#64 CQuery: parallel

Fechado
4 anos atrás foi aberto por ranides · 4 comentários

Review CQuery and prepare for correct parallelization.

Please remember about all atomic variables from #63.

Be prepared for derived classes which:

  • use internally parallel stream if possible
  • use internally parallel spliterator if possible

API for:

  • turn on/off query parallelization
  • check parallelization feature

Implement parallel processing for:

  • CQList
  • CQListSupplier
  • CQArray
  • CQArraySupplier
  • CQCollection
  • CQCollectionSupplier
  • CQAppend ?

Implement parallel processing for derived streams: prefer parallel if parent stream supports it.

Review CQuery and prepare for correct parallelization. Please remember about all atomic variables from #63. Be prepared for derived classes which: - use internally parallel stream if possible - use internally parallel spliterator if possible API for: - turn on/off query parallelization - check parallelization feature Implement parallel processing for: - CQList - CQListSupplier - CQArray - CQArraySupplier - CQCollection - CQCollectionSupplier - CQAppend ? Implement parallel processing for derived streams: prefer parallel if parent stream supports it.
ranides comentado 4 anos atrás
Proprietário
  • parallel for-each for list, array, collection and spliterator
- [x] parallel for-each for list, array, collection and **spliterator**
ranides comentado 4 anos atrás
Proprietário
  • coverage
- [ ] coverage
ranides comentado 4 anos atrás
Proprietário

todo: we should mark additional feature: "hasFastParallelEach" - it should be returned if implementation of "each" is paralell sensitive and can use internally more than one thread.

In such situations, when fast-parallel-each is supported, we should prefer direct #each call instead of some other generic methods from support package.

Maybe, but this is really maybe: write additional method forEachParallel & whileEachParallel?

todo: we should mark additional feature: "hasFastParallelEach" - it should be returned if implementation of "each" is paralell sensitive and can use internally more than one thread. In such situations, when fast-parallel-each is supported, we should prefer direct #each call instead of some other generic methods from `support` package. Maybe, but this is *really* maybe: write additional method `forEachParallel` & `whileEachParallel`?
ranides comentado 4 anos atrás
Proprietário

Please note: our coverage tests should not profile performance problems which are known and delegated to #68

Please note: our coverage tests should not profile performance problems which are known and delegated to #68
ranides citou este problema em um commit 4 anos atrás
Faça login para participar desta conversação.
Sem milestone
Não atribuída
1 participantes
Carregando...
Cancelar
Salvar
Ainda não há conteúdo.