Historique des commits

Auteur SHA1 Message Date
  Olivier Aumage 98f48e3fd4 reenable modular_eager_prefetching il y a 8 ans
  Olivier Aumage 7bcdfb8ec5 enter relaxed section when acquiring the fifo component lock on push also il y a 8 ans
  Olivier Aumage 0b9ff3f17c update sched_modular_eager policy with new synchro scheme il y a 8 ans
  Olivier Aumage eaebdeebf1 update dm* policies with new synchro scheme il y a 8 ans
  Olivier Aumage 3953060972 update pheft with new synchro scheme il y a 8 ans
  Olivier Aumage 0ff769fe15 update push_local_task synchro il y a 8 ans
  Olivier Aumage 055cd401b0 pseudo merge 20752 from trunk il y a 8 ans
  Olivier Aumage 10ec2f8eec remove synchro changes from trunk/, since development has moved to branch synchro/ il y a 8 ans
  Olivier Aumage 370ee64978 merge from trunk il y a 8 ans
  Olivier Aumage 44bd171763 lock/unlock inversion il y a 8 ans
  Olivier Aumage 428ed04158 update heteroprio with new synchro scheme il y a 8 ans
  Samuel Thibault d5af97f058 more portable way to write a bogus pthread_t il y a 8 ans
  Samuel Thibault 4b6405da7d more portable way to write a bogus pthread_t il y a 8 ans
  Corentin Salingue 975d60ffcf fix shutdown when using cg option il y a 8 ans
  Olivier Aumage a83e24a99e restore graph_record variable on destroy il y a 8 ans
  Samuel Thibault 79d060efd8 Make sure to always update exp_end when updating exp_start, to notably fix transfer predictions, thanks Lionel Eyraud for the patch! il y a 8 ans
  Samuel Thibault c50d47db20 Fix use of deprecated type il y a 8 ans
  Olivier Aumage 86a4e26742 fix initialization and cleaning of priorities in sched_ctx il y a 8 ans
  Olivier Aumage 19cb83f85b merge from trunk il y a 8 ans
  Olivier Aumage 7f6c965ee1 update graph_test policy with new synchro scheme il y a 8 ans
  Samuel Thibault 686f368580 Avoid running pivoting with non-shared cl_args memory il y a 8 ans
  Olivier Aumage 30cf4c32fe update ws and lws with new synchro scheme il y a 8 ans
  Olivier Aumage b4c7229c1b remove debugging code il y a 8 ans
  Corentin Salingue 7ed5090182 Use 4 cores per MPI slaves when checking il y a 8 ans
  Samuel Thibault b61e7fe0a0 Use starpu_pthread_equal instead of hardcoding equality il y a 8 ans
  Samuel Thibault 0d9bf41d6f Fix execution with only OpenCL devices il y a 8 ans
  Samuel Thibault f45c16240f Also use physical number of cpus, since the users may use specific CPU cores il y a 8 ans
  Samuel Thibault 0612e2cf83 fix array sizes, notably important when mixing cuda and opencl devices, in which case they are not numbered from 0 il y a 8 ans
  Olivier Aumage fa947f5c97 update peager with new synchro scheme il y a 8 ans
  Olivier Aumage 3c6301dc94 policy random is ok with new synchro scheme il y a 8 ans