Asp Forum - Re: How to achieve structure level locking in STL containers in C++

Wenwei Peng

4/22/2015 7:58:00 AM

? 2007?12?5???? UTC+8??1:45:02,grbgooglefan??:
> My application uses the caching heavily to store the data from
> databases & also the runtime orders information.
> All these caches are built on STL hash map, vectors & maps and data
> format is the structures.
> There are multiple threads accessing these caches simultaneously for
> reading the data as well as updating the data.
> Whenever any thread accesses the cache, it locks that cache & finds
> the required element. Does the actions required & unlocks the cache.
> I am finding that this is causing the other threads to wait for longer
> time because the locking is at cache level.
> I would like to make this locking quite finer & granular, in such a
> way that only the single structure which is to be updated is locked.
> This is somewhat same as the row level locking in databases.
>
> I am using the hash map, etc from default available STL library on
> Linux (libstdc++). I am using them as memory cache for server
> application.
> Most of these caches are built using the string as key and the
> structure as data. Flow of functions happens somewhat as below in the
> application:
> Thread 1 gets data from database using select query. Populates that
> data in the structures & pushes that structure on the cache.
> Thread 2 then picks up that structure & uses it for getting other live
> real time data from other service. Once that real time data is
> available, this Thread 2 keeps on updating that data in the structures
> in the cache.
> There Thread 3 which reads the data from this cache & uses for
> processing the requests from the client application.
>
> So all types of data structure operations - insertions, erasing,
> updating, reading are happening on this cache with the strctures.
> The cache holds more than 60000 data elements (structures) & searching
> for each required structure takes long time. So when one thread is
> trying to get the required structure, others have to wait & cannot do
> any productive work. Because they also need data from the same cache
> (even though for different type of actions).
>
> How do we achieve this level of granular locking?
> Thoughts please.

Title: The core of the core of the big data solutions -- Map
Author: pengwenwei
Email:
Language: c++
Platform: Windows, linux
Technology: Perfect hash algorithm
Level: Advanced
Description: Map algorithm with high performance
Section MFC c++ map stl
SubSection c++ algorithm
License: (GPLv3)

Download demo project - 1070 Kb
Download source - 1070 Kb

Introduction:
For the c++ program, map is used everywhere.And bottleneck of program performance is often the performance of map.Especially in the case of large data,and the business association closely and unable to realize the data distribution and parallel processing condition.So the performance of map becomes the key technology.

In the work experience with telecommunications industry and the information security industry, I was dealing with the big bottom data,especially the most complex information security industry data,all can’t do without map.

For example, IP table, MAC table, telephone number list, domain name resolution table, ID number table query, the Trojan horse virus characteristic code of cloud killing etc..

The map of STL library using binary chop, its has the worst performance.Google Hash map has the optimal performance and memory at present, but it has repeated collision probability.Now the big data rarely use a collision probability map,especially relating to fees, can’t be wrong.

Now I put my algorithms out here,there are three kinds of map,after the build is Hash map.We can test the comparison,my algorithm has the zero probability of collision,but its performance is also better than the hash algorithm, even its ordinary performance has no much difference with Google.

My algorithm is perfect hash algorithm,its key index and the principle of compression algorithm is out of the ordinary,the most important is a completely different structure,so the key index compression is fundamentally different.The most direct benefit for program is that for the original map need ten servers for solutions but now I only need one server.
Declare: the code can not be used for commercial purposes, if for commercial applications,you can contact me with QQ 75293192.
Download:
https://sourceforge.net/projects/pwwhas...

Applications:
First,modern warfare can’t be without the mass of information query, if the query of enemy target information slows down a second, it could lead to the delaying fighter, leading to failure of the entire war. Information retrieval is inseparable from the map, if military products use pwwhashMap instead of the traditional map,you must be the winner.

Scond,the performance of the router determines the surfing speed, just replace open source router code map for pwwHashMap, its speed can increase ten times.
There are many tables to query and set in the router DHCP ptotocol,such as IP,Mac ,and all these are completed by map.But until now,all map are using STL liabrary,its performance is very low,and using the Hash map has error probability,so it can only use multi router packet dispersion treatment.If using pwwHashMap, you can save at least ten sets of equipment.

Third,Hadoop is recognized as the big data solutions at present,and its most fundamental thing is super heavy use of the map,instead of SQL and table.Hadoop assumes the huge amounts of data so that the data is completely unable to move, people must carry on the data analysis in the local.But as long as the open source Hadoop code of the map changes into pwwHashMap, the performance will increase hundredfold without any problems.

Background to this article that may be useful such as an introduction to the basic ideas presented:
http://blog.csdn.net/chixinmuzi/article/detai...

comp.programming.threads

Re: How to achieve structure level locking in STL containers in C++

Wenwei Peng

x Login to ForumsZone