Commits

Branch selector

 master

User selector

All users

All time

Commit History

Commits on Jun 26, 2016

• Update README.md

  Erkaman

  authored
  816c48b

  Copy full SHA for 816c48b

• small cleanup

  Erkaman

  committed
  7cd78ed

  Copy full SHA for 7cd78ed

Commits on Jun 25, 2016

• Create LICENSE

  Erkaman

  authored
  2f4cc00

  Copy full SHA for 2f4cc00

• changed profile time to microseconds

  Erkaman

  committed
  d3fdd4b

  Copy full SHA for d3fdd4b

Commits on Jun 24, 2016

• replace tabs by spaces

  Erkaman

  committed
  67a0f17

  Copy full SHA for 67a0f17

• clean up

  Erkaman

  committed
  0b1b4d4

  Copy full SHA for 0b1b4d4

• added a different random data generator

  Erkaman

  committed
  b376b3e

  Copy full SHA for b376b3e

• more cleanup

  Erkaman

  committed
  c252482

  Copy full SHA for c252482

• removed cudpp and replaced it by chag

  Erkaman

  committed
  5119957

  Copy full SHA for 5119957

• cleanup

  Erkaman

  committed
  aac2630

  Copy full SHA for aac2630

Commits on Jun 23, 2016

• fixed so that we are only allocating plan once

  Erkaman

  committed
  0f846c8

  Copy full SHA for 0f846c8

Commits on Jun 22, 2016

• checkpoint

  Erkaman

  committed
  60cebfe

  Copy full SHA for 60cebfe

• optimized maskKernel, so that it uses shared memory

  Erkaman

  committed
  33b658c

  Copy full SHA for 33b658c

• now almost all arrays are global, and are allocated only once, in order to improve performance

  Erkaman

  committed
  c9e556a

  Copy full SHA for c9e556a

Commits on Jun 17, 2016

• simplified the rle implementation

  Erkaman

  committed
  f600eda

  Copy full SHA for f600eda

Commits on Jun 16, 2016

• added chag for slightly better performance

  Erkaman

  committed
  2f2853b

  Copy full SHA for 2f2853b

• tried making maskKernel more memory efficient

  Erkaman

  committed
  590e235

  Copy full SHA for 590e235

Commits on Jun 15, 2016

• add GPU profiling

  Erkaman

  committed
  c340348

  Copy full SHA for c340348

• implemented RLE on CPU

  Erkaman

  committed
  00d00d4

  Copy full SHA for 00d00d4

• separated host and device code for parle

  Erkaman

  committed
  e4ea6e1

  Copy full SHA for e4ea6e1

• decrease size from 2n to n

  Erkaman

  committed
  62046d1

  Copy full SHA for 62046d1

• now using HEMI to launch custom kernels

  Erkaman

  committed
  7f672ce

  Copy full SHA for 7f672ce

Commits on Jun 14, 2016

• now the algorithm uses CUDPP instead of hand-rolled kernels

  Erkaman

  committed
  9a8e33a

  Copy full SHA for 9a8e33a

Commits on Jun 8, 2016

• added units tests

  Erkaman

  committed
  3999a6b

  Copy full SHA for 3999a6b

Commits on Jun 7, 2016

• now d_blocksRunCount is computed without any crashes

  Erkaman

  committed
  311595d

  Copy full SHA for 311595d

Commits on Jun 6, 2016

• now you can specify the thread block size

  Erkaman

  committed
  14b6480

  Copy full SHA for 14b6480

• Merge branch 'master' of https://github.com/Erkaman/parle-cuda

  Erkaman

  committed
  f1d2ed1

  Copy full SHA for f1d2ed1

• replace all char:s with int:s

  Erkaman

  committed
  6c3058d

  Copy full SHA for 6c3058d

• intial commit

  Erkaman

  committed
  e9240ba

  Copy full SHA for e9240ba

Commits on Jun 5, 2016

• Create README.md

  Erkaman

  committed
  22ede44

  Copy full SHA for 22ede44