Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Mar 11, 2025

• signed 8-bit integer support

  

  verstatx

  authored and

  syurkevi

  committed
  cdbbc75

  Copy full SHA for cdbbc75

Commits on Jan 14, 2025

• Extend test for convolve2GradientNN function to verify zero padding fix in PR 3519 (#3631)

  christophe-murphy

  authored
  5c2ea29

  Copy full SHA for 5c2ea29

Commits on Jan 9, 2025

• 3580 bug investigate test failures when running with cuda 126 (#3588)

  Show description for f4edcf2
  

  christophe-murphy

  and

  FloopCZ

  authored
  f4edcf2

  Copy full SHA for f4edcf2

Commits on Jan 21, 2023

• Update convolve tests tolerances for floating point types

  umar456

  committed
  e5b1047

  Copy full SHA for e5b1047

Commits on Oct 4, 2022

• Update clang-format version on github workflow

  umar456

  committed
  e401fce

  Copy full SHA for e401fce

• Update deprecated macro from GTest. Add workaround for older versions

  umar456

  committed
  83edd09

  Copy full SHA for 83edd09

Commits on Aug 18, 2021

• Fix reference count if array used in JIT operations.

  Show description for a57b291
  

  umar456

  authored and

  pradeep

  committed
  a57b291

  Copy full SHA for a57b291

Commits on May 20, 2021

• Increase half type error tolerance to 0.07 for convolve tests

  

  pradeep

  committed
  34833d1

  Copy full SHA for 34833d1

Commits on Aug 12, 2020

• Fix problems in RNG for older compute architectures with fp16

  

  umar456

  authored and

  pradeep

  committed
  fe0c8d5

  Copy full SHA for fe0c8d5

Commits on Apr 28, 2020

• Create a test library to speed up test compilation

  

  umar456

  authored and

  pradeep

  committed
  fd01d59

  Copy full SHA for fd01d59

Commits on Apr 22, 2020

• Address all warnings with -Wall flags in GCC 9.3

  

  umar456

  authored and

  pradeep

  committed
  3f56eb7

  Copy full SHA for 3f56eb7

Commits on Apr 13, 2020

• Make cudnn dependency optional (#2836)

  Show description for fc99193
  

  syurkevi

  and

  umar456

  authored
  fc99193

  Copy full SHA for fc99193

Commits on Apr 8, 2020

• fix zero padding in convolve2NN (#2820)

  Show description for decde4e
  syurkevi

  authored
  decde4e

  Copy full SHA for decde4e

Commits on Feb 15, 2020

• Format source files per clang-format style

  

  pradeep

  committed
  2c819be

  Copy full SHA for 2c819be

Commits on Aug 29, 2019

• Dilated convolve2 and backward gradients (#2359)

  Show description for dae7c16
  

  syurkevi

  authored and

  umar456

  committed
  dae7c16

  Copy full SHA for dae7c16

Commits on May 30, 2019

• Fix convolve3 launch configuration in CUDA backend

  Show description for b5286ef
  

  pradeep

  authored and

  umar456

  committed
  b5286ef

  Copy full SHA for b5286ef

Commits on Jan 24, 2019

• Use SUPPORTED_TYPE_CHECK macro to test for supported type

  Show description for 9763d2f
  umar456

  committed
  9763d2f

  Copy full SHA for 9763d2f

Commits on Jan 1, 2019

• Update all source files based on clang-format rules

  

  umar456

  authored and

  pradeep

  committed
  7128623

  Copy full SHA for 7128623

Commits on Aug 1, 2018

• Using the new ASSERT*_ARRAY_EQ macros for tests

  Show description for ca61e34
  

  Mark Poscablo

  authored and

  pradeep

  committed
  ca61e34

  Copy full SHA for ca61e34

Commits on Jun 19, 2018

• Cleanup tests. Fix warnings when running with asan. (#2199)

  Show description for 7c6fa42
  

  mlloreda

  authored and

  umar456

  committed
  7c6fa42

  Copy full SHA for 7c6fa42

Commits on Aug 29, 2017

• reduce number of launched blocks

  

  syurkevi

  authored and

  pavanky

  committed
  66f010f

  Copy full SHA for 66f010f

Commits on Jun 19, 2017

• Add matching af_free_host and use vector instead of new in tests

  

  umar456

  authored and

  pavanky

  committed
  7cf4f9a

  Copy full SHA for 7cf4f9a

• convolve tests for complex types

  

  pradeep

  authored and

  mlloreda

  committed
  18278ea

  Copy full SHA for 18278ea

Commits on Jan 2, 2016

• Fix clang warnings (std::abs, pragma ignores)

  Show description for c2d7e42
  shehzan10

  committed
  c2d7e42

  Copy full SHA for c2d7e42

Commits on Nov 3, 2015

• Add intl/uintl to convolve and fftconvolve

  

  shehzan10

  authored and

  umar456

  committed
  401f130

  Copy full SHA for 401f130

Commits on Sep 22, 2015

• Added short (s16) and ushort (u16) types for CPU

  Show description for ebfe9e5
  shehzan10

  committed
  ebfe9e5

  Copy full SHA for ebfe9e5

Commits on Jun 8, 2015

• BUGFIX/TEST: fftConvolve now does multi dimensional batching properly

  Show description for 55bf057
  pavanky

  committed
  55bf057

  Copy full SHA for 55bf057

Commits on May 28, 2015

• TEST: Adding GFOR tests for convolve

  pavanky

  committed
  2631d68

  Copy full SHA for 2631d68

Commits on May 25, 2015

• API: rename af_weak_copy/af_destroy_array to af_retain/af_release

  Show description for ba75031
  umar456

  committed
  ba75031

  Copy full SHA for ba75031

Commits on May 15, 2015

• FEAT: convolve automatically switches to frequency domain when necessary

  pavanky

  committed
  9002e68

  Copy full SHA for 9002e68

Commits on May 14, 2015

• Making the expand parameter an enum for convolutions

  pavanky

  committed
  7990c2c

  Copy full SHA for 7990c2c

Commits on Apr 18, 2015

• Remove unnecessary template parameters from tests

  umar456

  committed
  4e7338a

  Copy full SHA for 4e7338a

Commits on Mar 19, 2015

• Revised documentation for convolve

  Show description for e379a90
  

  Pradeep

  committed
  e379a90

  Copy full SHA for e379a90

Commits on Mar 8, 2015

• Make tests C++03 complient.

  Show description for cc07a2c
  umar456

  committed
  cc07a2c

  Copy full SHA for cc07a2c

Commits on Mar 6, 2015

• FEAT: filter in convolutions is cast to the accum type

  pavanky

  committed
  b1e87b3

  Copy full SHA for b1e87b3