We have large (100,000+ elements) ordered vectors of structs (operator < overloaded to provide ordering):
std::vector < MyType > vectorMyTypes;
std::sort(vectorMyType.begin(), vectorMyType.end());
My problem is that we're seeing performance problems when adding new elements to these vectors while preserving sort order. At the moment we're doing something like:
for ( a very large set )
{
vectorMyTypes.push_back(newType);
std::sort(vectorMyType.begin(), vectorMyType.end());
...
ValidateStuff(vectorMyType); // this method expects the vector to be ordered
}
This isn't exactly what our code looks like since I know this example could be optimised in different ways, however it gives you an idea of how performance could be a problem because I'm sorting after every push_back.
I think I essentially have two options to improve performance:
Use a (hand crafted?) insertion sort instead of std::sort to improve the sort performance (insertion sorts on a partially sorted vector are blindingly quick)
Create a heap by using std::make_heap and std::push_heap to maintain the sort order
My questions are:
Should I implement an insertion sort? Is there something in Boost that could help me here?
Should I consider using a heap? How would I do this?
Edit:
Thanks for all your responses. I understand that the example I gave was far from optimal and it doesn't fully represent what I have in my code right now. It was simply there to illustrate the performance bottleneck I was experiencing - perhaps that's why this question isn't seeing many up-votes :)
Many thanks to you Steve, it's often the simplest answers that are the best, and perhaps it was my over analysis of the problem that blinded me to perhaps the most obvious solution. I do like the neat method you outlined to insert directly into a pre-ordered vector.
As I've commented, I'm constrained to using vectors right now, so std::set, std::map, etc aren't an option.
Ordered insertion doesn't need boost:
vectorMyTypes.insert(
std::upper_bound(vectorMyTypes.begin(), vectorMyTypes.end(), newType),
newType);
upper_bound provides a valid insertion point provided that the vector is sorted to start with, so as long as you only ever insert elements in their correct place, you're done. I originally said lower_bound, but if the vector contains multiple equal elements, then upper_bound selects the insertion point which requires less work.
This does have to copy O(n) elements, but you say insertion sort is "blindingly fast", and this is faster. If it's not fast enough, you have to find a way to add items in batches and validate at the end, or else give up on contiguous storage and switch to a container which maintains order, such as set or multiset.
A heap does not maintain order in the underlying container, but is good for a priority queue or similar, because it makes removal of the maximum element fast. You say you want to maintain the vector in order, but if you never actually iterate over the whole collection in order then you might not need it to be fully ordered, and that's when a heap is useful.
According to item 23 of Meyers' Effective STL, you should use a sorted vector if you application use its data structures in 3 phases. From the book, they are :
Setup. Create a new data structure by inserting lots of elements into it. During this phase, almost all operation are insertions and erasure. Lookups are rare on nonexistent
Lookup. Consult the data structure to find specific pieces of information. During this phase, almost all operations are lookups. Insertion and erasures are rare or nonexistent. There are so many lookups, the performance of this phase makes the performance of the other phases incidental.
Reorganize. Modify the content of the data structure. perhaps by erasing all the current data and inserting new data in its place. Behaviorally, this phase is equivalent to phase 1. Once this phase is completed, the application return to phase 2
If your use of your data structure resembles this, you should use a sorted vector, and then use a binary_search as mentionned. If not, a typical associative container should do it, that means a set, multi-set, map or multimap as those structure are ordered by default
Why not just use a binary search to find where to insert the new element? Then you will insert exactly into the required position.
If you need to insert a lot of elements into a sorted sequence, use std::merge, potentially sorting the new elements first:
void add( std::vector<Foo> & oldFoos, const std::vector<Foo> & newFoos ) {
std::vector<Foo> merged;
// precondition: oldFoos _and newFoos_ are sorted
merged.reserve( oldFoos.size() + newFoos.size() ); // only for std::vector
std::merge( oldFoos.begin(), oldFoos.end(),
newFoos.begin(), newFoos.end(),
std::back_inserter( merged );
// apply std::unique, if wanted, here
merged.erase( std::unique( merged.begin(), merged.end() ), merged.end() );
oldFoos.swap( merged ); // commit changes
}
Using a binary search to find the insertion location isn't going to speed up the algorithm much because it will still be O(N) to do the insertion (consider inserting at the beginning of a vector - you have to move every element down one to create the space).
A tree (aka heap) will be O(log(N)) to insert, much better performance.
See http://www.sgi.com/tech/stl/priority_queue.html
Note that a tree will still have worst case O(N) performance for insert unless it is balanced, e.g. an AVL tree.
Why not to use boost::multi_index ?
NOTE: boost::multi_index does not provide memory contiguity, a property of std::vectors by which elements are stored adjacent to one another in a single block of memory.
There are a few things you need to do.
You may want to consider making use of reserve() to avoid excessive re-allocing of the entire vector. If you have knowledge of the size it will grow to, you may gain some performance by doing resrve()s yourself (rather than having the implemetation do them automaticaly using the built in heuristic).
Do a binary search to find the insertion location. Then resize and shift everything following the insertion point up by one to make room.
Consider: do you really want to use a vector? Perhaps a set or map are better.
The advantage of binary search over lower_bound is that if the insertion point is close to the end of the vector you don't have to pay the theta(n) complexity.
If you want insert an element into the "right" position, why do you plan on using sort. Find the position using lower_bound and insert, using, well, `insert' method of the vector. That will still be O(N) to insert new item.
heap is not going to help you, because heap is not sorted. It allows you get get at the smallest element quickly, and then quickly remove it and get next smallest element. However, the data in heap is not stored in sort order, so if you have algorithms that must iterate over data in order, it will not help.
I am afraid you description skimmed to much detail, but it seems like list is just not the right element for the task. std::deque is much better suited for insertion in the middle, and you might also consider std::set. I suggest you explain why you need to keep the data sorted to get more helpful advice.
You might want to consider using a BTree or a Judy Trie.
You don't want to use contiguous memory for large collections, insertions should not take O(n) time;
You want to use at least binary insertion for single elements, multiple elements should be presorted so you can make the search boundaries smaller;
You do not want your data structure wasting memory, so nothing with left and right pointers for each data element.
As others have said I'd probably have created a BTree out of a linked list instead of using a vector. Even if you got past the sorting issue, vectors have the problem of fully reallocating when they need to grow, assuming you don't know your maximum size before hand.
If you are worried about a list allocating on different memory pages and causing cache related performance issues, preallocate your nodes in an array, (pool the objects) and insert these into the list.
You can add a value in your data type that denotes if it is allocated off the heap or from a pool. This way if you detect that your pool runs out of room, you can start allocating off the heap and throw an assert or something to yourself so you know to bump up the pool size (or make this a command line option to set.
Hope this helps, as I see you already have lots of great answers.
Related
I am looking for a data structure that is faster than C++'s unordered_map in my scenario.
I am storing in the map unsorted unique C-String char * (map.first) and integers (map.second). I can use around 10MB of memory for this data structure. Before I add a new item I need to check if it exists first. So, I am doing a ton of searches and a lot of inserts. The data structure will contain few items ( < 500 ) usually and then it will be deleted. So, I don't need to delete individual items.
I implemented my own AVL self balancing tree (seems to be really good for my case) but it was actually slower compared to std::unordered_map.
Do you know any data structure better than unordered_map in my case?
A good answer to this would be a combination of linear lookup and binary search.
Basically have a sorted vector of items which you can binary search in. This will have fantastic cache locality and will probably be quicker for the kind of sizes you're looking at. If you need to insert just push it onto a seperate unsorted vector. When you next need to search both do a linear search of the unsorted vector and a binary search of the sorted vector. When your unsorted vector gets big enough (say 10 but profiling will help here) then insert them to the back of the sorted vector and resort it then clear out the 'unsorted' vector.
This doesn't have the best complexity guarantees, but will likely be faster on modern hardware for the kind of sizes you're looking at (linear memory access are FAST and likely beat trees/lists until you get quite large).
Sorting the unsorted vector and then merging it into the sorted one would give a bit of a speed increase at the cost of complexity of code.
If Memory really doesn't matter you can create a huge vector<bool> and store if the given value was inserted in your AVL tree.
e.g. have a look at Counting sort. You could implement it like this.
It sounds like your use case calls for a set rather than a map. Do you really need a map for some reason not clear in the question? If not, an unordered_set would be a better choice and if you are dealing with a small enough range a vector<bool> as suggested by Thomas Sparber.
I want to know which data-structures are more efficient for iterating through their elements between std::set, std::map and std::unordered_set, std::unordered_map.
I searched through SO and I found this question. The answers either propose to copy the elements in a std::vector or to use Boost.Container, which IMHO don't answer my question.
My purpose is to keep in a container a big number of unique elements, that most of the time I want to iterate through them. Insertions and extractions are more rare. I want to avoid std::vector in combination with std::unique.
Lets consider set vs unordered_set.
The main difference here is the 'nature' of the iteration, that is the traversal of the set will give you the elements in order while traversing a range in an unordered set will give you a bunch of values in no particular order.
Suppose you want to traverse a range [it1, it2]. If we exclude the lookup time that's needed to find elements it1 and it2 there can be no direct mapping from one case to another since the elements in between are not guarrandeed to be the same even if you've used the same elements to construct the container.
There are cases however where something like this has meaning when e.g. you want to traverse a fixed number of elements (regardless of what they are) or when you need to traverse the whole container. In such cases you need to consider implementation mechanics :
Sets are usually implemented like Red–black trees (a form of binary search trees). Like all binary search trees allow efficient in-order traversal (LRR: left root right) of their elements. That is to traverse you pay the cost of pointer chasing (just like traversing a list).
Unordered sets on the other hand are hash tables and to my knowledge the STL implementation uses hashing with chaining. That means (in a very very high level) that what's used for the structure is a (contiguous) buffer where each element is the head of a chain (list) that contains the elements. The way the elements are layed out across those chains (buckets) and across the buffer will affect the traversal time, however you'll be chasing pointers once again jumping through differents lists as well this time. I don't think it'll vary significantly from the tree case but won't be any better for sure.
In any case micro tuning and benchmarking will give you the answer for your particular application.
The difference does not lie between the ordering or lack of one but in the backing container. If it's a contiguous memory it should be fast to iterate over, due to simple implementation of iterator and cache friendliness.
Unordered containers are usually stored as a vector of vectors (or a similar thing), while ordered containers are implemented using trees, but it is left for implementation after all. This would suggest that iterating over unordered version should be waster. However this is left for implementation after all, and I saw implementations (which bent rules a little to be fair) with different behaviour.
Generally speaking, container performance is quite a complex topic and usually has to be tested in actual application to get reliable answer. There is plenty on implemention-defined stuff that might affect the performance. I'd go with hash_set if I had to go in blind. Copying into a vector might also turn out a good option.
EDIT: As #TonyD said in it's comment, there is a rule, that disallows invalidating iterators during addition of element when the max_load_factor() is not exceeded, this practically rules out backing containers which are contiguous in memory.
Thus, copying everything into a vector seems like even more reasonable option. If you need to remove duplicates, a feasible option might be to use http://en.cppreference.com/w/cpp/algorithm/sort and have dupes easily ignored. I have heard that using vector and sort to have a sorted array (or vector) is quite often a used option in case of need for a container that needs to be sorter and is being iterated over more often than modified.
iterate from fastest to slowest should be : set > map > unordered_set > unordered_map;
set is a little lighter than map, and they are ordered with binary tree rule, so should be faster than unordered_ containers.
Interview question: There is a stream of Integers that arrives at specified intervals (say every 20 sec). Which Container of STL would you use to store them so that the Integers look sorted? My reply was map/set when there is no duplicate or multimap/multiset when there is duplicate. Any better answer if exists?
Use a multiset if you want to preserve duplicates. If you don't want to preserve duplicates, use a set.
If it's only being updated every 20 seconds, it probably doesn't matter a whole lot (unless it goes for so long that the set of integers becomes tremendously huge).
If you had data coming in a lot faster, there are alternatives that might be worth considering. One would be to use a couple of vectors. As data arrives, just push it onto one of the vectors. When you need to do an in-order traversal, sort that newly arrived data, and merge with the other vector of existing (already-sorted data). That'll give you results in order, which you can then write out to another vector, and start the same cycle again.
The big advantage here is that you're dealing with contiguous data instead of individually allocated nodes. Even with a possibility of three vectors in use at a time, your total memory usage is likely to be about equal (or possibly even less than) that of using a set or multiset.
Another possibility to consider (that's a bit of a hybrid between the two) would be something like a B+ tree. This is still a tree, so you can do in-order insertions with logarithmic complexity, but you have all the data in the leaf nodes (which are fairly large) so you get at least a reasonable amount of contiguous access as well.
To maintain a sorted list of integers streaming I would use std::priority_queue with any underlying container (vector or deque depending on the particular use).
You can keep push() ing to the priority_queue and use top() and pop() to retrieve in the sorted order.
Answer should be std::set . std::map<key, value> has to consider when there is a pairs of data as <key, value> and it need to be sorted according to the value of key
In same way if you have to consider duplicates, use std::multiset and std::multimap according to type of data.
For my application I am using STD vector. I am inserting to the vector at the end, but erasing from vector randomly i.e element can be erased from middle, front anywhere. These two are only requirement, 1)insert at the end 2) erase from anywhere.
So should I use STD List, since erasing does shifting of data. Or I would retain Vector in my code for any reason??
Please give comment, If Vector is the better option, how it would be better that List here?
One key reason to use std::vector over std::list is cache locality. A list is terrible in this regard, because its elements can be (and usually are) fragmented in your memory. This will degrade performance significantly.
Some would recommend using std::vector almost always. In terms of performance, cache locality is often more important than the complexity of insertion or deletion.
Here's a video about Bjarne Stroustrup's opinion regarding subject.
I would refer you to this cheat sheet, and the conclusion would be the list.
A list supports deletion at an arbitrary but known position in constant time.
Finding that position takes linear time, just like modifying a vector.
The only advantage of the list is if you repeatedly erase (or insert) at (approximately) the same position.
If you're erasing more or less at random, chances are that the better memory locality of the vector could win out in the end.
The only way to be sure is to measure and compare.
List is better in this case most probably. The advantage of a list over vector is that it supports deletion at arbitrary position with constant complexity. A vector would only be better choice if you require constant index operation of elements of the container. Still you have to take into consideration how is the element you would like to delete passed to your function for deletion. If you only pass an index, vector will be able to find the element in constant time, while in list you will have to iterate. In this case I would benchmark the two solution, but still I would bet on list performing better.
It depends on many factors and how are you using your data.
One factor: do you need an erase that maintains the order of the collection? or you can live with changing order?
Other factor: what kind of data is in the collection? (numbers: ints/floats) || pointers || objects
Not keeping order
You could continue using vector if the data is basic numbers or pointers, to delete one element you could copy the last element of the vector over the deleted position, then pop_back(). This way you avoid moving all the data.
If using objects, you could still use the same method if the object you need to copy is small.
Keeping order
Maybe List would be your friend here, still, some tests would be advised, depends on size of data, size of list, etc
I have the following scenario.The implementation is required for a real time application.
1)I need to store at max 20 entries in a container(STL Map, STL List etc).
2)If a new entry comes and 20 entries are already present i have to overwrite the oldest entry with the new entry.
Considering point 2, i feel if the container is full (Max 20 entries) 'list' is the best bet as i can always remove the first entry in the list and add the new one at last (push_back). However, search won't be as efficient.
For only 20 entries, does it really make a big difference in terms of searching efficiency if i use a list in place of a map?
Also considering the cost of insertion in map i feel i should go for a list?
Could you please tell what is a better bet for me ?
1)I need to store at max 20 entries in a container(STL Map, STL List etc). 2)If a new entry comes and 20 entries are already present i have to overwrite the oldest entry with the new entry.
This seems to me the job for boost::circular_buffer.
In general the term circular buffer refers to an area in memory which is used to store incoming data. When the buffer is filled, new data is written starting at the beginning of the buffer and overwriting the old.
The circular_buffer is a STL compliant container. It is a kind of sequence similar to std::list or std::deque. It supports random access iterators, constant time insert and erase operations at the beginning or the end of the buffer and interoperability with std algorithms. The circular_buffer is especially designed to provide fixed capacity storage. When its capacity is exhausted, newly inserted elements will cause elements either at the beginning or end of the buffer (depending on what insert operation is used) to be overwritten.
The circular_buffer only allocates memory when created, when the capacity is adjusted explicitly, or as necessary to accommodate resizing or assign operations. On the other hand, there is also a circular_buffer_space_optimized available. It is an adaptor of the circular_buffer which does not allocate memory at once when created, rather it allocates memory as needed.
For the fast search, I think that with just 20 elements (if their comparison isn't too complicated) you're ok with a "low-cost" container like this and normal linear search, in my opinion it would be difficult to achieve better performance with other STL containers.
Maintain order of insertion, or allow fast searching: choose one.
std::map is not an option here because it doesn't maintain the order of insertion. Besides, it's an associative container. You should choose between a list, a deque and a vector. In terms of performance your best bet is a list, since you can pop off an element from the back and insert a new one at the front (or vice-versa) without any shifting or performance penalty.
The cost of insertion in a map, just as a sidenote, isn't expensive it all: it's in the order of O(log n). Practically irrelevant in the case of 20 elements. The same holds for a std::set.
With only 20 elements, I would not worry much about which container you use. If you determine that the container chosen is in fact a detriment to the performance of your application, it should be relatively easy to swap out the container chosen and replace it with a more-efficient container later.
With that being said, for a large number of elements, the std::deque would probably give you the best all-around efficiency for what you are trying to accomplish. Unlike std::vector, std::deque allows for removal from the front without needing to move all of the other elements. Unlike std::list, std::deque allows for random access of its elements.
You just need to implement a priority queue. STL Map doesn't work.
It depends on the size of the elements.
I know from my own experience that for five integers an unordered array of integers searched with linear search is faster than a set, a list or insertion sort and binary search on an ordered array.
The O() notation of an unordered array may be much worse than any of the other options but the normally unseen C in O(N+C) + C is so much smaller.
A list, set or map (anything that uses dynamic memory and is linked by pointers) will be dominated by cache misses, memory allocations and indirect reference penalties.
You need a Priority Queue implemented on an array.
See the Binary Heap for an implementation.
Do you already know that this is a bottleneck?
My advice would be to first use what is more natural to read while programming and only optimize it when you see that the performance is not what you need.
My suggestion would be to make a circular buffer. But that only works if "old" is determined by when it was inserted, and not some field.
If you need to have a proper LRU, then you should probably go and look at something like http://www.codeproject.com/KB/recipes/LRUCache.aspx?fid=1000025&df=90&mpp=25&noise=3&sort=Position&view=Quick&fr=15
But with 20 entries as your max, it will be very hard to you to find a complex algorithm that is actually faster than the trivial lineary check of every element.