Commits · 6f782b6d68d1622657df98d99f5656cf92119d1f · tbgl / tools / arbd · GitLab

Snippets Groups Projects

Jan 06, 2021
- Macros to make NCCL optional · 98b2a95f
  cmaffeo2 authored 4 years ago
  
  98b2a95f
- Fixed -g argument so more than two GPUs can be provided · 176bbfab
  cmaffeo2 authored 4 years ago
  
  176bbfab
- Added profileStart to GrandBrownTown.cu and allowed nccl_broadcast to overlap... · 5bb84f6f
  cmaffeo2 authored 4 years ago
  
  Added profileStart to GrandBrownTown.cu and allowed nccl_broadcast to overlap with computeForce on GPU 0
  5bb84f6f
- Changed code slightly in attempt to mitigate impact of main loop when single GPU is used · c19ddc35
  cmaffeo2 authored 4 years ago
  
  c19ddc35
- Preliminary implementation of multi-gpu communication for nonbonded force... · b4d059e2
  cmaffeo2 authored 4 years ago
  
  Preliminary implementation of multi-gpu communication for nonbonded force calculation; energy kernel not yet implemented
  b4d059e2
- Use vectors in ComputeForce for device pointers (in preparation for multigpu) · 3b90d947
  cmaffeo2 authored 4 years ago
  
  3b90d947
- Refactored GPUManager so streams are owned by new GPU class that represents a single GPU · 1e528b5b
  cmaffeo2 authored 4 years ago
  
  1e528b5b
May 09, 2018
- Added CUDA streams to GPUManager class as static member data · 28ea36bd
  cmaffeo2 authored 6 years ago
  
  28ea36bd
May 09, 2017
- Made GPU assigment more intuitive · 7e2dc96f
  cmaffeo2 authored 8 years ago
  
  7e2dc96f
Nov 09, 2016
- Began reorganizing project · 3e2ac53a
  cmaffeo2 authored 8 years ago
  
  3e2ac53a
Nov 03, 2015
- initial commit: svn r22959 · ffdf082a
  cmaffeo2 authored 9 years ago
  
  ffdf082a