Commits · cbd05a45c38ce9ba2cc9de676e9bac739b02acb9 · TNL / tnl-dev · GitLab

This project is archived. Its data is read-only.

Aug 17, 2019
- Merge branch 'JK/multireduction' into 'develop' · cbd05a45
  Jakub Klinkovský authored Aug 17, 2019
  cbd05a45
- Implemented distributed prefix-sum · d13a2d18
  Jakub Klinkovský authored Aug 16, 2019
  d13a2d18
- PrefixSum: separate first and second phase for OpenMP implementation and... · 174ad5fd
  Jakub Klinkovský authored Aug 16, 2019
  174ad5fd
- CUDA prefix-sum: separated the implementation of the first and second phase · 2c40015f
  Jakub Klinkovský authored Aug 16, 2019
  
  2c40015f
- CUDA prefix-sum: moved gridShift from the first phase to the second phase · ac2ee07e
  Jakub Klinkovský authored Aug 16, 2019
  
  ac2ee07e
- Replaced static member variables in CudaPrefixSumKernelLauncher with static getters · 1fe62640
  Jakub Klinkovský authored Aug 15, 2019
  
  1fe62640
- Added default stream synchronizations after kernel launches in CudaPrefixSumKernel.h · af6d1d6b
  Jakub Klinkovský authored Aug 15, 2019
  
  af6d1d6b
- Removed volatile reduction from PrefixSum and updated the normal reduction operation · 8d0d2638
  Jakub Klinkovský authored Aug 14, 2019
  8d0d2638
- Added prefix-sum to BLAS benchmarks · 27631930
  Jakub Klinkovský authored Aug 14, 2019
  
  27631930
- Implemented parallel prefix-sum with OpenMP · 7cc55dee
  Jakub Klinkovský authored Aug 14, 2019
  7cc55dee
- Benchmarks: compute sample standard deviation of the measured computation times · 2bea9311
  Jakub Klinkovský authored Aug 14, 2019
  
  2bea9311
- Removed timing parameter from benchmarks · e6e6cf46
  Jakub Klinkovský authored Aug 14, 2019
  e6e6cf46
- Benchmarks: added scalar multiplication with BLAS · 232be124
  Jakub Klinkovský authored Aug 14, 2019
  
  232be124
- Optimized OpenMP thread counts for reduction and multireduction · 505d0b68
  Jakub Klinkovský authored Aug 14, 2019
  
  505d0b68
- Ugly workaround for nvcc's stupid modification of `new` expressions · 32c69a11
  Jakub Klinkovský authored Aug 12, 2019
  
  32c69a11
- Replaced custom lambda functions with instances of STL types where possible · 0a57393f
  Jakub Klinkovský authored Aug 12, 2019
  
  0a57393f
- Changed reduction operation to use functions with `return a + b` instead of `a += b` · e20a0930
  Jakub Klinkovský authored Aug 12, 2019
  e20a0930
- Removed VectorOperations class which is now useless · d0fc1bb7
  Jakub Klinkovský authored Aug 12, 2019
  d0fc1bb7
- Removed ReductionOperations.h · 1777e488
  Jakub Klinkovský authored Aug 11, 2019
  
  1777e488
- Removed volatile reduction completely · 13b89a71
  Jakub Klinkovský authored Aug 11, 2019
  
  13b89a71
- Found a way to avoid using volatile in CUDA reduction: __syncwarp() · cbc2fff9
  Jakub Klinkovský authored Aug 11, 2019
  cbc2fff9
- Rewritten multireduction with lambda functions · b74a24d2
  Jakub Klinkovský authored Aug 11, 2019
  
  b74a24d2
- Style changes in the code for reduction · e470040a
  Jakub Klinkovský authored Aug 11, 2019
  
  e470040a
- Removed reduction and multireduction declarations for MIC · 1a82b047
  Jakub Klinkovský authored Aug 11, 2019
  1a82b047
- Execute tests in parallel · 95b2d990
  Jakub Klinkovský authored Aug 17, 2019
  
  95b2d990
- Tests: forced using a unique file name in each test · a3ba2469
  Jakub Klinkovský authored Aug 17, 2019
  a3ba2469
- Disabled long double in tnl-diff, tnl-init, tnl-lattice-init and tnl-view · 307b6650
  Jakub Klinkovský authored Aug 17, 2019
  307b6650
Aug 16, 2019
- Enabled link-time optimizations · 402d5bab
  Jakub Klinkovský authored Aug 14, 2019
  
  402d5bab
- Removed unnecessary build options from CMakeLists.txt · ee29cf08
  Jakub Klinkovský authored Aug 14, 2019
  
  ee29cf08
- Cleaned up communicators in tnl-init · 9cda013b
  Jakub Klinkovský authored Aug 13, 2019
  
  9cda013b
- Fixed move-constructor in Array · 569ccc9a
  Jakub Klinkovský authored Aug 16, 2019
  
  569ccc9a
Aug 14, 2019
- Merge branch 'ndarray' into 'develop' · 3e785f01
  Jakub Klinkovský authored Aug 14, 2019
  3e785f01
- DistributedNDArray: added methods getLocalIndexer, getLocalView,... · 05634323
  Jakub Klinkovský authored Aug 14, 2019
  05634323
- Added overlaps to NDArrayIndexer · e4e780e1
  Jakub Klinkovský authored Aug 14, 2019
  e4e780e1
- Fixed StaticFor for loops with a large iterations count · daebbce1
  Jakub Klinkovský authored Aug 14, 2019
  
  daebbce1
- Removed conditional per-device Permutation and SliceInfo setting from NDArray and SlicedNDArray · 55ded6ad
  Jakub Klinkovský authored Apr 19, 2019
  
  55ded6ad
- Split NDArrayIndexer from NDArrayStorage and NDArrayView · de8f1eb8
  Jakub Klinkovský authored Apr 14, 2019
  
  de8f1eb8
- Added templated assignment operators for NDArray and NDArrayView · fd4d8429
  Jakub Klinkovský authored Apr 09, 2019
  fd4d8429
- Copied getSubarrayView method from NDArrayView to NDArray · 560aba80
  Jakub Klinkovský authored Apr 09, 2019
  
  560aba80
- Fixed setSize method in LocalBeginsHolder · e885aa29
  Jakub Klinkovský authored Apr 06, 2019
  
  e885aa29