This is a Multi-Producer Single-Consumer queue implementation specific to SBCL x86-64. It is probably one of the fastest of its kind. It implements the beautiful algorithm found at Non-intrusive MPSC node-based queue, and is intended for developing high-performance Actor systems.

Performance comparison with sb-concurrency:queue (8 producers, 1 consumer, 8M items; iMac Pro (2017), 8 Core Xeon, 32 GB RAM):

FAST-MPSC-QUEUE> (benchmark-fast)
Evaluation took:
  0.438 seconds of real time
  3.127986 seconds of total run time (2.892130 user, 0.235856 system)
  714.16% CPU
  1,398,546,106 processor cycles
  128,508,736 bytes consed

NIL
FAST-MPSC-QUEUE> (benchmark-sb-concurrency)
Evaluation took:
  1.650 seconds of real time
  14.386185 seconds of total run time (14.321212 user, 0.064973 system)
  [ Run times consist of 0.002 seconds GC time, and 14.385 seconds non-GC time. ]
  871.88% CPU
  5,267,885,814 processor cycles
  128,292,352 bytes consed

NIL

fast-mpsc-queue is able to process items at 18.2M/s while sb-concurrency:queue did 4.8M/s. That’s lots of speed!