Commit Graph

1036 Commits

Author SHA1 Message Date
Rasmus Munk Larsen
fa68342ef8 Move sigmoid functor to core. 2018-08-03 17:31:23 -07:00
Rasmus Munk Larsen
bcb29f890c Fix initialization order. 2018-08-03 10:18:53 -07:00
Mehdi Goli
3074b1ff9e Fixing the compilation error. 2018-08-03 17:13:44 +01:00
Mehdi Goli
01358300d5 Creating separate SYCL required PR for uncontroversial files. 2018-08-03 16:59:15 +01:00
Benoit Steiner
93b9e36e10 Merged in paultucker/eigen (pull request PR-431)
Optional ThreadPoolDevice allocator

Approved-by: Benoit Steiner <benoit.steiner.goog@gmail.com>
2018-08-01 19:14:34 +00:00
Benoit Steiner
17221115c9 Merged in codeplaysoftware/eigen-upstream-pure/eigen_variadic_assert (pull request PR-447)
Adding variadic version of assert which can take a parameter pack as its input.
2018-08-01 16:41:54 +00:00
Benoit Steiner
0360c36170 Merged in codeplaysoftware/eigen-upstream-pure/separating_internal_memory_allocation (pull request PR-446)
Distinguishing between internal memory allocation/deallocation from explicit user memory allocation/deallocation.
2018-08-01 16:13:15 +00:00
Mehdi Goli
c6a5c70712 Correcting the position of allocate_temp/deallocate_temp in TensorDeviceGpu.h 2018-08-01 16:56:26 +01:00
Benoit Steiner
45f75f1ace Merged in codeplaysoftware/eigen-upstream-pure/using_PacketType_class (pull request PR-449)
Enabling per device specialisation of packetSize.
2018-08-01 15:43:03 +00:00
Mehdi Goli
af96018b49 Using the suggested modification. 2018-08-01 16:04:44 +01:00
Mehdi Goli
b512a9536f Enabling per device specialisation of packetsize. 2018-08-01 13:39:13 +01:00
Mehdi Goli
3a197a60e6 variadic version of assert which can take a parameter pack as its input. 2018-08-01 12:19:14 +01:00
Mehdi Goli
d7a8414848 Distinguishing between internal memory allocation/deallocation from explicit user memory allocation/deallocation. 2018-08-01 11:56:30 +01:00
Mehdi Goli
9e219bb3d3 Converting ad-hoc inline keyword to EIGEN_STRONG_INLINE MACRO. 2018-08-01 10:47:49 +01:00
Benoit Steiner
edf46bd7a2 Merged in yuefengz/eigen (pull request PR-370)
Use device's allocate function instead of internal::aligned_malloc.
2018-07-31 22:38:28 +00:00
Paul Tucker
385f7b8d0c Change getAllocator() to allocator() in ThreadPoolDevice. 2018-07-31 13:52:18 -07:00
Gael Guennebaud
678a0dcb12 Merged in ezhulenev/eigen/tiling_3 (pull request PR-438)
Tiled tensor executor
2018-07-31 08:13:00 +00:00
Gael Guennebaud
679eece876 Speedup trivial tensor broadcasting on GPU by enforcing unaligned loads. See PR 437. 2018-07-31 10:10:14 +02:00
Eugene Zhulenev
966c2a7bb6 Rename Index to StorageIndex + use Eigen::Array and Eigen::Map when possible 2018-07-27 12:45:17 -07:00
Eugene Zhulenev
6913221c43 Add tiled evaluation support to TensorExecutor 2018-07-25 13:51:10 -07:00
Rasmus Munk Larsen
e478532625 Reduce the number of template specializations of classes related to tensor contraction to reduce binary size. 2018-07-27 12:36:34 -07:00
Eugene Zhulenev
d55efa6f0f TensorBlockIO 2018-07-23 15:50:55 -07:00
Eugene Zhulenev
34a75c3c5c Initial support of TensorBlock 2018-07-20 17:37:20 -07:00
Paul Tucker
d4afccde5a Add test coverage for ThreadPoolDevice optional allocator. 2018-07-19 17:43:44 -07:00
Eugene Zhulenev
c58b874727 PR430: Convert count to the reducer type in MeanReducer
Without explicit conversion Tensorflow fails to compile, pset1 template deduction fails.

cannot convert '((const Eigen::internal::MeanReducer<Eigen::half>*)this)
  ->Eigen::internal::MeanReducer<Eigen::half>::packetCount_'
(type 'const DenseIndex {aka const long int}')
to type 'const type& {aka const Eigen::half&}'
     return pdiv(vaccum, pset1<Packet>(packetCount_));
Honestly I’m not sure why it works in Eigen tests, because Eigen::half constructor is explicit, and why it stopped working in TF, I didn’t find any relevant changes since previous Eigen upgrade.

static_cast<T>(packetCount_) - breaks cxx11_tensor_reductions test for Eigen::half, also quite surprising.
2018-07-19 17:37:03 -07:00
Paul Tucker
4e9848fa86 Actually add optional Allocator* arg to ThreadPoolDevice(). 2018-07-16 17:53:36 -07:00
Paul Tucker
b3e7c9132d Add optional Allocator argument to ThreadPoolDevice constructor.
When supplied, this allocator will be used in place of
internal::aligned_malloc.  This permits e.g. use of a NUMA-node specific
allocator where the thread-pool is also restricted a single NUMA-node.
2018-07-16 17:26:05 -07:00
Rasmus Munk Larsen
3a9cf4e290 Get rid of alias for m_broadcast. 2018-07-13 16:24:48 -07:00
Rasmus Munk Larsen
4222550e17 Optimize the case where broadcasting is a no-op. 2018-07-13 16:12:38 -07:00
Gael Guennebaud
06eb24cf4d Introduce gpu_assert for assertion in device-code, and disable them with clang-cuda. 2018-07-13 16:04:27 +02:00
Eugene Zhulenev
6e654f3379 Reduce number of allocations in TensorContractionThreadPool. 2018-07-16 14:26:39 -07:00
Gael Guennebaud
7ccb623746 bug #1569: fix Tensor<half>::mean() on AVX with respective unit test. 2018-07-19 13:15:40 +02:00
Eugene Zhulenev
e3c2d61739 Assert that no output kernel is defined for GPU contraction 2018-07-18 14:34:22 -07:00
Eugene Zhulenev
79d4129cce Specify default output kernel for TensorContractionOp 2018-07-18 14:21:01 -07:00
Yuefeng Zhou
1eff6cf8a7 Use device's allocate function instead of internal::aligned_malloc. This would make it easier to track memory usage in device instances. 2018-02-20 16:50:05 -08:00
Gael Guennebaud
6cd6551b26 Add deprecated header files for TensorFlow 2018-07-12 10:50:53 +02:00
Deven Desai
876f392c39 Updates corresponding to the latest round of PR feedback
The major changes are

1. Moving CUDA/PacketMath.h to GPU/PacketMath.h
2. Moving CUDA/MathFunctions.h to GPU/MathFunction.h
3. Moving CUDA/CudaSpecialFunctions.h to GPU/GpuSpecialFunctions.h
    The above three changes effectively enable the Eigen "Packet" layer for the HIP platform

4. Merging the "hip_basic" and "cuda_basic" unit tests into one ("gpu_basic")
5. Updating the "EIGEN_DEVICE_FUNC" marking in some places

The change has been tested on the HIP and CUDA platforms.
2018-07-11 10:39:54 -04:00
Deven Desai
38807a2575 merging updates from upstream 2018-07-11 09:17:33 -04:00
Gael Guennebaud
6190aa5632 bug #1567: add optimized path for tensor broadcasting and 'Channel First' shape 2018-07-09 11:23:16 +02:00
Deven Desai
1bb6fa99a3 merging the CUDA and HIP implementation for the Tensor directory and the unit tests 2018-06-20 16:44:58 -04:00
Deven Desai
cfdabbcc8f removing the *Hip files from the unsupported/Eigen/CXX11/src/Tensor and unsupported/test directories 2018-06-20 12:57:02 -04:00
Deven Desai
7e41c8f1a9 renaming *Cuda files to *Gpu in the unsupported/Eigen/CXX11/src/Tensor and unsupported/test directories 2018-06-20 12:52:30 -04:00
Deven Desai
b6cc0961b1 updates based on PR feedback
There are two major changes (and a few minor ones which are not listed here...see PR discussion for details)

1. Eigen::half implementations for HIP and CUDA have been merged.
This means that
- `CUDA/Half.h` and `HIP/hcc/Half.h` got merged to a new file `GPU/Half.h`
- `CUDA/PacketMathHalf.h` and `HIP/hcc/PacketMathHalf.h` got merged to a new file `GPU/PacketMathHalf.h`
- `CUDA/TypeCasting.h` and `HIP/hcc/TypeCasting.h` got merged to a new file `GPU/TypeCasting.h`

After this change the `HIP/hcc` directory only contains one file `math_constants.h`. That will go away too once that file becomes a part of the HIP install.

2. new macros EIGEN_GPUCC, EIGEN_GPU_COMPILE_PHASE and EIGEN_HAS_GPU_FP16 have been added and the code has been updated to use them where appropriate.
- `EIGEN_GPUCC` is the same as `(EIGEN_CUDACC || EIGEN_HIPCC)`
- `EIGEN_GPU_DEVICE_COMPILE` is the same as `(EIGEN_CUDA_ARCH || EIGEN_HIP_DEVICE_COMPILE)`
- `EIGEN_HAS_GPU_FP16` is the same as `(EIGEN_HAS_CUDA_FP16 or EIGEN_HAS_HIP_FP16)`
2018-06-14 10:21:54 -04:00
Deven Desai
d1d22ef0f4 syncing this fork with upstream 2018-06-13 12:09:52 -04:00
Rasmus Munk Larsen
5418154a45 Fix oversharding bug in parallelFor. 2018-06-20 17:51:48 -07:00
Michael Figurnov
30fa3d0454 Merge from eigen/eigen 2018-06-07 17:57:56 +01:00
Gael Guennebaud
b3fd93207b Fix typos found using codespell 2018-06-07 14:43:02 +02:00
Michael Figurnov
4bd158fa37 Derivative of the incomplete Gamma function and the sample of a Gamma random variable.
In addition to igamma(a, x), this code implements:
* igamma_der_a(a, x) = d igamma(a, x) / da -- derivative of igamma with respect to the parameter
* gamma_sample_der_alpha(alpha, sample) -- reparameterization derivative of a Gamma(alpha, 1) random variable sample with respect to the alpha parameter

The derivatives are computed by forward mode differentiation of the igamma(a, x) code. Although gamma_sample_der_alpha can be implemented via igamma_der_a, a separate function is more accurate and efficient due to analytical cancellation of some terms. All three functions are implemented by a method parameterized with "mode" that always computes the derivatives, but does not return them unless required by the mode. The compiler is expected to (and, based on benchmarks, does) skip the unnecessary computations depending on the mode.
2018-06-06 18:49:26 +01:00
Deven Desai
8fbd47052b Adding support for using Eigen in HIP kernels.
This commit enables the use of Eigen on HIP kernels / AMD GPUs. Support has been added along the same lines as what already exists for using Eigen in CUDA kernels / NVidia GPUs.

Application code needs to explicitly define EIGEN_USE_HIP when using Eigen in HIP kernels. This is because some of the CUDA headers get picked up by default during Eigen compile (irrespective of whether or not the underlying compiler is CUDACC/NVCC, for e.g. Eigen/src/Core/arch/CUDA/Half.h). In order to maintain this behavior, the EIGEN_USE_HIP macro is used to switch to using the HIP version of those header files (see Eigen/Core and unsupported/Eigen/CXX11/Tensor)


Use the "-DEIGEN_TEST_HIP" cmake option to enable the HIP specific unit tests.
2018-06-06 10:12:58 -04:00
Benoit Steiner
e206f8d4a4 Merged in mfigurnov/eigen (pull request PR-400)
Exponentially scaled modified Bessel functions of order zero and one.

Approved-by: Benoit Steiner <benoit.steiner.goog@gmail.com>
2018-06-05 17:05:21 +00:00