This repository has long been aged by the proprietor. It is right now read-onIy.
Join GitHub nowadays
GitHub can be house to over 36 million designers working jointly to host and evaluate code, manage tasks, and develop software jointly.
Sign upFor example, itʼs common to pass a delegate to ConcurrentDictionary. NET Core where they use an optimized representation to reduce their size, which. NET objects, i.e. It will show you how the JITter has decided to arrange the fields.
Remarks
mentionedOct 20, 2016
It may become achievable to considerably enhance the overall performance of MemoryCache by using ConcurrentDictionary rather thanDictionary covered by aReaderWriterLockSIim l developed a quick prototype which shows significantly enhanced functionality, though it may not really become functionally correct in all scenarios.
The staying work can be to enhance and test the ConcurrentDictionary implementation to assure it't correct under higher load, has no competition situations, étc. Prototypé: https://github.cóm/aspnet/Caching/woods/mikeharder/concurrent-dictionary |
included the2 - Functioningcontent labelApril 27, 2016
referenced this problemApril 31, 2016
MergedConcurrent dictionary implementation for MemoryCache#246
added thepestcontent labelNov 3, 2016
includedenhancementand removedinsectlabelsNov 3, 2016
left a commentNovember 16, 2016
Please be sure to thoroughly test ConcurrentDictionary vs locking mechanism'ed Dictionary functionality and don't make this transformation blindly. Overall performance may rely a great deal on particular workload (amount of obtain vs set, etc.), also simple Keep track of may be faster than ReaderWriterLockSlim if most operations are writés. |
mentionedNov 17, 2016
@Dissimilis thanks a lot for the opinions but is definitely right now there any specific workload you are usually worried about? A several scenarios we have got benchmarked include 1) concurrent gets with a several collection to up-date the access and 2) concurrent sets and removes and possess seen main enhancements in both. Find the referenced Page rank for benchmark results. The move towardConcurrentDictionary furthermore takes benefit of the more fine-grained Iocking and will significantly improve overall performance when concurrently accessing different entries in the cache compared to the previous implementation which uses a oneReaderWriterLockSlimacross the whole dictionary. |
included3 - Completedand taken out2 - FunctioningbrandsNov 17, 2016
commentedNov 17, 2016
Can we upgrade the desk with the new figures? |
commentedNovember 17, 2016
|
Indication up for freeto subscribe to this discussion on GitHub. Currently possess an accounts? Sign in.