Code

CUDA optimized ASSET#351

Code

Merged

dizcza merged 21 commits intoNeuralEnsemble:masterNeuralEnsemble/elephant:masterfrom INM-6:cuda/assetINM-6/elephant:cuda/assetCopy head branch name to clipboardJan 29, 2021

+803-239

Commits

Commits on Sep 10, 2020

• CUDA optimized ASSET

  dizcza

  committed
  617b9b8

  Copy full SHA for 617b9b8

• wrap asset import

  dizcza

  committed
  a009c92

  Copy full SHA for a009c92

• check for is_cuda_available

  dizcza

  committed
  393eebc

  Copy full SHA for 393eebc

• python2

  dizcza

  committed
  248b74b

  Copy full SHA for 248b74b

• fixed _is_cuda_available

  dizcza

  committed
  ef4f07e

  Copy full SHA for ef4f07e

Commits on Oct 9, 2020

• double float everywhere

  dizcza

  committed
  b49a89c

  Copy full SHA for b49a89c

• Merge branch 'master' into cuda/asset

  dizcza

  committed
  64d7c20

  Copy full SHA for 64d7c20

• precision_printf

  dizcza

  committed
  933ef38

  Copy full SHA for 933ef38

• log_du always float

  dizcza

  committed
  ca6f99e

  Copy full SHA for ca6f99e

• upd doc

  dizcza

  committed
  9549983

  Copy full SHA for 9549983

Commits on Oct 12, 2020

• reduced branch divergence, atomicAdd misses

  dizcza

  committed
  539ae70

  Copy full SHA for 539ae70

• manually set CWR_LOOPS

  dizcza

  committed
  4fa486c

  Copy full SHA for 4fa486c

Commits on Oct 13, 2020

• fixed codefactor issues

  dizcza

  committed
  ddf39a5

  Copy full SHA for ddf39a5

• cuda header doc

  dizcza

  committed
  a27be5d

  Copy full SHA for a27be5d

Commits on Oct 14, 2020

• software implementation of atomicAdd for doubles; doubles are by default

  dizcza

  committed
  1a9fdbd

  Copy full SHA for 1a9fdbd

• get_cuda_capability_major

  dizcza

  committed
  378291f

  Copy full SHA for 378291f

• added test_precision

  dizcza

  committed
  148b004

  Copy full SHA for 148b004

Commits on Dec 6, 2020

• Merge branch 'master' into cuda/asset

  dizcza

  committed
  f5e0165

  Copy full SHA for f5e0165

Commits on Jan 13, 2021

• clip cuda P_total array

  dizcza

  committed
  7c15545

  Copy full SHA for 7c15545

Commits on Jan 21, 2021

• Merge branch 'master' into cuda/asset

  dizcza

  committed
  503e7ec

  Copy full SHA for 503e7ec

• reverted python2 compatibility

  dizcza

  committed
  d2ca4b1

  Copy full SHA for d2ca4b1