List view

Profile DLA-Future more thoroughly

Priority:Medium

TODO:Task

Type:Optimization

Status: Closed (completed).

#587 In eth-cscs/DLA-Future;

· by msimberg was closed on Dec 8, 2023

Introduce sub-band in SEP pipeline

Priority:Medium

TODO:Task

Type:Optimization

Status: Closed (completed).

#502 In eth-cscs/DLA-Future;

· by albestro was closed on Jan 3, 2023

Need of a "semi" blocking barrier for bulk

Priority:High

TODO:Task

Type:New FeatureNew feature or requestNew feature or request

Type:Optimization

Status: Closed (completed).

#833 In eth-cscs/DLA-Future;

· by albestro was closed on Jun 12, 2023

Reduce gemm complexity in the tridiagonal eigensolver

Priority:High

TODO:Task

Type:Optimization

Status: Closed (completed).

#916 In eth-cscs/DLA-Future;

· by rasolca was closed on Dec 12, 2023

Tridiagonal eigensolver, reduce gemm complexity taking advantage of the matrix shapes.

Priority:High

TODO:Epic

Type:Optimization

Type:Performance issue

Status: Closed (completed).

#917 In eth-cscs/DLA-Future;

· by rasolca was closed on Sep 26, 2024

Implement matrix redistribution algorithm

Priority:High

TODO:Task

Type:New FeatureNew feature or requestNew feature or request

Type:Optimization

Status: Closed (completed).

#919 In eth-cscs/DLA-Future;

· by rasolca was closed on Jul 26, 2023

implement gemm for the tridiagonal solver

Priority:High

TODO:Task

Type:New FeatureNew feature or requestNew feature or request

Type:Optimization

Status: Closed (completed).

#915 In eth-cscs/DLA-Future;

· by rasolca was closed on Dec 12, 2023

Test if CUDA RDMA works

Priority:Medium

TODO:Task

Type:Optimization

Status: Closed (completed).

#734 In eth-cscs/DLA-Future;

· by msimberg was closed on Jun 14, 2024

Update CUDA RDMA option to work with HIP

Priority:Medium

TODO:Task

Type:Optimization

Status: Closed (completed).

#735 In eth-cscs/DLA-Future;

· by msimberg was closed on Jun 14, 2024

Extend ReductionToBand to accept band size

Priority:Medium

TODO:Task

Type:Optimization

Status: Closed (completed).

#701 In eth-cscs/DLA-Future;

· by albestro was closed on Dec 23, 2022

Test using stackless threads

Priority:Low

TODO:Task

Type:Optimization

Status: Closed (completed).

#720 In eth-cscs/DLA-Future;

· by msimberg was closed on Dec 8, 2023

Extend BacktransformationReductionToBand to accept band size

Priority:Medium

TODO:Task

Type:Optimization

Status: Closed (completed).

#702 In eth-cscs/DLA-Future;

· by albestro was closed on Dec 13, 2022

Permute (locally) to get the optimal matrix shape

Priority:High

TODO:Task

Type:Optimization

Status: Closed (completed).

#914 In eth-cscs/DLA-Future;

· by rasolca was closed on Dec 12, 2023

Tridiagonal Solver (dist): Migrate permutation of local eigenvectors to GPU

TODO:Task

Type:New FeatureNew feature or requestNew feature or request

Type:Optimization

Status: Closed (completed).

#1058 In eth-cscs/DLA-Future;

· by albestro was closed on Jun 21, 2024

Extend <code>reduce</code> communication kernel to work directly with GPU memory

Priority:Low

TODO:Task

Type:Optimization

Status: Closed (completed).

#1094 In eth-cscs/DLA-Future;

· by albestro was closed on Mar 29, 2024

algo: enable sub-band in Backtransform ReductionToBand Distributed (both MC and GPU)

Type:Optimization

Status: Merged (completed).

eth-cscs/DLA-Futurenumber 746

#746 In eth-cscs/DLA-Future;

· by albestro was closed on Dec 13, 2022

Retiled matrix + refactoring bt_band_to_tridiag

Priority:High

Type:Optimization

Status: Merged (completed).

eth-cscs/DLA-Futurenumber 799

#799 In eth-cscs/DLA-Future;

· by rasolca was closed on Apr 14, 2023

Split application of HH reflector block in bt_band_to_tridiag

Priority:High

Type:Optimization

Status: Merged (completed).

eth-cscs/DLA-Futurenumber 807

#807 In eth-cscs/DLA-Future;

· by rasolca was closed on Apr 18, 2023

computeTFactor parallelise computation with bulk (MC Local and Distributed)

Priority:Low

Type:Optimization

Status: Closed (abandoned).

eth-cscs/DLA-Futurenumber 798

#798 In eth-cscs/DLA-Future;

· by albestro was closed on Feb 12, 2025

Add tile sum kernel

Priority:High

Type:Optimization

Status: Merged (completed).

eth-cscs/DLA-Futurenumber 796

#796 In eth-cscs/DLA-Future;

· by rasolca was closed on Feb 15, 2023

ReductionToBand parallelise panel computation with bulk (both MC and GPU, both Local and Distributed)

Priority:High

Type:Optimization

Status: Merged (completed).

eth-cscs/DLA-Futurenumber 797

#797 In eth-cscs/DLA-Future;

· by albestro was closed on Apr 17, 2023

tridiagonal eigensolver -&gt; improve perf on AMD gpus

Priority:High

Type:Optimization

Type:Performance issue

Status: Merged (completed).

eth-cscs/DLA-Futurenumber 817

#817 In eth-cscs/DLA-Future;

· by rasolca was closed on Mar 27, 2023

Introduce sub-band in the eigensolver

Priority:Medium

Type:Optimization

Status: Merged (completed).

eth-cscs/DLA-Futurenumber 752

#752 In eth-cscs/DLA-Future;

· by rasolca was closed on Jan 3, 2023

Remove the need of the distributed row permutation in the tridiagonal solver.

Priority:High

Type:Optimization

Status: Merged (completed).

eth-cscs/DLA-Futurenumber 831

#831 In eth-cscs/DLA-Future;

· by rasolca was closed on Apr 17, 2023

Remove unnecessary <code>barrier::arrive_and_wait</code> from <code>reduction_to_band</code>

Type:Optimization

Status: Closed (abandoned).

eth-cscs/DLA-Futurenumber 841

#841 In eth-cscs/DLA-Future;

· by msimberg was closed on Apr 21, 2023