Skip to content

amo: apply fast-path analysis and optimization  #117

Description

@minsii

Direct AMO mode still shows about 2x overhead than that of SOS

Benchmark: osu_oshm_atomics_all2one (shmem_int_finc -> MPI FOP)

        #SOS   #direct-amo
Theta/np=2 	14.88	39.22
Cori/np=2 	2.67    4.21

The current atomics check in OFI can be the cause. We need analyze the atomics path and apply optimizations similer to previous RMA (IPO inline, reduce instructions on AMO path)

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions