExchange.using("externalEditor", function ()
StackExchange.using("snippets", function ()
StackExchange.snippets.init();
);
);
, "code-snippets");
StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "1"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);
else
createEditor();
);
function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);
);
Alireza Shafaei is a new contributor. Be nice, and check out our Code of Conduct.
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f55438306%2fhow-to-avoid-run-time-checks-for-running-parts-of-code-that-become-unreachable-a%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
2 Answers
2
active
oldest
votes
2 Answers
2
active
oldest
votes
active
oldest
votes
active
oldest
votes
Ensuring that if statements are evaluated at compile time
C++17 introduces if constexpr, which does exactly this:
template<bool task_1, bool task_2, bool task_3>
void callback_function()
if constexpr (task_1)
std::cout << "Running task 1" << std::endl;
if constexpr (task_2)
std::cout << "Running task 2" << std::endl;
if constexpr (task_3)
std::cout << "Running task 3" << std::endl;
If you have optimizations enabled, if constexpr isn't necessary. Even if you use a regular if instead of if constexpr, because the bools are now templated, the compiler will be able to eliminate the if statements entirely, and just run the tasks. If you look at the assembly produced here, you'll see that even at -O1, there are no if statements in any of the callback functions.
We can now use callback_function directly as a function pointer, avoiding function<void()>:
int main()
using callback_t = void(*)();
callback_t func = callback_function<true, false, true>;
// Do stuff with func
We can also name the bools by assigning them to constexpr variables:
int main()
using callback_t = void(*)();
constexpr bool do_task1 = true;
constexpr bool do_task2 = false;
constexpr bool do_task3 = true;
callback_t func = callback_function<do_task1, do_task2, do_task3>;
// Do stuff with func
Automatically creating a lookup table of all possible callback functions
You mentioned choosing between different callback functions at runtime. We can do this pretty easily with a lookup table, and we can use templates to automatically create a lookup table of all possible callback functions.
The first step is to get a callback function from a particular index:
// void(*)() is ugly to type, so I alias it
using callback_t = void(*)();
// Unpacks the bits
template<size_t index>
constexpr auto getCallbackFromIndex() -> callback_t
constexpr bool do_task1 = (index & 4) != 0;
constexpr bool do_task2 = (index & 2) != 0;
constexpr bool do_task3 = (index & 1) != 0;
return callback_function<do_task1, do_task2, do_task3>;
Once we can do that, we can write a function to create a lookup table from a bunch of indexes. Our lookup table will just be a std::array.
// Create a std::array based on a list of flags
// See https://en.cppreference.com/w/cpp/utility/integer_sequence
// For more information
template<size_t... Indexes>
constexpr auto getVersionLookup(std::index_sequence<Indexes...>)
-> std::array<callback_t, sizeof...(Indexes)>
return getCallbackFromIndex<Indexes>()...;
// Makes a lookup table containing all 8 possible callback functions
constexpr auto callbackLookupTable =
getVersionLookup(std::make_index_sequence<8>());
Here, callbackLookupTable contains all 8 possible callback functions, where callbackLookupTable[i] expands the bits of i to get the callback. For example, if i == 6, then i's bits are 110 in binary, so
callbackLookupTable[6] is callback_function<true, true, false>
Using the lookup table at runtime
Using the lookup table is really simple. We can get an index from a bunch of bools by bitshifting:
callback_t getCallbackBasedOnTasks(bool task1, bool task2, bool task3)
// Get the index based on bit shifting
int index = ((int)task1 << 2) + ((int)task2 << 1) + ((int)task3);
// return the correct callback
return callbackLookupTable[index];
Example demonstrating how to read in tasks
We can get the bools at runtime now, and just call getCallbackBasedOnTasks to get the correct callback
int main()
bool t1, t2, t3;
// Read in bools
std::cin >> t1 >> t2 >> t3;
// Get the callback
callback_t func = getCallbackBasedOnTasks(t1, t2, t3);
// Invoke the callback
func();
Theconstexpris IMHO a red herring, the important part is the use of template parameters which are evaluated at compile time. Any halfway-decent compiler is able to figure this out even without theseconsexprhints.
– Ulrich Eckhardt
3 hours ago
I updated the answer explaining how to automatically generate a list of all possible callback functions. You can find the right one just by packing the bools into the bits of the index
– Jorge Perez
3 hours ago
The lookup table is a runtime solution. You can get the right function just by indexing into it
– Jorge Perez
3 hours ago
I added code showing exactly how to do that
– Jorge Perez
3 hours ago
1
(@$#*!) – can't find anything to criticise any more... Just kidding. But the automatically created lookup table is great, far better than my originally proposedswitchstatement...
– Aconcagua
3 hours ago
|
show 1 more comment
Ensuring that if statements are evaluated at compile time
C++17 introduces if constexpr, which does exactly this:
template<bool task_1, bool task_2, bool task_3>
void callback_function()
if constexpr (task_1)
std::cout << "Running task 1" << std::endl;
if constexpr (task_2)
std::cout << "Running task 2" << std::endl;
if constexpr (task_3)
std::cout << "Running task 3" << std::endl;
If you have optimizations enabled, if constexpr isn't necessary. Even if you use a regular if instead of if constexpr, because the bools are now templated, the compiler will be able to eliminate the if statements entirely, and just run the tasks. If you look at the assembly produced here, you'll see that even at -O1, there are no if statements in any of the callback functions.
We can now use callback_function directly as a function pointer, avoiding function<void()>:
int main()
using callback_t = void(*)();
callback_t func = callback_function<true, false, true>;
// Do stuff with func
We can also name the bools by assigning them to constexpr variables:
int main()
using callback_t = void(*)();
constexpr bool do_task1 = true;
constexpr bool do_task2 = false;
constexpr bool do_task3 = true;
callback_t func = callback_function<do_task1, do_task2, do_task3>;
// Do stuff with func
Automatically creating a lookup table of all possible callback functions
You mentioned choosing between different callback functions at runtime. We can do this pretty easily with a lookup table, and we can use templates to automatically create a lookup table of all possible callback functions.
The first step is to get a callback function from a particular index:
// void(*)() is ugly to type, so I alias it
using callback_t = void(*)();
// Unpacks the bits
template<size_t index>
constexpr auto getCallbackFromIndex() -> callback_t
constexpr bool do_task1 = (index & 4) != 0;
constexpr bool do_task2 = (index & 2) != 0;
constexpr bool do_task3 = (index & 1) != 0;
return callback_function<do_task1, do_task2, do_task3>;
Once we can do that, we can write a function to create a lookup table from a bunch of indexes. Our lookup table will just be a std::array.
// Create a std::array based on a list of flags
// See https://en.cppreference.com/w/cpp/utility/integer_sequence
// For more information
template<size_t... Indexes>
constexpr auto getVersionLookup(std::index_sequence<Indexes...>)
-> std::array<callback_t, sizeof...(Indexes)>
return getCallbackFromIndex<Indexes>()...;
// Makes a lookup table containing all 8 possible callback functions
constexpr auto callbackLookupTable =
getVersionLookup(std::make_index_sequence<8>());
Here, callbackLookupTable contains all 8 possible callback functions, where callbackLookupTable[i] expands the bits of i to get the callback. For example, if i == 6, then i's bits are 110 in binary, so
callbackLookupTable[6] is callback_function<true, true, false>
Using the lookup table at runtime
Using the lookup table is really simple. We can get an index from a bunch of bools by bitshifting:
callback_t getCallbackBasedOnTasks(bool task1, bool task2, bool task3)
// Get the index based on bit shifting
int index = ((int)task1 << 2) + ((int)task2 << 1) + ((int)task3);
// return the correct callback
return callbackLookupTable[index];
Example demonstrating how to read in tasks
We can get the bools at runtime now, and just call getCallbackBasedOnTasks to get the correct callback
int main()
bool t1, t2, t3;
// Read in bools
std::cin >> t1 >> t2 >> t3;
// Get the callback
callback_t func = getCallbackBasedOnTasks(t1, t2, t3);
// Invoke the callback
func();
Theconstexpris IMHO a red herring, the important part is the use of template parameters which are evaluated at compile time. Any halfway-decent compiler is able to figure this out even without theseconsexprhints.
– Ulrich Eckhardt
3 hours ago
I updated the answer explaining how to automatically generate a list of all possible callback functions. You can find the right one just by packing the bools into the bits of the index
– Jorge Perez
3 hours ago
The lookup table is a runtime solution. You can get the right function just by indexing into it
– Jorge Perez
3 hours ago
I added code showing exactly how to do that
– Jorge Perez
3 hours ago
1
(@$#*!) – can't find anything to criticise any more... Just kidding. But the automatically created lookup table is great, far better than my originally proposedswitchstatement...
– Aconcagua
3 hours ago
|
show 1 more comment
Ensuring that if statements are evaluated at compile time
C++17 introduces if constexpr, which does exactly this:
template<bool task_1, bool task_2, bool task_3>
void callback_function()
if constexpr (task_1)
std::cout << "Running task 1" << std::endl;
if constexpr (task_2)
std::cout << "Running task 2" << std::endl;
if constexpr (task_3)
std::cout << "Running task 3" << std::endl;
If you have optimizations enabled, if constexpr isn't necessary. Even if you use a regular if instead of if constexpr, because the bools are now templated, the compiler will be able to eliminate the if statements entirely, and just run the tasks. If you look at the assembly produced here, you'll see that even at -O1, there are no if statements in any of the callback functions.
We can now use callback_function directly as a function pointer, avoiding function<void()>:
int main()
using callback_t = void(*)();
callback_t func = callback_function<true, false, true>;
// Do stuff with func
We can also name the bools by assigning them to constexpr variables:
int main()
using callback_t = void(*)();
constexpr bool do_task1 = true;
constexpr bool do_task2 = false;
constexpr bool do_task3 = true;
callback_t func = callback_function<do_task1, do_task2, do_task3>;
// Do stuff with func
Automatically creating a lookup table of all possible callback functions
You mentioned choosing between different callback functions at runtime. We can do this pretty easily with a lookup table, and we can use templates to automatically create a lookup table of all possible callback functions.
The first step is to get a callback function from a particular index:
// void(*)() is ugly to type, so I alias it
using callback_t = void(*)();
// Unpacks the bits
template<size_t index>
constexpr auto getCallbackFromIndex() -> callback_t
constexpr bool do_task1 = (index & 4) != 0;
constexpr bool do_task2 = (index & 2) != 0;
constexpr bool do_task3 = (index & 1) != 0;
return callback_function<do_task1, do_task2, do_task3>;
Once we can do that, we can write a function to create a lookup table from a bunch of indexes. Our lookup table will just be a std::array.
// Create a std::array based on a list of flags
// See https://en.cppreference.com/w/cpp/utility/integer_sequence
// For more information
template<size_t... Indexes>
constexpr auto getVersionLookup(std::index_sequence<Indexes...>)
-> std::array<callback_t, sizeof...(Indexes)>
return getCallbackFromIndex<Indexes>()...;
// Makes a lookup table containing all 8 possible callback functions
constexpr auto callbackLookupTable =
getVersionLookup(std::make_index_sequence<8>());
Here, callbackLookupTable contains all 8 possible callback functions, where callbackLookupTable[i] expands the bits of i to get the callback. For example, if i == 6, then i's bits are 110 in binary, so
callbackLookupTable[6] is callback_function<true, true, false>
Using the lookup table at runtime
Using the lookup table is really simple. We can get an index from a bunch of bools by bitshifting:
callback_t getCallbackBasedOnTasks(bool task1, bool task2, bool task3)
// Get the index based on bit shifting
int index = ((int)task1 << 2) + ((int)task2 << 1) + ((int)task3);
// return the correct callback
return callbackLookupTable[index];
Example demonstrating how to read in tasks
We can get the bools at runtime now, and just call getCallbackBasedOnTasks to get the correct callback
int main()
bool t1, t2, t3;
// Read in bools
std::cin >> t1 >> t2 >> t3;
// Get the callback
callback_t func = getCallbackBasedOnTasks(t1, t2, t3);
// Invoke the callback
func();
Ensuring that if statements are evaluated at compile time
C++17 introduces if constexpr, which does exactly this:
template<bool task_1, bool task_2, bool task_3>
void callback_function()
if constexpr (task_1)
std::cout << "Running task 1" << std::endl;
if constexpr (task_2)
std::cout << "Running task 2" << std::endl;
if constexpr (task_3)
std::cout << "Running task 3" << std::endl;
If you have optimizations enabled, if constexpr isn't necessary. Even if you use a regular if instead of if constexpr, because the bools are now templated, the compiler will be able to eliminate the if statements entirely, and just run the tasks. If you look at the assembly produced here, you'll see that even at -O1, there are no if statements in any of the callback functions.
We can now use callback_function directly as a function pointer, avoiding function<void()>:
int main()
using callback_t = void(*)();
callback_t func = callback_function<true, false, true>;
// Do stuff with func
We can also name the bools by assigning them to constexpr variables:
int main()
using callback_t = void(*)();
constexpr bool do_task1 = true;
constexpr bool do_task2 = false;
constexpr bool do_task3 = true;
callback_t func = callback_function<do_task1, do_task2, do_task3>;
// Do stuff with func
Automatically creating a lookup table of all possible callback functions
You mentioned choosing between different callback functions at runtime. We can do this pretty easily with a lookup table, and we can use templates to automatically create a lookup table of all possible callback functions.
The first step is to get a callback function from a particular index:
// void(*)() is ugly to type, so I alias it
using callback_t = void(*)();
// Unpacks the bits
template<size_t index>
constexpr auto getCallbackFromIndex() -> callback_t
constexpr bool do_task1 = (index & 4) != 0;
constexpr bool do_task2 = (index & 2) != 0;
constexpr bool do_task3 = (index & 1) != 0;
return callback_function<do_task1, do_task2, do_task3>;
Once we can do that, we can write a function to create a lookup table from a bunch of indexes. Our lookup table will just be a std::array.
// Create a std::array based on a list of flags
// See https://en.cppreference.com/w/cpp/utility/integer_sequence
// For more information
template<size_t... Indexes>
constexpr auto getVersionLookup(std::index_sequence<Indexes...>)
-> std::array<callback_t, sizeof...(Indexes)>
return getCallbackFromIndex<Indexes>()...;
// Makes a lookup table containing all 8 possible callback functions
constexpr auto callbackLookupTable =
getVersionLookup(std::make_index_sequence<8>());
Here, callbackLookupTable contains all 8 possible callback functions, where callbackLookupTable[i] expands the bits of i to get the callback. For example, if i == 6, then i's bits are 110 in binary, so
callbackLookupTable[6] is callback_function<true, true, false>
Using the lookup table at runtime
Using the lookup table is really simple. We can get an index from a bunch of bools by bitshifting:
callback_t getCallbackBasedOnTasks(bool task1, bool task2, bool task3)
// Get the index based on bit shifting
int index = ((int)task1 << 2) + ((int)task2 << 1) + ((int)task3);
// return the correct callback
return callbackLookupTable[index];
Example demonstrating how to read in tasks
We can get the bools at runtime now, and just call getCallbackBasedOnTasks to get the correct callback
int main()
bool t1, t2, t3;
// Read in bools
std::cin >> t1 >> t2 >> t3;
// Get the callback
callback_t func = getCallbackBasedOnTasks(t1, t2, t3);
// Invoke the callback
func();
edited 2 hours ago
answered 4 hours ago
Jorge PerezJorge Perez
1,042414
1,042414
Theconstexpris IMHO a red herring, the important part is the use of template parameters which are evaluated at compile time. Any halfway-decent compiler is able to figure this out even without theseconsexprhints.
– Ulrich Eckhardt
3 hours ago
I updated the answer explaining how to automatically generate a list of all possible callback functions. You can find the right one just by packing the bools into the bits of the index
– Jorge Perez
3 hours ago
The lookup table is a runtime solution. You can get the right function just by indexing into it
– Jorge Perez
3 hours ago
I added code showing exactly how to do that
– Jorge Perez
3 hours ago
1
(@$#*!) – can't find anything to criticise any more... Just kidding. But the automatically created lookup table is great, far better than my originally proposedswitchstatement...
– Aconcagua
3 hours ago
|
show 1 more comment
Theconstexpris IMHO a red herring, the important part is the use of template parameters which are evaluated at compile time. Any halfway-decent compiler is able to figure this out even without theseconsexprhints.
– Ulrich Eckhardt
3 hours ago
I updated the answer explaining how to automatically generate a list of all possible callback functions. You can find the right one just by packing the bools into the bits of the index
– Jorge Perez
3 hours ago
The lookup table is a runtime solution. You can get the right function just by indexing into it
– Jorge Perez
3 hours ago
I added code showing exactly how to do that
– Jorge Perez
3 hours ago
1
(@$#*!) – can't find anything to criticise any more... Just kidding. But the automatically created lookup table is great, far better than my originally proposedswitchstatement...
– Aconcagua
3 hours ago
The
constexpr is IMHO a red herring, the important part is the use of template parameters which are evaluated at compile time. Any halfway-decent compiler is able to figure this out even without these consexpr hints.– Ulrich Eckhardt
3 hours ago
The
constexpr is IMHO a red herring, the important part is the use of template parameters which are evaluated at compile time. Any halfway-decent compiler is able to figure this out even without these consexpr hints.– Ulrich Eckhardt
3 hours ago
I updated the answer explaining how to automatically generate a list of all possible callback functions. You can find the right one just by packing the bools into the bits of the index
– Jorge Perez
3 hours ago
I updated the answer explaining how to automatically generate a list of all possible callback functions. You can find the right one just by packing the bools into the bits of the index
– Jorge Perez
3 hours ago
The lookup table is a runtime solution. You can get the right function just by indexing into it
– Jorge Perez
3 hours ago
The lookup table is a runtime solution. You can get the right function just by indexing into it
– Jorge Perez
3 hours ago
I added code showing exactly how to do that
– Jorge Perez
3 hours ago
I added code showing exactly how to do that
– Jorge Perez
3 hours ago
1
1
(@$#*!) – can't find anything to criticise any more... Just kidding. But the automatically created lookup table is great, far better than my originally proposed
switch statement...– Aconcagua
3 hours ago
(@$#*!) – can't find anything to criticise any more... Just kidding. But the automatically created lookup table is great, far better than my originally proposed
switch statement...– Aconcagua
3 hours ago
|
show 1 more comment
Short of JIT compilation, you can’t do better than your 2^n functions, although you can of course use a template to avoid writing them all out. You still need a (massive) if-else tree to dispatch to the correct implementation, so the source (not just the binary) still scales exponentially with the number of variables.
If you know something at compiletime, it's easy to template it. Especially in the OP's case.
– Jorge Perez
4 hours ago
add a comment |
Short of JIT compilation, you can’t do better than your 2^n functions, although you can of course use a template to avoid writing them all out. You still need a (massive) if-else tree to dispatch to the correct implementation, so the source (not just the binary) still scales exponentially with the number of variables.
If you know something at compiletime, it's easy to template it. Especially in the OP's case.
– Jorge Perez
4 hours ago
add a comment |
Short of JIT compilation, you can’t do better than your 2^n functions, although you can of course use a template to avoid writing them all out. You still need a (massive) if-else tree to dispatch to the correct implementation, so the source (not just the binary) still scales exponentially with the number of variables.
Short of JIT compilation, you can’t do better than your 2^n functions, although you can of course use a template to avoid writing them all out. You still need a (massive) if-else tree to dispatch to the correct implementation, so the source (not just the binary) still scales exponentially with the number of variables.
answered 4 hours ago
Davis HerringDavis Herring
8,8871736
8,8871736
If you know something at compiletime, it's easy to template it. Especially in the OP's case.
– Jorge Perez
4 hours ago
add a comment |
If you know something at compiletime, it's easy to template it. Especially in the OP's case.
– Jorge Perez
4 hours ago
If you know something at compiletime, it's easy to template it. Especially in the OP's case.
– Jorge Perez
4 hours ago
If you know something at compiletime, it's easy to template it. Especially in the OP's case.
– Jorge Perez
4 hours ago
add a comment |
Alireza Shafaei is a new contributor. Be nice, and check out our Code of Conduct.
Alireza Shafaei is a new contributor. Be nice, and check out our Code of Conduct.
Alireza Shafaei is a new contributor. Be nice, and check out our Code of Conduct.
Alireza Shafaei is a new contributor. Be nice, and check out our Code of Conduct.
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f55438306%2fhow-to-avoid-run-time-checks-for-running-parts-of-code-that-become-unreachable-a%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown

5
If you’re interested in performance, don’t use
std::functionwhen a function pointer will do.– Davis Herring
4 hours ago
5
Have you actually measured whether these conditional statements make a difference? This looks like a pretty pointless optimization effort to me. Or, if you're trying to solve an actual problem with this, it may be the wrong approach, a so-called "XY problem". Please, as a new user, also take the tour and read How to Ask.
– Ulrich Eckhardt
4 hours ago
If you go with with the 2nd approach, I believe, you will end up doing more checks then the 1st approch. Because each compound checks you are doing in the 2nd one will computationally cost you more than the 1st. I am not sure what are you trying to accomplish here, but if your concern is that for a false flag, the statements inside the block will take time to execute, then you don't have to worry about that. Because if the flag is false, the block will not take any execution time. And checking 1 by 1 will be cheaper than the combinations.
– ABM Ruman
4 hours ago
1
@ABMRuman He'd be doing more checks only once, not every time in the loop. If this is a long running application... One could safe quite a lot of checks if one combined the conditions inside an unsigned int/uint32_t/uint64_t (depending on number of checks) and select the function via a
switchstatement. The functions might be generated via a template function usingif constexprinside, so one wouldn't need to write all the functions explicitly.– Aconcagua
4 hours ago
Joining @UlrichEckhardt: You should first run a profiler to find the hottest spots to optimise. Optimising the called functions at the right places will most likely bring you much more performance gain than avoiding these view
ifs...– Aconcagua
3 hours ago