Code

Add GPU functionality#126

Code

Merged

computablee merged 90 commits intomaincomputablee/DotMP:mainfrom release/v2.0.0computablee/DotMP:release/v2.0.0Copy head branch name to clipboardNov 27, 2023

+1,939-29

Commits

Commits on Oct 13, 2023

• initial commit for GPU API

  computablee

  committed
  2529f7a

  Copy full SHA for 2529f7a

• rename ParallelTests to CPUTests

  computablee

  committed
  4d04a5f

  Copy full SHA for 4d04a5f

• add GPU tests

  computablee

  committed
  f3ddf4c

  Copy full SHA for f3ddf4c

• add reference to GPU class

  computablee

  committed
  a3a1d8d

  Copy full SHA for a3a1d8d

• add ILGPU as dependency

  computablee

  committed
  e31b723

  Copy full SHA for e31b723

• redo devicehandle api

  computablee

  committed
  2d790e1

  Copy full SHA for 2d790e1

• major updates, rewrites

  computablee

  committed
  c8f8556

  Copy full SHA for c8f8556

• final testing idea for base functionality

  computablee

  committed
  5a81011

  Copy full SHA for 5a81011

• add cecil as dependency

  computablee

  committed
  2b8e2b4

  Copy full SHA for 2b8e2b4

Commits on Oct 14, 2023

• new API for DotMP.GPU

  computablee

  committed
  7853b47

  Copy full SHA for 7853b47

• got basic tester working!

  computablee

  committed
  313c9ef

  Copy full SHA for 313c9ef

• rename namespace to include gpu

  computablee

  committed
  054a62d

  Copy full SHA for 054a62d

• remove using for testing

  computablee

  committed
  45197c8

  Copy full SHA for 45197c8

Commits on Oct 16, 2023

• update GPU tests for new GPU API

  computablee

  committed
  58474a8

  Copy full SHA for 58474a8

• implement new API for parallelfor bodies

  computablee

  committed
  2b3d7af

  Copy full SHA for 2b3d7af

Commits on Oct 17, 2023

• Merge pull request #95 from computablee/gpu

  Show description for 42bce7c
  computablee

  authored
  42bce7c

  Copy full SHA for 42bce7c

Commits on Nov 10, 2023

• test using new GPU data transfer API

  computablee

  committed
  61b93af

  Copy full SHA for 61b93af

• implement new memory model

  computablee

  committed
  60a1c9f

  Copy full SHA for 60a1c9f

• tidying up duplicate code

  computablee

  committed
  49706b3

  Copy full SHA for 49706b3

• remove unnecessary implicit operators

  computablee

  committed
  68092ff

  Copy full SHA for 68092ff

• add overloads for dispatching loops with 5 or 6 variables

  computablee

  committed
  0ac98e5

  Copy full SHA for 0ac98e5

• add parfor_dump to gitignore

  computablee

  committed
  b76af51

  Copy full SHA for b76af51

• add parallel for overload code gen

  computablee

  committed
  c2f591a

  Copy full SHA for c2f591a

• add overloads for up to 16 kernel variables

  computablee

  committed
  dd56407

  Copy full SHA for dd56407

• move python to python folder

  computablee

  committed
  82165da

  Copy full SHA for 82165da

• update python to only generate up to 13 data params

  computablee

  committed
  89b5e44

  Copy full SHA for 89b5e44

• add parfor overloads for up to 13 data parameters

  computablee

  committed
  7b90933

  Copy full SHA for 7b90933

• fix cardinality in documentation

  computablee

  committed
  8a87b5c

  Copy full SHA for 8a87b5c

• implement 2D arrays into GPUArray and Buffer objects

  computablee

  committed
  ef3035f

  Copy full SHA for ef3035f

• initial commit of GPU heat transfer benchmark

  computablee

  committed
  4eb4445

  Copy full SHA for 4eb4445

• add nocopy behavior

  computablee

  committed
  b5365b7

  Copy full SHA for b5365b7

• fix exception on OpenCL devices

  computablee

  committed
  c092eb9

  Copy full SHA for c092eb9

• better accelerator selection

  computablee

  committed
  f66b227

  Copy full SHA for f66b227

• fix copying back 2D arrays

  computablee

  committed
  79c4a35

  Copy full SHA for 79c4a35

• added start offset for index calculations

  computablee

  committed
  528cba0

  Copy full SHA for 528cba0

• get HeatTransferVerify running properly

  computablee

  committed
  6121c0d

  Copy full SHA for 6121c0d

• add LGPL license header

  computablee

  committed
  62b5ebe

  Copy full SHA for 62b5ebe

• prepare benchmark

  computablee

  committed
  d11913c

  Copy full SHA for d11913c

• remove dispose for benchmarking

  computablee

  committed
  8b12422

  Copy full SHA for 8b12422

• change to .net 6

  computablee

  committed
  20eebd6

  Copy full SHA for 20eebd6

Commits on Nov 11, 2023

• add attributes to prevent exceptions when collecting code coverage

  computablee

  committed
  708d9d5

  Copy full SHA for 708d9d5

• fix bug

  computablee

  committed
  2e328e7

  Copy full SHA for 2e328e7

• mark old single/ordered/critical regions as obsolete, implement new versions

  computablee

  committed
  f3b7735

  Copy full SHA for f3b7735

• remove sln files

  computablee

  committed
  429bde2

  Copy full SHA for 429bde2

• exclude obsolete methods from code coverage

  computablee

  committed
  fab2c47

  Copy full SHA for fab2c47

• add missing using

  computablee

  committed
  8b40582

  Copy full SHA for 8b40582

• use new critical/ordered/single methods

  computablee

  committed
  594b6ba

  Copy full SHA for 594b6ba

• testing forcollapse performance in heat transfer, will fully implement forcollapse later

  computablee

  committed
  5e90c89

  Copy full SHA for 5e90c89

• implement T4 template for acceleratorhandler

  computablee

  committed
  348ad0c

  Copy full SHA for 348ad0c

• ignore generated acceleratorhandler file

  computablee

  committed
  91f1692

  Copy full SHA for 91f1692

• add T4 stuff

  computablee

  committed
  6540825

  Copy full SHA for 6540825

• delete now unnecessary files

  computablee

  committed
  a8872d1

  Copy full SHA for a8872d1

• more autogen

  computablee

  committed
  cdf34c1

  Copy full SHA for cdf34c1

• revert collapse

  computablee

  committed
  267a789

  Copy full SHA for 267a789

• remove excess newlines

  computablee

  committed
  3d610b4

  Copy full SHA for 3d610b4

Commits on Nov 12, 2023

• get parallelfor t4 gen working

  computablee

  committed
  d03d31b

  Copy full SHA for d03d31b

• implement collapsed for loops

  computablee

  committed
  945f7e7

  Copy full SHA for 945f7e7

• remove erroneous comment line

  computablee

  committed
  d90279e

  Copy full SHA for d90279e

• test with 500x500 instead of 514x514

  computablee

  committed
  f22ac63

  Copy full SHA for f22ac63

• properly handle loops not divisible by block size

  computablee

  committed
  aeba3b8

  Copy full SHA for aeba3b8

• turn array bounds into off-256-divisble size for better testing

  computablee

  committed
  a93d9fa

  Copy full SHA for a93d9fa

Commits on Nov 13, 2023

• enable more optimizations

  computablee

  committed
  11d4baf

  Copy full SHA for 11d4baf

• add GPU kernel launch overhead benchmark

  computablee

  committed
  1c12e39

  Copy full SHA for 1c12e39

• begin progress towards better index integration

  computablee

  committed
  f2b2360

  Copy full SHA for f2b2360

• add test for forcollapse

  computablee

  committed
  1f20188

  Copy full SHA for 1f20188

• tidy up calls

  computablee

  committed
  141aafd

  Copy full SHA for 141aafd

• implement caching of indices

  computablee

  committed
  b5ca731

  Copy full SHA for b5ca731

• add support for 3D buffers

  computablee

  committed
  2275b97

  Copy full SHA for 2275b97

• migrate to new index technique

  computablee

  committed
  81c39ba

  Copy full SHA for 81c39ba

• WIP

  computablee

  committed
  af259ec

  Copy full SHA for af259ec

• new index technique via index caching

  computablee

  committed
  7e4070a

  Copy full SHA for 7e4070a

• update benchmarks

  computablee

  committed
  a6d0072

  Copy full SHA for a6d0072

• run dotnet format

  computablee

  committed
  ca0b54a

  Copy full SHA for ca0b54a

Commits on Nov 14, 2023

• add support for .NET 8

  computablee

  committed
  1c72a2f

  Copy full SHA for 1c72a2f

Commits on Nov 16, 2023

• comments, optimizations

  computablee

  committed
  8432ed7

  Copy full SHA for 8432ed7

• temporary

  computablee

  committed
  97571c5

  Copy full SHA for 97571c5

• add optimizations

  computablee

  committed
  f87b687

  Copy full SHA for f87b687

Commits on Nov 26, 2023

• Merge pull request #125 from computablee/gpu

  Show description for 15e814b
  computablee

  authored
  15e814b

  Copy full SHA for 15e814b

• resolve merge conflicts

  computablee

  committed
  3ea76d2

  Copy full SHA for 3ea76d2

• fix merge conflicts

  computablee

  committed
  9df57a0

  Copy full SHA for 9df57a0

• fix compile issues

  computablee

  committed
  e0b139e

  Copy full SHA for e0b139e

• uncomment 1D and 3D arrays

  computablee

  committed
  4561480

  Copy full SHA for 4561480

• fix linting errors

  computablee

  committed
  8f9450d

  Copy full SHA for 8f9450d

• fix constructor

  computablee

  committed
  ca35b92

  Copy full SHA for ca35b92

• fix bug with getnumthreads

  computablee

  committed
  ffc09ad

  Copy full SHA for ffc09ad

• add test for 3D buffers and 3D collapsed for loops

  computablee

  committed
  b5dc84c

  Copy full SHA for b5dc84c

Commits on Nov 27, 2023

• remove debug console writeline

  computablee

  committed
  cc314e9

  Copy full SHA for cc314e9

• fix 3D allocation (ZY->XY stride)

  computablee

  committed
  6541e09

  Copy full SHA for 6541e09

• omit unreachable code

  computablee

  committed
  b349701

  Copy full SHA for b349701

• modify site publishing to build templates before running doxygen

  computablee

  committed
  6fbab81

  Copy full SHA for 6fbab81