Skip to content

mpmc_result__1.11.9.md

Latest commit

 

History

History
57 lines (45 loc) · 2.35 KB

mpmc_result__1.11.9.md

File metadata and controls

57 lines (45 loc) · 2.35 KB
Raw

Benchmark: Multi-producer multi-consumer (MPMC) performance of AtomicChannels.jl

Date: 2026-05-06 09:39:23
AtomicChannels.jl Version 1.0.2
Julia Version 1.11.9
Commit 53a02c0720c (2026-02-06 00:27 UTC)
Build Info:
Official https://julialang.org/ release
Platform Info:
OS: Linux (x86_64-linux-gnu)
CPU: 128 × AMD Ryzen Threadripper PRO 7985WX 64-Cores
WORD_SIZE: 64
LLVM: libLLVM-16.0.6 (ORCJIT, znver4)

Case 1: Low capacity to encourage contention

This benchmark evaluates the performance of data operations (put and take), without task switching (when threads>1).

threads workers capacity items Channel/ms AtomicChannel/ms speedup
64 64 4 100000 1398.5 143.42 9.751x
32 32 4 100000 678.46 74.36 9.124x
16 16 4 100000 405.61 55.16 7.353x
8 8 1 100000 384.79 55.46 6.939x
4 4 1 100000 147.87 28.6 5.169x
2 2 1 100000 146.33 26.65 5.491x
1 2 1 100000 116.0 104.72 1.108x

Case 2: Higher capacity with minimal contention

This benchmark mimics a lowest contention scenario where the contention is only from task switching, not data operations.

threads workers capacity items Channel/ms AtomicChannel/ms speedup
64 64 256 100000 126.33 28.71 4.4x
32 32 256 100000 88.66 24.63 3.6x
16 16 256 100000 55.28 26.61 2.078x
8 8 256 100000 43.86 25.34 1.731x
4 4 256 100000 36.79 18.32 2.008x
2 2 256 100000 22.19 18.1 1.226x
1 2 256 100000 13.96 9.82 1.421x

Case 3: Varying worker (task) counts to mimic different levels of concurrency

This benchmark mimics varying concurrency levels on both data operations and task switching.

threads workers capacity items Channel/ms AtomicChannel/ms speedup
32 512 256 100000 47.16 23.39 2.017x
32 256 256 100000 43.26 19.98 2.165x
32 128 256 100000 45.85 23.58 1.944x
32 64 256 100000 52.13 24.88 2.095x
32 32 256 100000 163.46 37.71 4.335x
32 16 256 100000 70.11 23.83 2.941x
32 8 256 100000 47.96 27.51 1.743x