NSIS-ka
A free C++ implementation of NSIS protocols

Opened 11 years ago

Last modified 10 years ago

#18 new task

Reduction of threading

Reported by: bless Owned by: bless
Priority: minor Milestone:
Component: GIST Version:
Keywords: Cc:

Description

The performance is not optimal unless we have enough CPUs so that each thread is ready to process an internal message. This would form a perfect pipeline. Currently, it looks like a bursty processing behavior due to the internal message processing queues and too coarse granular thread switching. The strategy should be better as follows: each thread processes a message completely instead of passing it between different threads. This would reduce the latency for each message.

Change History (1)

comment:1 Changed 10 years ago by bless

  • Priority changed from major to minor
Note: See TracTickets for help on using tickets.