Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

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

• Fix the MatrixMultiplyBatch test so that we are testing the result

  

  umar456

  authored and

  pradeep

  committed
  a70a00f

  Copy full SHA for a70a00f

Commits on Feb 21, 2020

• Fix CPU OpenCL blas batching

  umar456

  committed
  ca90115

  Copy full SHA for ca90115

• Fix matmul on Intel OpenCL when passing same array as input

  Show description for 99bffc9
  umar456

  committed
  99bffc9

  Copy full SHA for 99bffc9

Commits on Feb 15, 2020

• Format source files per clang-format style

  

  pradeep

  committed
  2c819be

  Copy full SHA for 2c819be

Commits on Jan 28, 2020

• Delegate all non-batch matmul to cuBLAS GEMM API

  Show description for e9ce3f2
  

  pradeep

  committed
  e9ce3f2

  Copy full SHA for e9ce3f2

Commits on Aug 26, 2019

• Add support for gemm/dot for CPU using conversion to float (#2624)

  Show description for 3dc3da1
  

  umar456

  authored and

  pradeep

  committed
  3dc3da1

  Copy full SHA for 3dc3da1

Commits on Jul 12, 2019

• Fix half ComputeType for CudaCompute 6.1 GPUs

  

  WilliamTambellini

  authored and

  umar456

  committed
  f596f8c

  Copy full SHA for f596f8c

Commits on Jul 8, 2019

• Fix several leaks in tests

  

  umar456

  authored and

  pradeep

  committed
  73093cd

  Copy full SHA for 73093cd

Commits on Jul 5, 2019

• af_matmul for f16

  

  WilliamTambellini

  authored and

  umar456

  committed
  4621bd6

  Copy full SHA for 4621bd6

Commits on Jun 28, 2019

• Initial f16 support (#2413)

  Show description for a23c00b
  umar456

  authored
  a23c00b

  Copy full SHA for a23c00b

Commits on Jun 20, 2019

• Add af_gemm

  Show description for 4159b1a
  

  Mark Poscablo

  authored and

  umar456

  committed
  4159b1a

  Copy full SHA for 4159b1a

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 Dec 27, 2018

• Use mkl gemm batch in CPU backend when appropriate

  Show description for 5cc9967
  

  pradeep

  committed
  5cc9967

  Copy full SHA for 5cc9967

Commits on Dec 19, 2018

• Update error tolerances for batched matmul. Required to pass on 6.x

  Show description for 4d003fc
  

  umar456

  authored and

  mark-poscablo

  committed
  4d003fc

  Copy full SHA for 4d003fc

Commits on Oct 18, 2018

• Add broadcast batching for matmul

  

  pradeep

  authored and

  umar456

  committed
  42d64f5

  Copy full SHA for 42d64f5

Commits on Aug 2, 2018

• Incorporated the ASSERT_*_NEAR macros to some of the tests

  Show description for 5705a53
  

  mark-poscablo

  authored and

  umar456

  committed
  5705a53

  Copy full SHA for 5705a53

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 Dec 1, 2017

• FEAT: Add support for batched matrix multiply

  

  pavanky

  authored and

  umar456

  committed
  9447258

  Copy full SHA for 9447258

Commits on Aug 23, 2017

• BUGIFX: Fixes issue with gemv when the rhs is an indexed vector

  pavanky

  committed
  e82f0e5

  Copy full SHA for e82f0e5

Commits on Sep 24, 2015

• Added enviornment variable to disable multi gpu tests

  shehzan10

  committed
  d639289

  Copy full SHA for d639289

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 19, 2015

• STYLE: Changing dim_type to dim_t

  Show description for b685fa6
  pavanky

  committed
  b685fa6

  Copy full SHA for b685fa6

Commits on May 14, 2015

• SOLVE, MATMUL and INVERSE now use af_mat_prop

  Show description for 6099075
  pavanky

  committed
  6099075

  Copy full SHA for 6099075

• Changing af_blas_transpose to af_transpose_t

  Show description for 8595c86
  pavanky

  committed
  8595c86

  Copy full SHA for 8595c86

Commits on Apr 1, 2015

• Create one cuBlasHandle per GPU. Tests

  umar456

  committed
  7e868b6

  Copy full SHA for 7e868b6

Commits on Mar 9, 2015

• Remove additional c++11 features from test

  umar456

  committed
  9725a7e

  Copy full SHA for 9725a7e

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

• Faster DBN convergence. Test updates

  umar456

  committed
  dadc5e8

  Copy full SHA for dadc5e8

Commits on Feb 25, 2015

• Destroy af_array at the end of tests

  pavanky

  committed
  1b093a2

  Copy full SHA for 1b093a2

Commits on Dec 16, 2014

• Changing af::af_cfloat to af::cfloat for C++ API

  pavanky

  committed
  5654c9c

  Copy full SHA for 5654c9c

Commits on Dec 11, 2014

• Adding noDoubleTests condition to all tests

  shehzan10

  committed
  4b5a66e

  Copy full SHA for 4b5a66e