[arrayfire] annotated tag upstream/v3.1.0+dfsg1 created (now c4e6e86)

Ghislain Vaillant ghisvail-guest at moszumanska.debian.org
Mon Sep 21 19:29:50 UTC 2015


This is an automated email from the git hooks/post-receive script.

ghisvail-guest pushed a change to annotated tag upstream/v3.1.0+dfsg1
in repository arrayfire.

        at  c4e6e86   (tag)
   tagging  54749215e4def3bce369bc09ab6cdddb6a3828d5 (commit)
  replaces  upstream/3.0_beta
 tagged by  Ghislain Antony Vaillant
        on  Tue Sep 15 01:40:19 2015 +0100

- Log -----------------------------------------------------------------
Upstream version v3.1.0+dfsg1
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=6Ejk
-----END PGP SIGNATURE-----

Brian Kloppenborg (36):
      Add OS and architecture information to generated installer files.
      Fix OpenCL not found for examples.
      Remove FindArrayFire, script is automatically generated.
      Add source package. Create using
      Merge pull request #579 from pavanky/ocl_helpers
      Doc for using ArrayFire with external OpenCL code
      Merge pull request #583 from pentschev/err_cufft_fix
      Add self-extracting zip for Linux.
      Add DEB and RPM packages for Linux.
      Build DEB and RPM packages only on Linux.
      Pull packging out of main CMake file.
      Remove package directory, move packaging to CPack.txt file.
      Add installation documentation for Windows, OSX, and Linux.
      Moved installation instructions to Doxygen. Updated links.
      Fix markdown for Doxygen.
      Add Windows (MSVS) and Linux (CMake/Make) usage instructions.
      Add lapacke dependency for Debian-based distributions.
      Document AF CMake variables. Add non-standard install instructions.
      Add lapack(e) dependency on Fedora.
      Add compiliation instructions to examples. This resolves #606.
      Only build examples if full backend is found, fixes #738.
      Fix missing includes in standalone compilation.
      Sort member functions alphabetically.
      Revamp the getting started guide.
      Remove extra FindOpenCL.cmake file.
      Fix incorrect array construction.
      Fix missing init of vector.
      Skip complex<float> object creation in favor of direct init.
      Move example install directory.
      Fix typo in installer mode.
      Fix path for CMake config files.
      Added citation and acknowledgements.
      Move acknowledgements to separate file.
      Point all acknowledgement documents to GitHub repo.
      Automatically enable and build CUDA and OpenCL backends by default.
      Suppress FIND output for non-essential libraries.

Filipe Maia (14):
      Fix missing dereference.
      Ensure af_get_last_error and af_err_to_string are exported by including the public header.
      Add missing definitions for array_proxy operators
      Add missing constant() declaration
      Add support for s64 and u64 arrays
      Requires some knowledge of array_proxy_impl. Fixes #787
      Implement CUDA complex dot product
      Add complex dot product for CPU backend
      Add complex dot product for OpenCL backend
      Add forgotten instatiations
      A few functions (e.g. setunique) were not being exported to the library
      Fix double free issue.
      Add test for index copy assignment.
      Add move constructor and move assignment op

Gaëtan Lehmann (7):
      simplify freeimage cmake code
      set the lib version
      fix 1000+ warnings about unused function with clang
      replace tabs with 4 spaces
      fix the missing external projects byproducts
      fix the condition used to set the external projects by products
      fix missing include

Gaëtan Lehmann (1):
      download and build the opencl dependencies with cmake

Ghislain Antony Vaillant (10):
      TYPO: fix ArrayFire URL
      add module centralizing all install paths
      use new install path variables
      make paths overridable
      fix #828: remove unnecessary include of a cmake module
      fix #957: remove unused dtype trait for size_t
      update cl.hpp to upstream version 1.2.7
      fix instantiation of Platform objects
      fix instantiation of cl::Platform objects
      dfsg clean

Keno Fischer (5):
      Allow building against 64bit index OpenBlas
      Add support for BLAS symbol renaming
      Also use the void* interface for MKL on windows
      Don't link the installed version of forge on Linux/OSX
      Merge pull request #1 from umar456/JuliaComputing-kf/openblas64

Michael Carilli (1):
      Changed THREADS to 256 and BLOCKS to 64 in random.hpp.

Michael Nowotny (1):
      Added new example: heston_model in financial

Muneer Macbook (3):
      Added a mutex in the memory alloc and dealloc of the CPU backend to facilitate multi-threaded use
      Fixed memory leak in cpu/where.cpp
      Temporary workaround for incorrect weak copy symantics of af_array

Nathan Jackson (1):
      Fix race condition in reduce_first_kernel.

Pavan Yalamanchili (551):
      Merge pull request #106 from arrayfire/devel
      Merge pull request #148 from arrayfire/devel
      Merge pull request #181 from arrayfire/devel
      Merge pull request #259 from arrayfire/devel
      Merge pull request #319 from arrayfire/devel
      Merge pull request #372 from arrayfire/devel
      Merge pull request #381 from arrayfire/devel
      Merge pull request #464 from arrayfire/devel
      Merge pull request #512 from arrayfire/devel
      Merge pull request #517 from shehzan10/devel
      Merge branch 'general_index' of https://github.com/9prady9/arrayfire into general_index
      Merge pull request #521 from mcarilli/randufix
      EXAMPLE: Cleaning up machine learning examples
      EXAMPLE: Added softmax regression
      FEAT: Changing indexing to the generalized APIs
      TEST: Adding test for a(idx) = b where idx is array
      FEAT: array::isbool() is now implemented
      BUGFIX: Fixed bug in seq to array casting inside gfor
      STYLE: Cleaning up rainfall example
      BUGFIX: Logical operations now return b8 instead of u8
      BUGFIX: Making sure the array index is locked
      TEST: Test for logical assignment
      TEST: added for scoped out indexing with arrays
      Merge pull request #523 from 9prady9/batch4gfor
      Merge pull request #534 from glehmann/lib-version
      STYLE: Removing AF_VERSION_MINOR output in af_info()
      Merge branch 'cmake-freeimage' of https://github.com/glehmann/arrayfire into devel
      Merge pull request #535 from umar456/coverall
      Merge pull request #536 from shehzan10/devel
      FEAT: Adding combinations of arrays and sequences for indexing
      TEST: Additional tests for gen_index
      TEST: added tests for gen_assign
      FEAT: Adding functions to expose memory info and garbage collection
      Add the id of the device the memory is allocated on
      BUGFIX: Decrment the used buffer and byte count only once
      BUGFIX: Binary operations with scalars create proper types
      BUGFIX: Cast logical and bitwise operators to the right type
      TEST: get the proper output type for binary operations
      Merge pull request #540 from 9prady9/conv_batch4gfor
      Merge pull request #543 from umar456/rm_del
      BUGFIX: for accum along non-first dimension
      Retiring warps early for accum along first dimension
      Merge pull request #547 from umar456/cppcheck
      PERF: Minor improvements to accum in CUDA
      BUGFIX: in accum for OpenCL backend
      Merge pull request #551 from arrayfire/revert-550-devel
      Revert "Revert "download and build the opencl dependencies with cmake""
      Merge branch 'devel' of https://github.com/shehzan10/arrayfire into devel
      BUILD: Making sure boost compute path is included properly
      Merge pull request #555 from umar456/multigpublas
      Merge pull request #559 from 9prady9/fft_fixes
      BUGFIX: The 3rd and 4th dimensions offsets were flipped for identity
      TEST: enabling tests for 3D identity
      Merge branch 'umar456-identity' into devel
      BUGFIX: Maintaining the proper data type in binary functions
      Merge pull request #564 from pgovind/identity_fix
      Call garbageCollect() when cufft plan creation fails and try again
      Call garbageCollect() when CLFFT plan creation fails and try again
      Renaming cuda/fft.cu to cuda/fft.cpp
      Removing trailing whitespaces
      Merge pull request #569 from pentschev/err_cufft
      Merge pull request #570 from umar456/reduce_refactor
      Changes to fix warnings on gcc
      Style changes in array.cpp
      BUGFIX: Assignment inside GFOR with start and end points
      BUGFIX: batched mode assignment inside GFOR
      TEST: Adding tests for GFOR
      Style fixes in src/api/c/assign.cpp
      Merge branch 'arrayfire/devel' into lin_algebra
      Changes to remove compiler warnings
      TEST: Adding tests for solve
      Cleaning up error handling for cublas functions
      Merge pull request #575 from umar456/xcode
      Cleaning up cublas and cusolve managers
      Fixing memory leaks for linear algebra in cuda backend
      BUGFIX: QR decomposition for CUDA backend when M >= N
      FEAT: Adding solve for CUDA backend
      FFTWL libraries are not required
      Moving FindMKL to FindCBLAS.cmake
      Merge pull request #577 from umar456/rel_dir
      Merge pull request #578 from pentschev/err_clfft
      Helper functions in OpenCL to create af::array from cl_mem
      Removing unnecessary cl_device_id check
      Merge pull request #580 from umar456/assign64
      Adding support to retain and release cl mem objects
      Adding an optional retain parameter for getContext and getQueue
      Merge pull request #582 from munnybearz/threading
      Merge pull request #587 from pentschev/fftconvolve
      STYLE: Removing swtich case from FFT
      STYLE: Cleaning up the FFT code in CPU backend
      STYLE: cleaning up fft_common in CUDA backend
      Fixing a warning when compiling cufft
      STYLE: Fixing fft_common in OpenCL backend
      STYLE: cleanup computeDims and computePaddedDims in all backends
      Merge pull request #593 from shehzan10/join
      Merge pull request #602 from umar456/diagonal
      Merge pull request #599 from pentschev/fftconvolve_cuda_cache
      FEAT: Adding multi-dimensional batch support for fft
      TEST: Updating the fft tests with batch mode support
      Merge remote-tracking branch 'upstream/devel' into fft_fixes
      BUILD: Fixing uncaught merge conflicts
      BUGFIX: Median should now use floating point numbers for outputs
      BUGFIX: Median was using the improper count for number of elements
      TEST: Added tests to cover all cases of median
      BUGFIX: Changing the behavior of indexing to relfect old arrayfire
      BUGFIX: Assignment of scalars to values now uses to the storage type
      BUGFIX: Removing the evaluations before they are done
      BUGFIX: Fixed assign when both out and rhs are vectors
      TEST: Adding tests for assignment when special vector cases are involved
      TEST: Adding tests for indexing when inputs are vectors
      Moving dim4 to the common backend area
      BUGFIX: calculating ndims from dim4 properly when elements() == 0
      TEST: Re-writing the random test properly after earlier bugfix
      BUGFIX: Adding proper dim asserts for all data creation functions
      STYLE: Making sure af_destroy_array is called on non empty array
      Added proper error checking for special indexing cases
      STYLE: Moving dim asserts in data creation functions to a single place
      Copying FindLAPACK from cmake distribution to CMakeModules
      Merge pull request #612 from umar456/array_attrib
      Merge pull request #618 from pentschev/add_missing_opencl_checks
      Importing the first set of clmagma functions
      FEAT: LU decomposition for OpenCL backend
      Merge remote-tracking branch 'upstream/devel' into lin_algebra
      Merge pull request #619 from umar456/index_refactor
      BUGFIX: Making sure the lu_split kernel is working properly
      BUGFIX: Bugfixes in the magma code for getrf
      BUGFIX: input and output were flipped in magma/transpose.cpp
      BUGFIX: Fixing out of bound acceses in createPivot when M > N
      Merge pull request #620 from pentschev/remove_exec_permissions
      Merge remote-tracking branch 'upstream/devel' into lin_algebra
      Adding CLBLAS_CHECK to clblas.cpp and magma_blas.h
      Making sure initBlas is called before calling magma functions
      BUGFIX: Call the clblas routines only for non zero dimension lengths
      FEAT: real and imag now work with non complex numbers
      TEST: Adding extended tests for LU decomposition
      Merge pull request #623 from pentschev/fftconvolve_tests_large
      Adding magma files necessary for cholesky decomposition
      FEAT: Adding cholesky for OpenCL
      Removing unused variable from CPU and CUDA cholesky functions
      STYLE: Reduce redundant code in LU for all backends
      STYLE: Reduce redundant code in cholesky for all backends
      STYLE: Reuse qr_inplace inside qr for CPU backend
      Merge pull request #624 from pentschev/fftconvolve_opencl_cache
      Adding triangle matrix extraction to CUDA
      Adding triangle matrix extracting to CPU backend
      Adding triangle matrix extraction to OpenCL backend
      BUGFIX: Extracting lower triangle now works as expected in all backends
      Out of place cholesky now returns triangular matrix for all backends
      TEST: Adding tests for cholesky for large matrix sizes
      Merge remote-tracking branch 'upstream/devel' into lin_algebra
      Merge pull request #628 from umar456/cpp_idioms
      Merge pull request #629 from pentschev/fix_ml_examples
      Remove the need for volatile memory by always using __syncthreads();
      Adding shuffle instructions for __CUDA_ARCH__ > 300 in reduce
      Removing volatile memory and race conditions by adding ireduce
      Removing warnings from test/assign.cpp
      Merge pull request #635 from umar456/index_fix
      Merge pull request #639 from pentschev/fix_scan_osx
      Merge pull request #641 from pentschev/opencl_osx_fixes
      Adding magma files required for QR decomposition
      FEAT: Adding QR decomposition for OpenCL backend
      Updating err_clblas
      TEST: Adding tests for QR decomposition
      TEST: Fixing tests in LU decomposition
      FEAT: Solve for square systems
      Merge pull request #640 from umar456/proxy
      Merge pull request #647 from umar456/simplify_idx
      Merge pull request #648 from bkloppenborg/install_docs
      BUGFIX / STYLE: Cleaning up of indexing code to fix minor bugs
      STYLE: Minor style changes to indexing structs and classes
      Merge pull request #651 from umar456/lookup
      Merge pull request #652 from pentschev/increase_fft_coverage
      FEAT: Solve for non square systems in OpenCL
      Minor style changes
      FEAT: inverse for opencl backend
      STYLE changes in CPU and CUDA inverse functions
      STYLE minor changes to cuda/solve.cu
      Fixing a minor bug in swapdblk
      TEST: Updated solve tests
      BUGFIX: memory leak in opencl linear algebra routine
      TEST: Added tests for inverse
      Merge pull request #654 from umar456/idx_errchk
      FEATURE / STYLE: Changes to af::exception. Added af_err_to_string
      Moving non index functions from index.h to data.h
      af_print no longer uses af_reorder
      BUGFIX: Fixing a minor bug in CUDA solve
      TEST: Changes to dense linear algebra tests
      BUGFIX: Reduce the local memory useage for transpose_inplace in OpenCL
      Removing default option from C API
      Fixing error checks in cholesky and solve
      FEAT: Adding lower and upper for all backends
      BUGFIX: base_type of intl and uintl is now fixed
      TEST: tests for lower and upper triangle matrices
      Removing new and delete from cuda and opencl linear algebra functions
      *Inplace --> *InPlace
      Removing commented out test
      Random number generators now have unified states for all types
      Merge pull request #659 from bkloppenborg/usage_instructions
      FEAT: Adding setSeed and getSeed for all backends
      TEST: for setSeed and getSeed
      BUGFIX: af_print now prints 1D arrays properly
      Merge remote-tracking branch 'upstream/devel' into minor_fixes
      Merge pull request #660 from shehzan10/devel
      TEST: Updating random test to make sure there are no clashes
      Merge pull request #666 from FilipeMaia/fix_get_last_error
      Merge pull request #667 from shehzan10/devel
      Merge pull request #669 from FilipeMaia/fix_non_exported_symbols
      BUGFIX in random number generation for multiple GPUS in CUDA backend
      Exporting the symbols from dim4.hpp
      Instantiating array::unlock, array_proxy::unlock()
      Fixing error handling in seqToDims
      Changing af_blas_transpose to af_transpose_t
      Changing enums in af_pad_type
      FEAT: 2D spatial convolution now supported until 17x17
      Making the expand parameter an enum for convolutions
      Making all the inputs consts
      SOLVE, MATMUL and INVERSE now use af_mat_prop
      Adding argument checks for matmul, dot, inverse and solve
      FEAT: convolve automatically switches to frequency domain when necessary
      Fixing compile warnings in tests
      FEAT: Chaining matrix multiplications
      FEAT: IIR and FIR for all backends
      Revert "BUGFIX: af_print now prints 1D arrays properly"
      Revert "af_print no longer uses af_reorder"
      BUGFIXES: fixing batch mode in IIR filter for all backends
      Merge pull request #675 from FilipeMaia/where_64bit_int_support
      FEAT: Add complex support for fftconvolve
      TEST: Adding complex tests for fftconvolve
      TEST: Initial tests for IIR filter
      FEAT: Add short circuit code for iir when only a0 is available
      TEST: Adding tests for iir when only a0 is available
      BUGFIX in iir for all backends
      TEST: Adding more tests for iir filter
      Merge pull request #679 from arrayfire/var_tests
      Merge remote-tracking branch 'upstream/devel' into features
      Fix warnings in test/var.cpp
      BUGFIX: Fixed random number generation for CPU backend
      Merge pull request #683 from pentschev/add_log2_support
      STYLE: Changing dim_type to dim_t
      Merge pull request #686 from shehzan10/devel
      Merge pull request #689 from munnybearz/memleak
      API: dim_t is now a signed type
      BUILD: Making the default build type to be Release
      BUGFIX: IIR now uses lower local memory in opencl backend
      Updating licenses and copyright
      Updating the gtest submodule commit
      BUGFIX: fixed memleak in indexing
      Merge pull request #695 from shehzan10/devel
      STYLE: af_err enums now have hard coded values
      TEST: Check if output type is double in var tests
      FEAT,TEST: Adding 64 bit int support and tests for mean
      FEAT: Adding 64 bit support for stdev, corrcoef, covariance
      Merge pull request #699 from munnybearz/indexmemleak
      STYLE: changing AF_ERR_NOMEM to AF_ERR_NO_MEM
      STYLE: Fixing the way debug info is displayed when building opencl programs
      Added comments to differentiate af_err enum ranges
      Merge branch 'gs_tests' of https://github.com/umar456/arrayfire into devel
      STYLE: removing templates from gen_indexing
      Changing the geqrf to geqrf3 to reflect function name from magma
      Merge pull request #701 from bkloppenborg/example_docs
      FEAT/TEST: lower and upper support making the diagonal == 1
      FEAT: Adding QR in place for OpenCL backend
      Changing the API of linear algebra functions
      TEST: Adding tests for luInPlace, qrInPlace, choleskyInPlace
      Fixing a minor style issue for solve in CPU backend
      Merge branch 'devel' of https://github.com/shehzan10/arrayfire into lapack
      Fixing matrix_manipulation test
      First draft of lapack documentation
      DOC: Updated lapack documentation with code snippets
      DOC: cleaning up af/opencl.h documentation
      DOC: Adding documentation for missing parameters
      API: Removing domain parameter from separable convolution
      Merge branch 'devel' of https://github.com/shehzan10/arrayfire into lapack
      Merge pull request #706 from umar456/rowcol
      Merge pull request #705 from shehzan10/devel
      Removing redundant enum: af_source_t was just a clone of af_source
      Marking arrayfire constants to be externs
      Adding documentation for enums in af/defines
      Moving unsanitized C++ API to be inside __cplusplus checks
      API: changing (af_pad_type, padType) --> (af_border_type, borderType)
      Fixing compile warnings
      DOC: Adding documentation for IIR and FIR
      Merge pull request #710 from pentschev/hamming
      Moving af_features to be an internal structure
      BUILD fix for test/hamming.cpp
      Functions that don't support GFOR / batch mode now return errors
      Changing the messages af::exception generates
      BUGFIX: Fix for assign with linear indexing
      TEST: Adding tests for linear assign and index
      BUGFIX: Fix GFOR support for convolve in CPU backend
      BUGFIX: Convolve in OpenCL now work when filters are sub arrays
      TEST: Adding GFOR tests for convolve
      BUGFIX/TEST: medfilt in CPU backend now works inside GFOR
      BUGFIX/TEST: morph fixed for GFOR in CPU backend.
      BUGFIX/TEST: Fixing bug inside GFOR for bilateral in CPU backend
      BUGFIX/TEST: histogram inside GFOR fixed for CPU backend
      BUGFIX/TEST: fixed meanshift in GFOR for CPU backend
      BUGFIX/TEST: Fixed lower and upper in OpenCL backend for sub arrays
      TEST: GFOR tests for resize, diagonal and transpose
      Merge pull request #718 from umar456/api_update
      Inplace LU now returns pivot in lapack compliant format
      Fixing warning messages during compilation
      FEAT: solveLU has been added to all backends
      TEST: Adding tests for solveLU
      DOCS: Adding documentation details for solveLU
      Fixing the cuda error check in cuda/platform.cpp
      Minor fixes to solveLU documentation
      Adding getrs to backend/lapacke.cpp
      Merge pull request #725 from shehzan10/devel
      Merge pull request #724 from pentschev/hamming_fixes
      Merge pull request #723 from 9prady9/fixes
      Fixes for building on OSX
      Minor style changes
      FEAT: Adding support for triangle matrix solve
      TEST: Adding tests for solve
      TEST: Updating documentation for af::solve
      Merge remote-tracking branch 'upstream/devel' into solvers
      Removing unnecessary const from af_lu_inplace
      BUILD fixes for OSX
      BUGFIX: solve with upper triangular matrices fixed for NVIDIA gpus
      Merge pull request #732 from umar456/array_docs
      Merge pull request #734 from unbornchikken/pr1
      BUGFIX: anytrue / alltrue now return b8 instead of u8
      FEAT: Adding det and rank for all backends
      Updating documentation for rank and det
      Merge pull request #722 from ghisvail/enh/cmake-install-paths
      FEAT: Adding support for matrix norm in all backends
      FEAT: Add missing element wise functions
      Adding array.nonzeros()
      FEAT: Adding minfilt and maxfilt
      Adding "tests" for missing functions
      Fixing up documentation for image processing and lapack
      Merge pull request #737 from umar456/style_docs
      Change uint to unsigned for fixing builds on windows
      Merge pull request #741 from bkloppenborg/improve_docs
      array.device<T>() now hands complete control of the memory to user
      BUGFIX: cpu memory manager incorrectly calculating total_bytes
      FEAT: Adding functions to get and set memory chunk resolution
      TEST: Adding tests for memory manager
      Fixing code for af/opencl.h
      TEST: Updating memory test for indexing and assignment
      BUILD fix for OSX.
      Fixing fractal example
      BUGFIX: Fixing the reset methods in backend/CPU/TNJ/*
      Removing exceptions being thrown from clean up code
      Adding a specialziation for array.device<cl_mem>() in af/opencl.h
      BUGFIX: element wise operations on sub arrays work now in CPU backend
      Merge pull request #748 from bkloppenborg/example_install_path
      Merge pull request #747 from shehzan10/devel
      Merge pull request #749 from 9prady9/ImageEditingExample
      Moving header files to fix build issues
      Merge pull request #751 from shehzan10/devel
      Merge pull request #752 from pentschev/fftconvolve_warnings
      Fixing issues with doxygen
      Changes to AF_VERSION*
      Fixing documentation to have pre-requisites
      Fixing path of the installed examples directory
      Removing LINK_INTERFACE_LIBRARIES and INTERFACE_LINK_LIBRARIES
      Merge pull request #756 from pentschev/doc_typos
      Merge pull request #754 from shehzan10/devel
      Merge pull request #755 from umar456/constant
      Merge pull request #753 from umar456/proxy_fix
      Fixing build failures for OpenCL on a few machines
      Merge branch 'dep_func' of https://github.com/umar456/arrayfire into devel
      Commenting out broken test
      BUILD: Adding INSTALLER_MODE flags for all backends
      Fixing documentation and release notes
      Fixing CPack.txt
      Fixing CPack.txt
      Merge pull request #773 from shehzan10/devel
      Merge pull request #770 from umar456/basic_c
      Merge pull request #774 from bkloppenborg/cmake_fix
      Merge pull request #777 from pentschev/vision_docs
      Merge pull request #778 from pentschev/matchtemplate_vision
      Merge branch 'fftconvolve_fix' of https://github.com/pentschev/arrayfire into hotfixes-3.0.1
      BUGFIX/TEST: fftConvolve now does multi dimensional batching properly
      BUGFIX/TEST: fftConvolve now does multi dimensional batching properly
      Merge pull request #789 from FilipeMaia/array_proxy_constructor
      Merge pull request #808 from shehzan10/hotfixes-3.0.1
      Merge pull request #807 from pentschev/ml_examples_fix
      Adding using_on_osx document
      Update release notes and version details
      Merge pull request #814 from arrayfire/hotfixes-3.0.1
      Merge pull request #817 from FilipeMaia/complex_dot_product
      Merge pull request #819 from FilipeMaia/missing_header
      Merge pull request #820 from shehzan10/CUDA-fixes
      Merge pull request #831 from ghisvail/hotfix/unnecessary-include
      Merge pull request #835 from 9prady9/dog
      Merge pull request #841 from shehzan10/gradient_fix
      Merge branch 'unwrap' of shehzan10/arrayfire into devel
      Merge pull request #842 from umar456/pod
      Merge pull request #843 from umar456/offset_check
      FEAT: Implementing array::lock() and array::unlock()
      DOCS: Adding documentation for constant
      Adding documentation for array::scalar
      DOCS: Updating arith functions to specify their input limitations
      FEAT / TEST: Adding af::copy()
      Display if a backend is enabled or not when building examples and tests
      DOCS: Fixing documentation for lock and unlock
      Merge pull request #852 from shehzan10/resize_lower
      Merge pull request #851 from shehzan10/linear_fix
      Merge pull request #849 from bkloppenborg/citation
      Merge pull request #855 from umar456/long_long
      Merge pull request #853 from shehzan10/nearestNeighbour
      DOCS: Updating release notes
      Merge pull request #857 from arrayfire/hotfixes-3.0.2
      Merge remote-tracking branch 'upstream/master' into devel
      Updating version number
      Merge branch 'master' into devel
      BUFIX: Remove memory leak in af::copy()
      Updating forge tag to fix build issuess with ninja
      FEAT: Adding function to get use_count of shared pointers
      PERF: Do not make copies if the number of references is only 1
      Renaming enums for convolve batch modes
      BUGFIX: Ignoring NaN values in min and max for all backends
      Merge pull request #866 from JuliaComputing/kf/forgerpath
      Merge pull request #867 from JuliaComputing/kf/blasrename
      Merge pull request #873 from shehzan10/more-lower
      Merge pull request #865 from JuliaComputing/kf/openblas64
      FEAT: Added support to substitute nan values for sum and product
      Adding missing instantiations for compat functions
      TEST: Adding tests for reductions when using NaNs
      BUGFIX: Fixing casting to and from complex numbers in CPU backend
      Adding more operator overloading for af::cfloat and af::cdouble
      BUGFIX: Making sure c32/c64 imitate f32/f64 when operating with scalars
      TEST: Add mixed type tests with complex inputs
      PERF: Improvements for non linear JIT kernels in OpenCL backend
      TEST: Adding batched mode tests
      PERF: Speeding up JIT for 3D arrays in OpenCL backends
      BUGFIX: median of all elements is now fixed
      PERF: Improvements to tile when tiling along singleton dimensions
      PERF: Improvements to CUDA JIT for non linear 3D and 4D arrays
      FEAT: Adding support for non overlapping batched convolution
      TEST: Adding tests for non overlapping convolves
      DOCS: Updating the documentation for convolution
      Adding missing license for a few files
      Updating version to 3.1.0
      Merge pull request #891 from 9prady9/assets_changes
      Merge pull request #892 from 9prady9/cmake_fixes
      Merge pull request #893 from shehzan10/devel
      BUGFIX: Check for NULL values when allocating memory on CPU backend
      Ensure CUDA and OpenCL return proper errors when out of memory
      TEST: Adding test to trash the memory manager and see if it recovers
      FEAT,TEST: Adding sigmoid function for all backends
      Adding the option to remove tests from ctest
      BUGFIX/TEST: Fixing bug in rank. Added appropriate tests
      BUGFIX/TEST: Fixing not for C API. Added relevant tests.
      BUGFIX: Fixing a bug in randn for CPU backend
      BUGFIX: Fixing setSeed for randu
      TEST: Updating and fixing the randu/randn tests
      Merge pull request #882 from klemmster/cusolver_svd
      TEST: Updating random tests to properly reset seeds
      TEST: Fixing out of bounds access in fft tests
      BUGFIX in randn for apple systems
      Renaming rank test to rank_dense
      Merge pull request #910 from shehzan10/devel
      Merge pull request #909 from bkloppenborg/autobuild-backends
      BUGFIX: Fixed issues with mixed real and complex types
      Fixing the checks for skew
      BUGFIX: conjg no longer errors out for real inputs
      af_scale now checks for default parameters properly
      Extended support for interleaved convolution
      Updating the COPYRIGHT.md document
      Updating the language in COPYRIGHT document
      Merge pull request #922 from 9prady9/match_template_example
      Merge pull request #921 from 9prady9/susan
      Moving af_array info gathering functions from util.h to array.h
      Making fft_inplace consistent across all backends
      FEAT / TEST: Adding support for inplace fft
      Moving general fft implementation to src/api/c
      Merge pull request #931 from 9prady9/sat
      Merge pull request #929 from 9prady9/cuda_default_stream
      Merge pull request #928 from pavanky/minor_changes
      Removing consts from the fftInplace API
      FEAT/TEST: Adding R2C and C2R FFT transforms for all backends
      DOCS: Adding documentation for real to complex transforms
      Fixing a minor issue in ArrayFireConfigVersion.cmake file
      Fixing issue in documentation
      Cleaning up multiply_inplace in cpu backend
      Enabling memory manager back in cuda backend
      Merge pull request #937 from 9prady9/wind_resize
      FEAT: Adding select for CPU backend
      FEAT: Select added for CUDA backend
      FEAT: Select for opencl backend
      FEAT: replace for all backends
      TEST: Adding tests for select and replace
      FEAT: adding complex support for exp
      Binary operations with floating point scalars default to single precision
      BUGFIX: Fixing offset issue with CPU element wise operations
      PERF: improvements to element wise operations in CPU backend
      Merge pull request #942 from 9prady9/ycbcr
      Merge branch 'devel' into svd
      Changes to style and fix compile errors
      Merge pull request #943 from shehzan10/stream
      Style changes to code in unwrap
      FEAT: Adding support to unwrap along rows as well as columns
      BUGFIX: Fixed a bug for unwrap in all backends
      FEAT/TEST/DOC: Adding wrap for CPU backend
      FEAT: wrap for CUDA backend
      Adding atomics.hpp file for CUDA that can be used in the future
      Moving the kernel cache map to a centralized location
      FEAT: wrap for OpenCL backend
      Removing faulty test
      Cleaning up unwrap code in OpenCL by using cache store
      Merge pull request #952 from 9prady9/cpuinfo
      Fixing the compile error on windows
      Cleaning up cpu blas / lapack in OpenCL backend
      Fixes to supress annoying compiler warnings in OpenCL backend
      Adding functions from clMagma necessary for OpenCL SVD:
      Initial support for SVD in OpenCL backend
      Adding proper error checking in magma
      BUGFIX: in array_proxy::get() const
      Merge pull request #961 from 9prady9/cl_hpp_fixes
      Merge pull request #953 from umar456/fix_951
      Fixing svd params to reflect clmagma
      Merge pull request #962 from shehzan10/update-deps
      Fixing the output of af::info() for OpenCL backend
      Updating documentation and adding version guards for for 3.1
      Merge pull request #948 from pentschev/sift
      Reorganizing non free build process.
      Changing build flags to build non free algorithms
      Templated options are now runtime compile options for opencl reductions
      Templated options are now runtime compile options for opencl convolutions
      Templated options are now runtime compile options for opencl indexed min/max
      Templated options are now runtime compile options for opencl scan
      Templated options are now runtime compile options for opencl nearest neighbor
      Removing unnecessary switch case from opencl ireduce
      Templated options are now runtime compile options for opencl FAST
      Splitting up opencl sort_by_key files to compile in parallel
      Splitting sort_by_key across too many files slows down compile times
      Fixing a bug introduced a couple of commits ago in OpenCL SIFT
      Merge branch 'devel' into nonfree_fixes
      Merge pull request #968 from shehzan10/opencl_fixes
      Merge remote-tracking branch 'upstream/devel' into svd
      Work around for issues in OpenCL svd
      API clean up and adding support for complex numbers for SVD
      Fixing various typos and bug fixes for SVD in CUDA and OpenCL
      TEST: for SVD
      DOCS: Updating the documentation for SVD
      Adding version guards for svd
      Adding more pragma directives to supress GCC warnings
      TEST: updating SVD tests to contain all four floating point data types
      Fixing svd example to reflect the change in API
      Revert "Updated boost compute version tags"
      Compilation fixes for OSX
      Use xGESVD instead of xGESDD for ARM platforms
      Merge pull request #970 from shehzan10/rel_31
      FEAT: Adding support for linear assignment in C API
      Restore original shape after flattening input for linear indexing
      FEAT: Adding support for linear indexing in C API
      Merge pull request #974 from arrayfire/devel

Peter Andreas Entschev (144):
      Added Hamming Distance API and CUDA backend
      Changed regions to accept b8 input only
      Fixed regions unit tests to use b8 as input
      Updated regions docs
      Merge pull request #563 from pavanky/binary_fixes
      Added CUFFT_CHECK() to check for cuFFT errors
      Changed CUDA FFT functions to use CUFFT_CHECK()
      Added fftconvolve() C API
      Added fftconvolve() C++ API
      Added fftconvolve() prototypes to signal.h
      Added documentation for fftconvolve()
      Added template parameter to fftconvolve() C API
      Added expand flag to fftconvolve()
      Added CLFFT_CHECK() to check for clFFT errors
      Merge branch 'devel' into fftconvolve
      Added fftconvolve() support for CUDA
      Added CUDA version check to err_cufft.hpp
      Added fftconvolve() support for OpenCL
      Added unit tests for fftconvolve()
      Added fftconvolve() support for CPU
      Merge remote-tracking branch 'upstream/devel' into fftconvolve
      Remove usage of max() in fftconvolve()
      Fixed bugs on fftconvolve()
      Fixed fftconvolve() bug for large input sizes
      Added fftconvolve() interface to fft() plan cache
      Updated OpenCL and CPU fftconvolve() templates
      Merge remote-tracking branch 'upstream/devel' into devel
      Removed unused template parameters from CUDA fftconvolve()
      Added u64 support to join()
      Merge remote-tracking branch 'upstream/devel' into add_join_u64
      Added s64 support to join()
      Moved large FFT unit tests to fft_large.cpp
      Added missing checks to OpenCL functions
      Removed execution permissions of source files
      Fixed size calculation of packed array in fftconvolve()
      Added large unit tests for fftconvolve()
      Fixed wrong fftconvolve() large unit tests
      Exposed fft_common() in OpenCL backend
      Integrated OpenCL fftconvolve() with fft() plan cacher
      Fixed wrong labels printing in ML examples
      Merge pull request #630 from 9prady9/win_fixes
      Fixed FAST on Mac OS X
      Merge pull request #631 from umar456/assign_tests
      Fixed build warnings
      Merge pull request #637 from shehzan10/compute_update
      Fixed scan for OS X.
      Added missing calls to OPENCL_DEBUG_FINISH()
      Added missing OpenCL exception handling
      Fixed assign()/index() on OSX
      Added missing POST_LAUNCH_CHECK() calls on CUDA
      Moved conv_image() to testHelpers.hpp
      Fixed meanshift() unit tests for double data type
      Merge pull request #643 from shehzan10/af_write
      Increase resize() test coverage
      Removed unecessary checks in resize()
      Added new test helper for output only functions
      Added tests for gaussiankernel()
      Removed test helper for output only functions
      Changed way how gaussiankernel() size is handle in tests
      Updated test data
      Increased fft() test coverage
      Fixed fft() tests to properly call dft() wrapper
      Fixed af_mul_t OpenCL operation
      Added log2() support
      Added missing copyright headers
      Merge pull request #685 from ghisvail/typo/copyrights-af-url
      Merge remote-tracking branch 'origin/hamming_distance' into devel
      Changed dim_type to dim_t in hamming_matcher
      Added CPU backend for hamming_matcher()
      Added OpenCL backend for hamming_matcher()
      Make proper use of local memory for OpenCL hamming_matcher()
      Fixed race condition in CUDA hamming_matcher()
      Reusing test condition in OpenCL hamming_matcher() kernel
      Added missing syncs to hamming_matcher()
      Added unit tests for hamming_matcher()
      Updated test data
      Added Hamming matcher documentation
      Moved computer vision functions to vision.h
      Prevent CUDA hamming_matcher from allocating additional device memory
      Added more syncs to hamming_matcher()
      Changed C++ hamming_matcher() to hammingMatcher()
      Added C++ unit test for hammingMatcher()
      Prevent OpenCL hamming_matcher from allocating additional device memory
      Cleaned up some hamming_matcher code
      Fixed memory leak in unit test helper's conv_image()
      Fixed OpenCL hamming_matcher() local memory size query
      Fixed hamming_matcher() OpenCL kernel for Intel devices
      Changed CUDA hamming_matcher switch case to if condition
      Added FAST feature detector example
      Changed exception handling in examples to stderr.
      Fixed CUDA fftconvolve warning
      Added missing return statement to OpenCL getQueue()
      Fixed several docs typos and confusing sentences
      Moved CV documentation to vision.dox
      Added missing Hamming matcher documentation
      Moved matchTemplate definition to vision.h
      Merge remote-tracking branch 'upstream/hotfixes-3.0.1' into vision_docs
      Fixed matchTemplate includes
      Fixed fftconvolve() bug, resulting in wrong output
      Fixed wrong indexing in some machine learning examples
      Added C API for Harris corner detector
      Added C++ API for Harris corner detector
      Added CPU backend for Harris corner detector
      Added CUDA backend for Harris corner detector
      Added OpenCL backend for Harris corner detector
      Added Harris corner detector unit tests
      Added documentation and function definitions for Harris
      Fixed FAST unit tests
      Updated data
      Removed unused variables from OpenCL's sort_index
      Merge remote-tracking branch 'upstream/devel' into harris
      Fixed bug affecting Harris on AMD GPUs
      Added SIFT prototypes and parameter documentation
      Added C API for SIFT
      Added C++ API for SIFT
      Added CUDA backend for SIFT
      Added OpenCL backend for SIFT
      Added CPU backend for SIFT
      Added AF_ERR_NONFREE to defines.h
      Added BUILD_NONFREE option to CMake
      Added SIFT Copyright information
      Added OpenSIFT License
      Added SIFT documentation
      Updated test data
      Added SIFT unit tests
      Made SIFT image indexing more readable in CPU backend
      SIFT fix for CUDA on Windows, made it more readable
      Made SIFT image indexing more readable in OpenCL backend
      Templated SIFT gaussianElimination() in CPU and CUDA backends
      Added missing CUDA_LAUNCH and THRUST_SELECT to SIFT
      Improved CUDA SIFT coalescing and performance
      Improved OpenCL SIFT coalescing and performance
      Passing shared size memory dynamically to CUDA SIFT
      Moved OpenCL's conv2Helper to kernel directory
      Improved SIFT OpenCL code
      Using pre-defined constants for workgroup sizes in CUDA SIFT
      Using 3D arrays for Gaussian/DoG pyramids in CUDA SIFT
      Using cudaMemsetAsync for SIFT
      Fixed OpenCL SIFT bug causing segmentation faults on Intel
      Added missing buffer freeing call to OpenCL SIFT
      Fixed CUDA SIFT on unused memory buffer
      Changed SIFT unit test to use std::stable_sort()
      Moved syncthreads/barriers out of thread conditionals
      Moving CUDA SIFT syncthreads calls out of thread conditionals

Pradeep (45):
      FEATURE: generalized indexing function
      Renamed af_assign to af_assign_seq
      FEATURE: af_assign_gen
      af_assign_gen cuda backend implementation
      af_assign_gen opencl backend implementation
      Merge branch 'devel' into general_index
      Added 4th dimension batch for erode,dilate
      Added 4th dimension batch for erode3d,dilate3d
      Revised documentation for convolve
      Added 4th dimension batch for bilateral
      Added 4th dimension batch for meanshift
      Added 4th dimension batch for histogram
      Added 4th dimension batch for median filter
      matchTemplate 4th dimension batch support
      4th dimension batch support for rgb-gray transformations
      Enabled 4th dim batch support for hsv-rgb conversions
      Added 4th dim batch support for 2d separable convolve
      Added 4-connectivity code sample
      batch support for indexed arrays in separable 2d convolution
      BUGFIX: batch support for indexed arrays in morphological functions
      batch support for indexed arrays in bilateral, meanshift
      batch support for indexed arrays in histogram
      batch support for indexed arrays in medfilt function
      batch support for indexed arrays in matchTemplate
      batch suppor for indexed arrays in sobel functions
      Moved cuda::trimIndex to utility header
      BUGFIX: fixed a condition check in cpp wrapper for convolve
      Renamed ConvolveBatchKind::[ONE2ALL to ONE2MANY]
      BUGFIX: added .as(u8) call for input in regions
      Multidimensional batch support for convolve
      Changed the shared memory loading pattern in 3d convolve
      HOTFIX: fixes normalization factor bug in ifft
      additional cpp convenience wrappers for fft and ifft
      renamed unified fft wrapper API
      BUGFIX corrected size_t format specifier
      BUGFIX fixed 3rd dimension index in cpu histogram implementation
      BUGFIX fixed Windows Debug mode issue
      Windows specific fixes for graphics
      Changes to reflect forge API changes
      Removed unncessary CheckGL calls
      OSX graphics fixes
      Replaced static arrays of cl::program/kernels with maps
      OS X interop context creation changes
      Corrected no graphics enum in graphics functions
      Additional Image Processing Examples

Pradeep Garigipati (22):
      Merge pull request #160 from arrayfire/devel
      Merge pull request #223 from arrayfire/devel
      Merge pull request #515 from bkloppenborg/cmake_opencl_fix
      Merge pull request #519 from shehzan10/devel
      Merge pull request #537 from pavanky/indexing
      Merge pull request #550 from shehzan10/devel
      Merge pull request #560 from pentschev/regions_b8
      Merge branch 'plot' into 'graphics'
      Merge branch 'border' into 'graphics'
      Merge pull request #603 from pavanky/fft_fixes
      Merge branch 'api_changes' into 'graphics'
      Merge branch 'cpu_hist' into 'graphics'
      Merge pull request #634 from pavanky/bug_fixes
      Merge pull request #645 from pentschev/increase_resize_coverage
      Merge pull request #649 from pavanky/index_cleanup
      Merge pull request #671 from pavanky/fixes
      Merge pull request #863 from pavanky/bugfix
      Merge pull request #906 from pavanky/bugfixes
      Merge pull request #936 from pavanky/fft
      Merge pull request #941 from pavanky/new_funcs_31
      Merge pull request #945 from pavanky/jit_fixes
      Merge pull request #944 from pavanky/minor

Prashanth Govindarajan (4):
      Build fix for Windows
      FEAT Added plot to cpu backend
      AF changes for forge borders and ticks
      AF CPU backend changes for forge 1dhistogram

Richard Klemm (5):
      Add SVD API
      Add SVD Cuda backend
      Add SVD CPU Backend
      Add SVD OpenCL Stump
      Add SVD example

Shehzan Mohammed (327):
      Merge pull request #130 from arrayfire/devel
      Added image graphics to CPU and CUDA backend
      Merge branch 'devel' into graphics
      Fix for no graphics builds
      Move graphics include files into GRAPHICS_FOUND in CMake
      Merge branch 'devel' into graphics
      Added path for GLEWmx for Tegra
      Added OpenGL error checks, better window close handling
      Added Conway's Game of Life example
      Added OpenGL errors, better window closing to CPU
      Added AF_ERR_GL_ERROR and error checking in graphics
      Merge branch 'devel' into graphics
      Merge branch 'devel' into graphics
      Merge branch 'devel' into graphics
      Update image API to accept scaling factors
      Merge branch 'devel' into graphics
      Merge branch 'devel' into graphics
      Merge pull request #249 from arrayfire/devel
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Added support for AFGFX
      Merge remote-tracking branch 'origin/devel' into graphics
      Using colorspace conv in graphics. Fixed memory leaks
      Mirroring change in handle names from AFGFX
      Merge remote-tracking branch 'origin/devel' into graphics
      Fixed memory leak in image c-api
      Merge remote-tracking branch 'upstream/devel' into graphics
      Merge remote-tracking branch 'upstream/devel' into graphics
      Using direct backend calls for join, tile reorder in af_image
      Merge remote-tracking branch 'origin/devel' into graphics
      Fix opencl build errors:
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'upstream/devel' into graphics
      Merge remote-tracking branch 'upstream/devel' into graphics
      Added pretty version of conway
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'upstream/graphics' into graphics
      Better templating convert_and_copy_image in graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'upstream/graphics' into graphics
      Changed graphics names to match Forge nomenclature
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'upstream/graphics' into graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'upstream/graphics' into graphics
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge pull request #514 from bkloppenborg/cmake_packaging
      Merge remote-tracking branch 'origin/devel' into graphics
      Merge remote-tracking branch 'upstream/graphics' into graphics
      BUGFIX Fixed bug in range
      Added test for range
      API Add overloaded api for tile
      BUGFIX Fixed tiling bug in iota
      Added test for iota
      Updated README.md with binary downloads link
      Using FindOpenCL from CMake 3.2
      Merge branch 'build-opencl-dependencies' of https://github.com/glehmann/arrayfire into devel
      Update clFFT commit id
      Merge pull request #545 from 9prady9/fft_api_additions
      Revert "download and build the opencl dependencies with cmake"
      Using commits as version for Boost.Compute external build
      Fixed building clBLAS external
      Fixed building clFFT external
      Using wget if cmake downloads empty file for boost compute
      Formatting and case in cmake files
      Merge remote-tracking branch 'origin/devel' into graphics
      Change download link
      Merge pull request #553 from glehmann/fix-ep-byproducts
      Merge remote-tracking branch 'upstream/graphics' into graphics
      Renaming lib_deps to FreeImage_LIBS
      Added Linear Algebra for CUDA using cusolver
      Fix compiler errors for lapack
      Added Cholesky for CPU
      Added LU to CPU
      Fixed dimensions in QR CUDA
      Added QR to CPU
      Added brief linear algebra example files
      Fixed cholesky upper-lower issue
      Fixed return arrays for lapack functions
      Returning index value array for LU as pivot
      Added CPP tests for linear algebra
      Added inplace transpose API
      Merge branch 'name_changes' into 'graphics'
      Added framework for solve
      Added CPU implementation for solve
      Merge pull request #571 from pavanky/gfor_fix
      Merge branch 'lin_algebra' of github.com:alltheflops/arrayfire into lin_algebra
      Added TRSM to cuda backend
      Added convert pivot option to cuda lu
      Solve CUDA and CPU fixes. Work in progress
      Compilation fixes for trsm
      Change lapack api leading dims to use strides
      Added convert_pivot option for LU to all backends
      CMake fixes for CUDA 7 with CMake 3.2. Change WITH_LINEAR_ALGEBRA to WITH_<backend>_LINEAR_ALGEBRA
      Remove AFAPI from cpp/lapack.cpp
      Added inverse to all backends
      Merge branch 'graphics' into plot
      Split plot from image
      Merge remote-tracking branch 'mule/opencl_image' into plot
      Merge remote-tracking branch 'upstream/devel' into plot
      Move CUDA InteropManager into its own file
      Move OpenCL InteropManager into its own file
      Added plot to CUDA backend
      Updated plot example
      Added plot to OpenCL backend
      Removed size from fg_plot2d in accordance with change in forge
      FEAT Added 3 and 4 array join functions
      Changed af_join3/4 into af_join_many
      Added CPP test for join many
      Fixes for number of arrays in join
      Merge pull request #609 from pentschev/add_join_u64
      Merge pull request #617 from pentschev/move_fft_large_tests
      Update boost.compute repository and commit
      Merge pull request #636 from pentschev/fix_build_warnings
      Added write and af_write_array functions to all backends
      Added tests for write
      Merge pull request #642 from pentschev/fix_meanshift_tests
      Merge remote-tracking branch 'upstream/devel' into lin_algebra
      Fix for min max in magma files
      Fixes for qr and cholesky tests
      Changes to accomodate f77 blas on windows for lapack
      Adding AF_CPU definition to CPU backend
      Separate variable for f77 blas for CPU and OpenCL
      Fixes for CUDA linear algebra on windows
      Add AFAPI to proxy class for windows
      Remove setenv from inverse test
      Using MKL CBLAS, LAPACKE
      Increase tolerance for LU float test
      Added provisional lapacke wrapper for lapack
      Fixes for lapacke wrapper on OSX
      Fixes for lapack tests
      Added return in swapblk opencl, fixed size of qr test
      Fix CUDA cusolver library find in CMake
      Fix compiler warnings in tests
      Disable lapacke.cpp for OSX CUDA
      Added class for FreeImage such that init and deinit are called once
      Reduced size for triangle tests - Used > 1.5GB mem
      Added options for static FreeImage in CMAKE
      Calling freeimage init/deinit in constructor/destructor
      Updated FindFreeImage.cmake to handle switch of lib type
      Added COPYRIGHT.md for software credits and corresponding licenses
      Fix freeimage_include_path
      CMake Fix for broken build due to freeimage static
      Fixed Freeimage initialization bug
      Changed FI, MKL toggle variables from *_USE_* to USE_*
      Using simpler FindLAPACKE script
      Add double test disabling to random getseed
      Type corrections in COPYRIGHT.md
      Added default arguments to translate, scale, skew
      Fixed linear interpolation for transform/rotate
      Merge remote-tracking branch 'upstream/devel' into devel
      Merge branch 'devel' of github.com:arrayfire/arrayfire into devel
      Build fixed for windows for b685fa6
      Merge remote-tracking branch 'upstream/devel' into graphics
      Fix Forge CMake scripts
      Warning fixes in hist and morphing
      Building forge as external project
      Build fixes for rotate and abs
      abs fix for size_t on tegra
      Merge remote-tracking branch 'upstream/devel' into graphics
      Fixed to build_forge for building on windows
      Installs forge includes/libs with install command
      Divide VS projects into VS Filters for Tests and Examples
      Update build_forge.cmake
      Add Find Package GLFW back for include directories
      Change translate test verification to ratio
      Fixes for Solve OpenCL for NVIDIA GPUs
      Disable solve/qr test failing on windows opencl
      Fixed qr and solve test to compile
      Merge remote-tracking branch 'upstream/devel' into graphics
      Fix build_forge to build properly on windows
      Adding BUILD_GRAPHICS Option to find/build Forge
      CMakeList Tabbing/Formatting
      Added GLEW and GLFW along with licenses to copyright
      Changing forge-external to forge-ext
      Change macro for dim_t definition
      Expanded macro for definition of dim_t
      Adding LP64 macro for dim_t
      Merge remote-tracking branch 'upstream/devel' into graphics
      Compilation fixes and getting started test fixes
      Documentation for data.h
      Documentation for util.h
      Documentation for print
      Documentation for inplace transpose
      Fix double tests in getting started and matrix manip
      Added c32/64 s/u/32/64 to resize
      DOC: Added documentation for image
      Compilation fixes in resize for osx and windows
      CPack filename based on graphics support
      Compilation fix for afHost in conway examples
      Change conway loop condition to window.close()
      Fixing double tests in getting started
      Making the triangle test slightly smaller in size
      Removing forge headers from install command
      Merge pull request #711 from 9prady9/no_gfx_fix
      Option to use static GLEWmx
      Remove print statements
      Add option to use system GTest
      Fixing GLEWmx for build forge
      Fix __builtin_popcount for windows
      Fix e5d2cf9
      Add AF_DISABLE_GRAPHICS to disable OpenCL window creation on init
      Correct alternate functions for deprecated functions
      Replace deprecated calls with new APIs
      Fix reduce test files
      Added more paths to FindGLFW
      Add descriptions to Conway, update conway_pretty
      Merge pull request #726 from 9prady9/examples
      Added wrapper for setTitle for window
      Added arrayfire pro tips page
      Renamed arrayfire_pro_tips to configuring_arrayfire_environment
      Add option for user to set relative test data directory
      More uint to unsigned
      Merge pull request #739 from 9prady9/cmap_additions
      Merge pull request #742 from bkloppenborg/examples_fix
      Merge pull request #745 from pavanky/fractal_fixes
      Adding capability to build multiple compute versions
      Change CUDA_COMPUTE_CAPABILITY to CUDA_DETECTED_COMPUTE
      Added CUDA_COMPUTE_DETECT Option to disable auto detection
      For single compute, set PTX to compute version
      Replace deprecated functions with new API
      Deprecated functions will throw errors in examples
      Disable cuda compute check from being run everytime cmake is run
      Fix deprecated warning to error for visual studio
      Fix headers for c-only compilation
      Merge pull request #757 from pavanky/final_fixes
      Tagging build_forge for af3.0
      Added Step 0 to windows to run pre-built executables
      Merge pull request #759 from arrayfire/devel
      Fixing conflicts between std::array and af::array
      Move snprintf/static macros from defines.h to backend
      Disable RPM and DEB packages
      Merge pull request #776 from 9prady9/gfx_changes
      Change forge branch to master
      Add defines.hpp to opencl files where snprintf is required on windows
      Enable multiple CUDA computes to be detected and enabled
      Disable fallback computes if any CUDA_COMPUTE_XY is set
      Change build status to reflect master status
      Disable fallback only if any compute is set to ON
      Merge branch 'fix-missing-include' of git://github.com/glehmann/arrayfire into multiple_computes
      Merge pull request #796 from umar456/osx_install
      Merge pull request #798 from umar456/fractal
      Optimized compute detection code
      Merge pull request #803 from bkloppenborg/hotfixes
      Merge remote-tracking branch 'upstream/hotfixes-3.0.1' into devel
      Merge branch 'private_lib' of git://github.com/umar456/arrayfire into hotfixes-3.0.1
      Fixed examples cmake for unix
      Merge remote-tracking branch 'upstream/devel' into hotfixes-3.0.1
      Fix header file in basic_c
      Merge pull request #805 from 9prady9/gfx_image_fixes
      Add gitter label
      Fix CUDACheckCompute when error is returned
      Minor typo fixes in windows doc
      Fixed gtest git link
      Merge pull request #811 from 9prady9/printf_cleanup
      Merge pull request #810 from umar456/osx_install
      Merge pull request #815 from arrayfire/hotfixes-3.0.1
      Remove cache path from NVVM path
      Function return type fix for blas
      Compilation fix for windows when graphics disabled
      FEAT Add CPU backend for unwrap function
      Added CUDA backend for Unwrap
      Added OpenCL backend for Unwrap
      Bugfixes, comments
      Adding padding for strides > 1
      Added test for unwrap, updated data submodule
      Added documentation for unwrap
      Changing behavior of unwrap using padding
      Changed unwrap tests to new behavior
      Updated documentation for unwrap
      Unwrap: Remove duplicate asserts, add intl, uintl to test
      Remove unused lapack definition
      BUGFIX for gradient when single element is in new block
      Added AF_INTERP_LOWER and implementation for resize
      Merge pull request #847 from pavanky/docs
      Fixing bug in linear interpolation functions
      FEAT Added nearest neighbour with SSD, SAD and SHD
      Added tests for nearest neighbour
      Added documentation for nearest neighbour
      Merge remote-tracking branch 'upstream/devel' into nearestNeighbour
      Fix double compilation
      Remove redefition warning for blas
      Added options for dotc and dotu to dot function
      Bug fixes for nearest neighbour and hamming
      Added tests for dot
      Compilation and warning fixes
      Add lower interpolation to rotate and transform
      Allow users to set precision when using print
      Changed af_print macro, added documentation
      Fix print macro
      Change build labels to be for devel branch
      Added noDoubleTest for nearest neighbour and dot test
      Added AF_API_VERSION
      Added 64-bit integer type support for functions
      Added print errors to documentation
      Merge pull request #915 from pavanky/fixes
      Fix type in linux doc
      Fix signed-usigned comp warnings
      Fixes to print functions
      FEAT Added saveArray and readArray functions for file read/write
      Removed af_print_array_c. af::print now calls af_print_array_p
      saveArray returns index of array
      Rename af_print_array_p to af_print_array_gen
      FEAT added to string function
      FEAT Added image IO using memory functions and tests
      Fix conjugate transpose for vectors
      Changed saveImageMem API. Added image format enum
      Add typedef af_image_format af::imageFormat
      Updated boost compute version tags
      Remove set_scalar(x, 0) instructions
      Remove unused opencl/kernel/set files
      Fix typo
      Merge pull request #969 from pavanky/svd
      Added missing 3.1 version gaurds
      BUGFIX SVD use gesdd only with MKL, use gesvd with atlas
      Updated release notes for v3.1.0
      Version guards for complex operators
      Added SIFT license info to release notes
      SVD using gesdd on Apple
      Merge pull request #971 from pavanky/assign
      Merge pull request #973 from 9prady9/upstream_updates

Umar Arshad (186):
      Merge pull request #400 from arrayfire/devel
      Remove macro templates in constant. Use Gtest templates
      Merge pull request #527 from 9prady9/docs_improvements
      Add code coverage flags for UNIX platforms
      Add coverall configuration file
      Merge pull request #531 from pavanky/indexing
      Merge pull request #530 from pavanky/softmax
      Added coveralls target to cmake
      Remove old token from coveralls config
      Merge pull request #539 from pavanky/memory
      Remove delete calls
      vector -> unique_ptr for unitilized data. Removed init loops
      Merge pull request #546 from glehmann/fix-unused-function
      Changes based on cppcheck static analysis
      Merge pull request #548 from pavanky/scan_fixes
      Create one cuBlasHandle per GPU. Tests
      Merge pull request #557 from bkloppenborg/cmake_packaging
      Identity support for s64 and u64. Unit Tests
      Fix identity for char and complex types
      Identity unit tests for C++
      Remove unnecessary operator overloads for complex in OpenCL
      Sort functions in docs. Update coveralls
      Merge pull request #568 from pavanky/fft_mem_fix
      Minor cpu/reduce refactor
      Add unit test coverage information in the README
      Fix CMake to create Xcode projects
      Print error messages based on environment vars
      Remove iostream.
      Use relitive path for unit test data path
      Add s64 and u64 support for assign
      Merge pull request #581 from bkloppenborg/docs_external_code
      Remove unnecessary template parameters from tests
      Use explicit for af::array constructors
      Typed tests for alltrue and anytrue
      Diagonal Tests
      Fix char ZERO option bugs in diagonal. Tests. Refactor
      Added support for s64 and u64 to diagonal. Tests
      Merge pull request #604 from pavanky/bug_fixes
      Merge pull request #605 from pavanky/index_fixes
      BUGFIX: adjust a few is* functions
      Unit tests for several af::array member funcitons
      Adjusted failure tests for constant and random
      Minor indexing refactor
      Destructors must not throw exceptions
      Merge pull request #610 from pavanky/dim4
      Merge branch 'devel' of github.com:arrayfire/arrayfire into array_attrib
      Scalar is not a vector
      Simplify isRow/isColumn
      Remove gen_index member function
      Move initBlas to the blas header.
      Including missing headers for call_once in blas.hpp
      Merge pull request #622 from bkloppenborg/linux_installers
      Use non-member non-friend functions to perform binary operations
      Use forward declarations instead of including headers
      Merge branch 'devel' into proxy
      Include array.h in opencl.h: forward decl doesn't work
      Indexing using proxy class. Initial implementation.
      Assign unit tests
      Assignment tests for row and col member functions.
      Assign unit tests
      Assignment tests for row and col member functions.
      Update test data commit
      Merge pull request #632 from pentschev/fix_fast_osx
      Fixed faulty float comparison
      Merge branch 'devel' into proxy
      Unit tests for array_proxy to array_proxy assignment
      Fixes indexed array to indexed array assignment
      Simplify indexing. Fix const correctness.
      Remove unused functions. Cleanup
      Image save tests
      Merge pull request #650 from pentschev/gaussiankernel_tests
      Add s64/u64 type support for lookup
      array_proxy member func tests. Test additional types
      Tests assignment with the slice function.
      Test rows and cols functions.
      remove scalar test.
      Error checks for indexing operations
      Merge pull request #653 from bkloppenborg/install_docs
      Merge pull request #657 from alltheflops/lin_algebra
      Remove unit tests from docs examples
      Merge pull request #664 from pavanky/minor_fixes
      Merge pull request #673 from shehzan10/devel
      Merge pull request #672 from FilipeMaia/fix_missing_operators
      Merge pull request #674 from pavanky/minor_features
      Add tests for where for u64 and s64
      Add support to var for u64 and s64.
      Additional var tests
      Merge pull request #680 from pavanky/features
      Merge pull request #684 from pavanky/api_changes
      Merge pull request #691 from pavanky/bug_fixes
      Test Getting started code examples
      Remove unused gfor parameter from array constructor
      Move matrix_manipulation examples to unit tests.
      Add missing constructors for af::array
      Fix indexing ND using linear indexing
      Partial fix for linear indexing on ND array
      Create indexing unit tests for indexing documentation.
      Apply Rule of 3/5 to array_proxy class.
      Merge branch 'devel' into gs_tests
      BUGFIX: Memory leak in array_proxy operator array()
      BUGFIX/STYLE: Refactor moddims/assign;  Memory leak in assign.
      API: rename af_weak_copy/af_destroy_array to af_retain/af_release
      Merge pull request #698 from pavanky/osx_fixes
      Document index and enums in defines.h
      Merge pull request #703 from pavanky/lapack
      DOCS: Added docs for operator(). Style updates in array.h
      Fix const correctness for row(s)/col(s)
      Merge branch 'devel' into array_docs
      Merge pull request #708 from pavanky/docs
      Merge pull request #707 from arrayfire/graphics
      Merge pull request #714 from pavanky/fixes
      renamed uint to unsigned
      Merge pull request #716 from shehzan10/devel
      API: Update set and reduce API
      API: fftconvolve, meanshift, colorspace, erode/dilate
      API: histequal->histEqual, deviceprop->deviceInfo
      API: Add old names to compatibility header
      Create a deprecated macro.
      Merge pull request #720 from pavanky/bug_fixes
      Merge pull request #717 from shehzan10/devel
      Merge branch 'devel' into array_docs
      DOCS: Additional operator overload docs
      Merge pull request #729 from pavanky/solvers
      Merge pull request #730 from pentschev/fast_example
      Merge pull request #728 from shehzan10/devel
      DOCS: Document rest of array.h. Fix doxygen labels
      Merge pull request #733 from shehzan10/devel
      DOCS: Style updates. Deprecated list
      Updated doxygen to show deprecated list.
      Fix documentation warnings
      Merge fixes
      Merge branch 'pavanky-missing_functions' into devel
      Convert constant into a template
      Added the row(s)/col(s)/slice(s) member functions to proxy
      std::complex -> af_cfloat/af_cdouble
      Add deprecated functions from AF2.1
      Merge branch 'devel' into dep_func
      c++ checks in complex
      Added test to check the validity of the headers in C
      Remove git folder form docs folder
      Resolve ambiguous pow error on osx cuda(6.5)
      Add missing parameter to CREATE_TESTS macro
      Updated fractal example
      Merge pull request #797 from shehzan10/multiple_computes
      Make all link interfaces private
      Fix cuda linking.
      Fix cmake errors in linux for private linking
      Remove references to installer_mode
      Created OSX Installer
      Make osx installer based on targets. Style
      Finalize osx Installer
      Put guards around osx install scripts
      Renamed cpack.txt to cpack.cmake
      Update the installer name
      Make dim4 a POD object
      Make ArrayInfo a POD object
      Add static checks for POD for dim4 and ArrayInfo
      Make Array<T> a standard layout type
      Ensure Array<T> is standard layout using static_asserts
      Revert "Make dim4 a POD object"
      Reduce size of Array<T> by rearranging mem vars
      Add static checks to make sure ArrayInfo is the first mem var
      Merge pull request #848 from pavanky/new_additions
      FEAT: intl/uintl for random
      FEAT: intl/uintl support for all reduce functions
      TEST: Refactor reduce tests. Test intl/uintl
      BUILD: Fix redefinition warning in blas.
      Revert "BUILD: Fix redefinition warning in blas."
      Merge pull request #870 from pavanky/refs
      STYLE: Remove macros; Simplify templates;
      Merge pull request #875 from pavanky/reduce-nan
      Merge pull request #874 from pentschev/harris
      Merge pull request #877 from pavanky/cplx_fixes
      Merge pull request #889 from pavanky/features-3.1
      Merge pull request #895 from FilipeMaia/devel
      Merge pull request #897 from pavanky/minor_changes
      Merge pull request #933 from 9prady9/ker_fixes
      Merge pull request #935 from 9prady9/cudaMem_stream
      Create instances for const index member functions
      Merge pull request #954 from shehzan10/imageio_mem
      Merge pull request #955 from pavanky/wrap
      Merge pull request #958 from shehzan10/ctrans_fix
      Merge pull request #959 from ghisvail/bugfix/remove-unused-dtype-traits
      Merge pull request #966 from pavanky/compile_fixes
      Merge pull request #964 from pavanky/fixes_310
      Merge pull request #965 from pavanky/nonfree_fixes

pradeep (123):
      Code cleanup
      BUGFIX: used a wrong offset for 0th dim in kernels
      Updated contribution guidelines with new wiki page link
      BUGFIX: corrected padding kernel offset
      Unit test for fft on padded Arrays
      HOTFIX:Corrected kernel window lengths meanshift
      Renamed files with image related functions to image
      cuda image rendering resource manager changes
      opencl backend graphics
      BUGFIX: Corrected PBO binding for image cpu
      BUGFIX: fixed vbo index in plot, cuda & opencl backends
      Changes to reflect API change in Forge
      BUGFIX: build/compile changes
      Merge branch 'devel' into graphics
      Updated assets submodule
      Merge remote-tracking branch 'upstream/graphics' into graphics
      Replaced loadFont with loadSystemFont call
      Merge branch 'graphics' into cpu_hist
      Modified histogram draw function
      CUDA backend copy_hist function
      OpenCL backend copy_hist function
      Style fixes for string macro concatenation in sprintf
      BUGFIX fixed signed-unsigned comparison warnings
      BUGFIX added /bigobj CXX flag for opencl on windows
      Added missing cpp wrappers for set functions
      Merge branch 'devel' into graphics
      Merge remote-tracking branch 'upstream/graphics' into graphics
      Added multiview examples: edge filters and morphing
      Added titles for multiple view mode graphics examples
      Modified CATCHALL macro to handle forge exceptions
      ArrayFire Graphics API changes
      BUGFIX fixed histogram draw params in example
      Modified histogram & plot examples to use render loop
      Added default axes labels for plot, hist functions
      Merge branch 'devel' into graphics
      Merge remote-tracking branch 'upstream/graphics' into graphics
      bug fixes: related to convolve API change and OpenCL headers
      BUGFIX fixed graphics namespace when not needed
      Removed uncessary dependency GLFW from ArrayFire
      Merge branch 'devel' into graphics
      Merge branch 'devel' into graphics
      Added copy-via-host fallback option for opencl graphics
      Merge branch 'devel' into graphics
      Style fixes
      Fixed segfault error when graphics is not used for opencl
      Changed af::Window::operator() to return reference
      Added checks for graphics calls
      Fixed histequal to return result with same dims as input
      BUGFIX fixed ForgeManager caching mechanism
      BUGFIX: corrected af_min_t to min for char specilization
      Added image processing examples
      Added set position function to af::Window
      Offseted window start position in pyramids examples
      Modified harris example to use graphics when appropriate
      Added fractal examples
      Fixed math header in optical flow example
      Modified fractal array to be normalized before rendering
      Added colormap attribute to af::Window
      Added colormap option to examples to make them look pretty
      Corrected doxygen group tag for matchTemplate C++ API
      Documentation for af::Window class and graphics C API
      Merge branch 'devel' into cmap_additions
      BUGFIX in jit opencl
      Graphics changes to reflect changes in upstream library forge
      Fixed glew header search hints for cmake
      Removed the need to Find glfw/glew from build_forge script
      updated external project forge tag
      Fix for displaying images of type uchar, int and uint
      Removed a printf from opencl backend
      Changed forge tag to af3.0.1 in build_forge.cmake
      FEAT: Difference of Gaussians
      Changes in examples to reflect asset modifications
      turned off clFFT examples in external project build
      fix forge dependency target errors in cpu, opencl backends
      Made cpu::Array constructor consistent with CUDA & OpenCL
      Fixed filenames for the files used in computer vision examples
      SUSAN Corner Detector
      Added error checks non-array parameters of susan API
      CUDA backend for SUSAN dectector
      OpenCL backend for SUSAN dectector
      fixes: typos; specific to windows; additional unit test
      Corrected memory allocation bug in cpu backend for orb, where
      Removed uncessary corner sorting for SUSAN
      matchTemplate example
      matchTemplate fix in opencl to support indexed template images
      Removed uncessary copy in matchTemplate example
      Added heat colormap to display disparity values for matching
      Changed CUDA/OpenCL kernels to use zero leading dimension
      typo fix in cuda SUSAN kernel
      SUSAN CUDA/OpenCL: Added bound checks to load shared/local Memory
      Replaced static shared memory with dynamic in SUSAN CUDA kernel
      Changed default cuda stream to be non-zero
      Wrapped cuda kernel launches with CUDA_LAUNCH macro
      FEAT: Summed Area Tables (sat, af_sat) a.k.a integral images
      Added stream parameter for upstream{thrust, cufft, cublas} calls
      Added check for skipping double type test when not supported
      Added uintl, intl support for jit operations in cuda backend
      thrust fixes for cuda stream selection on cuda < 7.0
      shared/local memory loading fix
      namespace fix for POST_LAUNCH_CHECK macro: cuda backend
      Replaced cuda Memcopy/Memset with async versions
      typo fix in opencl morph kernel
      graphics window set size functions
      style fixes in graphics examples
      YCbCr <-> RGB conversion functions
      unit tests for YCbCr <-> RGB conversion
      Documentation for ycbcr_rgb conversion functions
      Modified colorspace function to handle new colorspace - YCbCr
      Modifed colorspace function wrapper code for efficiency
      Replaced padArray calls with Jit operations in ycbcr conversions
      Updated forge tag
      modified cpu::getInfo to display CPU information
      Fixes for cpu backend getInfo on Tegra platform
      Fixed missing header error for windows platform
      adding GL headers in platform.cpp
      Moved GL headers in platform.cpp inside WITH_GRAPHICS block
      OpenCL backend af_info function string fixes
      Documentation for CUDA backend specific API
      Documentation fixes for 3.1 release
      Updated forge upstream tag for 3.1 release
      Documentation for missing index header functions
      Added missing docs for complex and opencl backend specific fns
      Added missing docs for operator%, array::H and array::T

unbornchikken (2):
      This fixes that serious memory leak in OpenCL backend, and gets rid of unnecessary CLBuffer referencing.
      Retain buffer added.

xumbu (1):
      Fixed 'snprintf' definition conflict in Visual Studio 2015

-----------------------------------------------------------------------

No new revisions were added by this update.

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-science/packages/arrayfire.git



More information about the debian-science-commits mailing list