Matches in SemOpenAlex for { <https://semopenalex.org/work/W4384705435> ?p ?o ?g. }
Showing items 1 to 68 of
68
with 100 items per page.
- W4384705435 abstract "MPI Neighborhood collectives are used for non-traditional collective operations involving uneven distribution of communication amongst processes such as sparse communication patterns. They provide flexibility to define the communication pattern involved when a neighborhood relationship can be defined. PETSc, the Portable, Extensible Toolkit for Scientific Computation, used extensively with scientific applications to provide scalable solutions through routines modeled by partial differential equations, utilizes neighborhood communication patterns to define various structures and routines.We propose GPU-aware MPI Neighborhood collective operations with support for AMD and NVIDIA GPU backends and propose optimized designs to provide scalable performance for various communication routines. We evaluate our designs using PETSc structures for scattering from a parallel vector to a parallel vector, scattering from a sequential vector to a parallel vector, and scattering from a parallel vector to a sequential vector using a star forest graph representation implemented with nonblocking MPI neighborhood alltoallv collective operations. We evaluate our neighborhood designs on 64 NVIDIA GPUs on the Lassen system with Infiniband networking, demonstrating30.90% improvement against a GPU implementation utilizing CPU-staging techniques, and 8.25% improvement against GPU-aware point-to-point implementations of the communication pattern. We also evaluate on 64 AMD GPUs on the Spock system with slingshot networking and present 39.52% improvement against the CPU-staging implementation of a neighborhood GPU vector type in PETSc, and 33.25% improvement against GPU-aware point-to-point implementation of the routine." @default.
- W4384705435 created "2023-07-20" @default.
- W4384705435 creator A5017245523 @default.
- W4384705435 creator A5024879682 @default.
- W4384705435 creator A5034293705 @default.
- W4384705435 creator A5057085084 @default.
- W4384705435 date "2023-05-01" @default.
- W4384705435 modified "2023-09-26" @default.
- W4384705435 title "Designing and Optimizing GPU-aware Nonblocking MPI Neighborhood Collective Communication for PETSc<sup>*</sup>" @default.
- W4384705435 cites W1569090332 @default.
- W4384705435 cites W2024565525 @default.
- W4384705435 cites W2025388037 @default.
- W4384705435 cites W2047102972 @default.
- W4384705435 cites W2097215900 @default.
- W4384705435 cites W2787701215 @default.
- W4384705435 cites W2963753177 @default.
- W4384705435 cites W2971519680 @default.
- W4384705435 cites W3132234403 @default.
- W4384705435 cites W3166649773 @default.
- W4384705435 cites W3170377433 @default.
- W4384705435 cites W3199181189 @default.
- W4384705435 cites W4250705075 @default.
- W4384705435 cites W4284964190 @default.
- W4384705435 doi "https://doi.org/10.1109/ipdps54959.2023.00070" @default.
- W4384705435 hasPublicationYear "2023" @default.
- W4384705435 type Work @default.
- W4384705435 citedByCount "0" @default.
- W4384705435 crossrefType "proceedings-article" @default.
- W4384705435 hasAuthorship W4384705435A5017245523 @default.
- W4384705435 hasAuthorship W4384705435A5024879682 @default.
- W4384705435 hasAuthorship W4384705435A5034293705 @default.
- W4384705435 hasAuthorship W4384705435A5057085084 @default.
- W4384705435 hasConcept C111919701 @default.
- W4384705435 hasConcept C173608175 @default.
- W4384705435 hasConcept C199360897 @default.
- W4384705435 hasConcept C26713055 @default.
- W4384705435 hasConcept C2781030343 @default.
- W4384705435 hasConcept C41008148 @default.
- W4384705435 hasConcept C459310 @default.
- W4384705435 hasConcept C48044578 @default.
- W4384705435 hasConcept C83283714 @default.
- W4384705435 hasConcept C854659 @default.
- W4384705435 hasConceptScore W4384705435C111919701 @default.
- W4384705435 hasConceptScore W4384705435C173608175 @default.
- W4384705435 hasConceptScore W4384705435C199360897 @default.
- W4384705435 hasConceptScore W4384705435C26713055 @default.
- W4384705435 hasConceptScore W4384705435C2781030343 @default.
- W4384705435 hasConceptScore W4384705435C41008148 @default.
- W4384705435 hasConceptScore W4384705435C459310 @default.
- W4384705435 hasConceptScore W4384705435C48044578 @default.
- W4384705435 hasConceptScore W4384705435C83283714 @default.
- W4384705435 hasConceptScore W4384705435C854659 @default.
- W4384705435 hasLocation W43847054351 @default.
- W4384705435 hasOpenAccess W4384705435 @default.
- W4384705435 hasPrimaryLocation W43847054351 @default.
- W4384705435 hasRelatedWork W1865359956 @default.
- W4384705435 hasRelatedWork W1969934278 @default.
- W4384705435 hasRelatedWork W1970172627 @default.
- W4384705435 hasRelatedWork W2035975621 @default.
- W4384705435 hasRelatedWork W2131186568 @default.
- W4384705435 hasRelatedWork W2156222394 @default.
- W4384705435 hasRelatedWork W2509326634 @default.
- W4384705435 hasRelatedWork W2906842697 @default.
- W4384705435 hasRelatedWork W4206430307 @default.
- W4384705435 hasRelatedWork W4289520059 @default.
- W4384705435 isParatext "false" @default.
- W4384705435 isRetracted "false" @default.
- W4384705435 workType "article" @default.