EPAM B2BITS CME MDP Handler C++  5.4.2
Public Member Functions | Public Attributes | List of all members
Cme::Mdp::MarketDataServiceThreadOptions Struct Reference

Market data service thread options. More...

#include <B2BITS_CmeMdpMarketDataService.h>

Collaboration diagram for Cme::Mdp::MarketDataServiceThreadOptions:
Collaboration graph
[legend]

Public Member Functions

 MarketDataServiceThreadOptions ()
 

Public Attributes

bool threadPoolEnabled_
 
u32 threadPoolSize_
 Number of shared threads in the pool. More...
 
bool spinningEnabled_
 Defines whether to use spinning to reduce latency at the cost of CPU load. More...
 
u64 affinityMask_
 
ThreadPriority priority_
 Defines thread priority. More...
 
u32 dblBFeedThreadCount_
 Number of threads to process data from Feed-B, if socket type is Myricom DBL and threadPoolEnabled_ is ON. More...
 
bool dblSpinningInOneThread_
 

Detailed Description

Market data service thread options.

Constructor & Destructor Documentation

Cme::Mdp::MarketDataServiceThreadOptions::MarketDataServiceThreadOptions ( )
inline

Member Data Documentation

u64 Cme::Mdp::MarketDataServiceThreadOptions::affinityMask_

Bitwise mask of CPUs to bind threads to. 0 indicates no thread binding.

u32 Cme::Mdp::MarketDataServiceThreadOptions::dblBFeedThreadCount_

Number of threads to process data from Feed-B, if socket type is Myricom DBL and threadPoolEnabled_ is ON.

bool Cme::Mdp::MarketDataServiceThreadOptions::dblSpinningInOneThread_

Limits access to DBL device only to one thread at time. Usefull if:

  1. MarketDataServiceSocketOptions::socketType_ is stMyricomDBL
  2. MarketDataServiceThreadOptions::spinningEnabled_ is ON
  3. and MarketDataServiceThreadOptions::threadPoolEnabled_ is ON Minimizes overal CPU usage. If ON, only one thread uses CPU at 100%; otherwise all threads in pool will use CPU at 100%.
ThreadPriority Cme::Mdp::MarketDataServiceThreadOptions::priority_

Defines thread priority.

bool Cme::Mdp::MarketDataServiceThreadOptions::spinningEnabled_

Defines whether to use spinning to reduce latency at the cost of CPU load.

bool Cme::Mdp::MarketDataServiceThreadOptions::threadPoolEnabled_

Defines whether to use a pool of shared threads for all connections or use a dedicated thread for each connection.

u32 Cme::Mdp::MarketDataServiceThreadOptions::threadPoolSize_

Number of shared threads in the pool.