I implemented the same algorithm both on CPU and GPU using C++ and CUDA C. In order to check if the results are correct I check if the 2 arrays of double calculated by both are the same with a precision of 1.0E-8 . And the result is that the 2 arrays are not the same. Is it possible to have different results because I'm working on two different entity or there might be some errors in my code?
I have to increase to 1.0E-3 to consider them equal!
Unless dev_s is thread local, you're going to have multiple threads writing to the same memory location here:
dev_s[i] = dev_s[i-1] + 0.5*(dev_matrix[index]+dev_matrix[index-1]);
Change this: [MarshalAs(UnmanagedType.ByValTStr, SizeConst = 36)] private string iu; to this: [MarshalAs(UnmanagedType.LPStr)] private string iu; Note that this code is good only to pass a string in the C#->C++ direction. For the opposite direction (C++->C#) it is more complex, because C# can't easily deallocate C++ allocated memory. Other important thing:...
Plenty of solutions are possible. A geometric approach would detect that the one moving blob is too big to be a single passenger car. Still, this may indicate a car with a caravan. That leads us to another question: if you have two blobs moving close together, how do you...
c++,templates,c++11,metaprogramming
You can do this without recursion by simply expanding the parameter pack directly into a std::tuple: template<My_enum... Enums> struct Tuple { using type = std::tuple<typename Bind_type<Enums>::type...>; }; To answer your question more directly, you can declare a variadic primary template, then write two specializations: for when there are at least...
your code works for me. But you used cv::waitKey(0) which means that the program waits there until you press a keyboard key. So try pressing a key after drawing, or use cv::waitKey(30) instead. If this doesnt help you, please add some std::cout in your callback function to verify it is...
c++,string,c++11,memory,standards
Section 21.4.1.5 of the 2011 standard states: The char-like objects in a basic_string object shall be stored contiguously. That is, for any basic_string object s, the identity &*(s.begin() + n) == &*s.begin() + n shall hold for all values of n such that 0 <= n < s.size(). The two...
c++,algorithm,parallel-processing,c++14
Parallel prefix sum is a classical distributed programming algorithm, which elegantly uses a reduction followed by a distribution (as illustrated in the article). The key observation is that you can compute parts of the partial sums before you know the leading terms.
It is very bad, accessing deleted objects as if they were not deleted will in the general case crash. There is no guarantee that the memory is still mapped inside the process and it could result in a virtual memory page fault. It is also likely that the memory will...
c++,templates,generic-programming
This depends on what you want the behaviour (protocol) of your class to be. Since you're logging into the error stream there, I assume you consider this an error condition to call pop() on an empty stack. The standard C++ way of signalling errors is to throw an exception. Something...
Basically, you are finding all permutations of the array using a recursive permutation algorithm. There are 4 things you need to change: First, start your loop from pos, not 0 Second, swap elements back after recursing (backtracking) Third, only test once you have generated each complete permutation (when pos =...
c++,templates,template-specialization
The specialization still needs to be a template template argument. You passed in a full type. You want: template <class Type, class Engine> class random_gen<std::uniform_real_distribution, Type, Engine> { ... }; Just std::uniform_real_distribution, not std::uniform_distribution<Type>. ...
c++,templates,constructor,explicit-instantiation
When the constructor is a template member function, they are not instantiated unless explicitly used. You would see the code for the constructor if you make it a non-template member function. template<typename T> class test { public: /*** template<typename T> test(T param) { parameter = param; }; ***/ test(T param)...
c++,opengl,opengl-es,integer,shader
These integers are handles.This is a common idiom used by many APIs, used to hide resource access through an opaque level of indirection. OpenGL is effectively preventing you from accessing what lies behind the handle without using the API calls. From Wikipedia: In computer programming, a handle is an abstract...
Speaking as someone who's had to do exactly what you're talking about a number of time, rr got it basically right, but I would change the emphasis a little. For file versioning, text is basically the winner. Since you're using an hdf5 library, I assume both serializing and parsing are...
This code works fine for a right angled triangle - * ** *** But I guess you want a triangle like this - * *** ***** Try this - #include <iostream> using namespace std; int main() { int i, j, k, n; cout << "Please enter number of rows you...
Your code makes no sense, why are you passing someStruct twice? For the reference part, you should have something like: void names(someStruct &s) { // <<<< Pass struct once as a reference cout << "First Name: " << "\n"; cin >> s.firstname; cout << "Last Name: " << "\n"; cin...
c++,polar-coordinates,cartesian-coordinates
You are converting to cartesian the points which are in cartesian already. What you want is: std::cout << "Cartesian Coordinates:" << std::endl; std::cout << to_cartesian(to_polar(a)) << std::endl; std::cout << to_cartesian(to_polar(b)) << std::endl; //... Edit: using atan2 solves the NaN problem, (0, 0) is converted to (0, 0) which is fine....
Your first problem is C++ name mangling. If you run nm on your .so file you will get something like this: nm test.so 0000000000000f40 T __Z3funv U _printf U dyld_stub_binder If you mark it as C style when compiled with C++: #ifdef __cplusplus extern "C" char fun() #else char fun(void)...
The simplest thing you can do is to use a for/while loop. A loop will basically repeat the same instruction for a number of n steps or until a certain condition is matched. The solution provided is pretty dummy, if you want to read the first name and last name...
The valid range of indices of an array with N elements is [0, N-1]. Thus instead of for example this loop for (int i=1; i <= n; i++) ^^^^ ^^^^^^ you have to write for ( int i = 0; i < n; i++ ) As you used operator new...
What you're trying to do makes little sense. We have subclass<int>. It is convertible to int&, but also to a lot of other reference types. char&. bool&. double&. The ambiguity arises from the fact that all the various overloads for operator<< that take any non-template argument are viable overload candidates...
Your error is actually coming from: array.push_back(day); This tries to put a copy of day in the vector, which is not permitted since it is unique. Instead you could write array.push_back( std::move(day) ); however the following would be better, replacing auto day...: array.emplace_back(); ...
Mentioned solution with fseek is good. However, it can be very slow for large matrices (as disks don't like random access, especially very far away). To speed up things, you should use blocking. I'll show a basic concept, and can explain it further if you need. First, you split your...
If you can use boost library you could simple do it like this: string date("2015-11-12"); string format("%Y-%m-%d"); date parsedDate = parser.parse_date(date, format, svp); You can read more about this here. If you want a pure C++ solution you can try using struct tm tm; std::string s("2015-11-123"); if (strptime(s.c_str(), "%Y-%m-%d", &tm))...
cuda,gpgpu,gpu-programming,cublas
The only actual problem in your code is here: cudaMalloc( &d_x,sizeof(d_x) ); sizeof(d_x) is just the size of a pointer. You can fix it like this: cudaMalloc( &d_x,sizeof(x) ); If you want to find out if a CUBLAS API call is failing, then you should check the return code of...
Converting comments into answer: condition_variable::wait(lock, pred) is equivalent to while(!pred()) wait(lock);. If pred() returns true then no wait actually takes place and the call returns immediately. Your first wake is from the notify_one() call; the second "wake" is because the second wait() call happens to execute after the Stop() call,...
You could use std::promise and std::future (or their boost counterparts if your are not yet on C++11). The idea is to store a std::shared_ptr<std::promise<bool>> with the current sequence id as a key in the map whenever a request is sent. In the blocking send function you wait for the corresponding...
The A[32] in the method is actually just a pointer to A. Therefore, sizeof is the size of *int. Take the following test code: void szof(int A[32]) { std::cout << "From method: " << sizeof(A) << "\n"; } int main(int argc, char *argv[]) { int B[32]; std::cout << "From main:...
The main reason to do something like this is to specialize void integerA(int x) to do something else. That is, if the programmer provides as input argument an int to member function abc::integerA then because of the C++ rules instead of instantiating the template member function the compiler would pick...
Your program is not ill-formed because <vector> is guaranteed to include <initializer_list> (the same is true for all standard library containers) §23.3.1 [sequences.general] Header <vector> synopsis #include <initializer_list> ... Searching the standard for #include <initializer_list> reveals the header is included along with the following headers <utility> <string> <array> <deque> <forward_list>...
The fanciest way I've seen to perform what you want is straight from the boost filesystem tutorial. In this particular example, the author appends the filename/directory to the vector and then utilizes a std::sort to ensure the data is in alphabetical order. Your code can easily be updated to use...
c++,inline,private,member,protected
The Compiler can Access everything. The restrictions are only valid for the programmer. This means there are no restrictions for the Compiler to Access any variables! At the end every variable is just translated to an address which can be accessed. So for the Compiler it is no Problem to...
There are no operator[] of std::map which is const, you have to use at or find: template<> struct Record::getDispatcher<std::string> { static std::string impl(Record const& rec, std::string& const field) { return rec.fieldValues_.at(field); // throw if field is not in map. } }; or template<> struct Record::getDispatcher<std::string> { static std::string impl(Record const&...
Removing BreakBeforeBraces: Allman Seems to do what you want (for me). I'm using SVN clang though. Although you probably wanted it there for a reason. According to the clang-format docs, the AllowShortBlocksOnASingleLine should do exactly what you want (regardless of brace style). This might be a bug in clang-format....
Preferred option: change isPrime to take a long (and pass *it to it). Secondary option: pass &*it instead of it. Your original code doesn't work because it is an iterator (which is a class) whereas the function expected long int * and there is no implicit conversion from iterator to...
c++,inheritance,constructor,subclass,superclass
This map: typedef map<string, Object> obj_map; only stores Object objects. When you try to put an Image in, it is sliced down and you lose everything in the Image that was not actually part of Object. The behaviour that you seem to be looking for is called polymorphism. To activate...
You're not using the function setText correctly. The canonical prototype is text(QString & subtype, Mode mode = Clipboard) const from the documentation. What you want to do is assemble your QString ahead of time and then use that to populate the clipboard. QString message = QString("Just a test text. And...
Yes, those objects still exist and you must delete them. Alternatively you could use std::vector<std::unique_ptr<myObject>> instead, so that your objects are deleted automatically. Or you could just not use dynamic allocation as it is more expensive and error-prone. Also note that you are misusing reserve. You either want to use...
Yes you can as of C++11 standard. Wiki article. Also a quick empirical verification: using namespace std; class A { public: A() { cout << "Hello "; } A(int x) : A() { cout << "World!" << endl; } }; int main() { A a(1); return 0; } Prints: Hello...
Use stoi, it's the modern C++ version of C's atoi. Update: Since the original answer text above the question was amended with the following error message: ‘stoi’ was not declared in this scope Assuming this error was produced by g++ (which uses that wording), this can have two different causes:...
You forgot the return statement in the function bool func(const pair<int,pair<int,int> >&i , const pair<int,pair<int,int> >&j ) { i.second.first < j.second.first ; } Write instead bool func(const pair<int,pair<int,int> >&i , const pair<int,pair<int,int> >&j ) { return i.second.first < j.second.first ; } Also you should include header <utility> where class std::pair...
If you want a sequence of int, then use a vector<int>. Using the key_char string, the values of the chars in it will serve as the initial value of the ints. std::vector<int> key_num(key_char.begin(), key_char.end()); Then, iterate over each character of key_num and convert it to the equivalent int value for...
If this is interview question or something , and you have to do it anyways , you can do this like ,below code . derive from std::stack , and overload [] operator #include <iostream> #include <algorithm> #include <stack> #include <exception> #include <stdexcept> template <typename T> class myStack:public std::stack<T> { public:...
As @101010 hints at: pay is a string, while hours_day is a float, and while some languages allow you to multiply strings with integers, c++11 (or any other flavor of c) doesn't, much less allow strings and floats to be multiplied together.
Its not the fastest method but you can do this: #include <string> #include <sstream> #include <iostream> template<typename ValueType> std::string stringulate(ValueType v) { std::ostringstream oss; oss << v; return oss.str(); } int main() { std::cout << ("string value: " + stringulate(5.98)) << '\n'; } ...
The header file provides enough information to let you declare variables. And for that matter to just compile (but not link) code. When you link, the linker has to resolve e.g. function references such as a reference to ServerConnection::getLicenceRefused, by bringing in the relevant machine code. You have to tell...
For your android problem you can use fb-adb which "propagates program exit status instead of always exiting with status 0" (preferred), or use this workaround (hackish... not recommended for production use): def run_exe_return_code(run_cmd): process=subprocess.Popen(run_cmd + '; echo $?',stdout=subprocess.PIPE,shell=True) (output,err)=process.communicate() exit_code = process.wait() print output print err print exit_code return exit_code...
You should use the random header. #include <random> std::default_random_engine generator; std::uniform_int_distribution dist(0, 5); int StringIndex = dist(generator); std::string ChosenString = characters[StringIndex]; The above will generate a random index into your array. If you want to limit the range, change the constructor of dist, for example (dist(0,2) would only allow for...
Your issue is that std::deque (and other standard containers) doesn't just take a single template argument. As well as the stored type, you can specify an allocator functor type to use. If you don't care about these additional arguments, you can just take a variadic template template and be on...
I think you just misspelled CFLAGS in CFLAGES=-c -Wall I'm guessing this is the case since g++ ../src/main.cpp -I ../include/ does not have the -c option...
(Edited away first "answer", this is an actual attempt at an answer) My guess: QList<Msg> messages() const { return _messages; } It's returning a copy of the QList _messages, rather than a reference to it. I'm not sure that it would give the results you're seeing, but it looks wrong...