Hello Julia Community, especially Dagger and HPC developers. I hope you are doing well this summer and have achieved your goals! I’m here to share: Dagger’s MPI Implementation status My Google Summer of Code at Julia conclusion Some great data visualization about Dagger’s MPI Implementation I wou…

[image] yanzin00: In GFLOPS results Figure 2, MPI sustains significantly higher throughput than TCP I cannot see this in the figure. In the left figure TCP has 40GFLOPS with 25 processors, and MPI has close to zero GFLOPS. Perhaps there is a mistake in the figures?

Productive Scalable Distributed Task Scheduling Using an MPI-based Backend for Dagger

Community

jblaschke September 8, 2025, 2:57am 4

Why is MPI so much slower than TCP? Can you share some details about your MPI setup? Also this looks like you’re running single-node only tests, or am I missing something?

2 Likes

Topic		Replies	Views
Mentor for HPC GSoC project Community	2	822	February 28, 2017
Dagger, No speed increase, parallel computing General Usage	5	1765	January 26, 2018
Julia vs Python's Dask: Known speed comparisons? Julia at Scale question , parallel , distributed	13	5099	October 15, 2019
Faster MPI Integration in Dagger Community announcement , parallel , mpi , dagger	9	464	July 27, 2025
Help Optimizing Slow Distributed Array Computation Performance distributed	7	1938	November 14, 2018

Productive Scalable Distributed Task Scheduling Using an MPI-based Backend for Dagger

Related topics