Mark D Ryan
bc615e4585
Re-enable FMA for fast sqrt functions
2018-07-30 13:21:00 +02:00
Mark D Ryan
96b030a8e4
Re-enable FMA for fast sqrt functions
...
This commit re-enables the use of FMA for the FAST sqrt functions.
Doing so improves the performance of both algorithms. The float32
version is now 88% the speed of the original function, while the
double version is 90%.
2018-07-30 10:19:51 +01:00
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
Rasmus Munk Larsen
2ebcb911b2
Add pcast packet op for NEON.
2018-07-26 14:28:48 -07:00
Christoph Hertzberg
397b0547e1
DIsable static assertions only when necessary and disable double-promotion warnings in that case as well
2018-07-26 00:01:24 +02:00
Christoph Hertzberg
5e79402b4a
fix warnings for doc-eigen-prerequisites
2018-07-24 21:59:15 +02:00
Christoph Hertzberg
5f79b7f9a9
Removed several shadowing types and use global Index typedef everywhere
2018-07-25 21:47:45 +02:00
Christoph Hertzberg
44ee201337
Rename variable which shadows class name
2018-07-25 20:26:15 +02:00
Gustavo Lima Chaves
705f66a9ca
Account for missing change on commit "Remove SimpleThreadPool and..."
...
"... always use {NonBlocking}ThreadPool". It seems the non-blocking
implementation was me the default/only one, but a reference to the old
name was left unmodified. Fix that.
2018-07-23 16:29:09 -07:00
Christoph Hertzberg
fd4fe7cbc5
Fixed issue which made documentation not getting built anymore
2018-07-24 22:56:15 +02:00
Christoph Hertzberg
636126ef40
Allow to filter out build-error messages
2018-07-24 20:12:49 +02:00
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
Gael Guennebaud
2c2de9da7d
Merged in glchaves/eigen (pull request PR-433)
...
Move cxx11_tensor_uint128 test under an EIGEN_TEST_CXX11 guarded block
2018-07-23 19:38:55 +00:00
Gael Guennebaud
4ca3e48f42
fix typo
2018-07-23 16:51:57 +02:00
Gael Guennebaud
c747cde69a
Add lastN shorcuts to seq/seqN.
2018-07-23 16:20:25 +02:00
Gustavo Lima Chaves
02eaaacbc5
Move cxx11_tensor_uint128 test under an EIGEN_TEST_CXX11 guarded
...
block
Builds configured without the -DEIGEN_TEST_CXX11=ON flag would fail
right away without this, as this test seems to rely on those language
features. The skip under compilation with MSVC was kept.
2018-07-20 16:08:40 -07:00
Eugene Zhulenev
2bf864f1eb
Disable type traits for stdlibc++ <= 4.9.3
2018-07-20 10:11:44 -07:00
Gael Guennebaud
de70671937
Oopps, EIGEN_COMP_MSVC is not available before including Eigen.
2018-07-20 17:51:17 +02:00
Gael Guennebaud
56a750b6cc
Disable optimization for sparse_product unit test with MSVC 2013, otherwise it takes several hours to build.
2018-07-20 08:36:38 -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
Gael Guennebaud
2424e3b7ac
Pass by const ref.
2018-07-19 18:48:19 +02:00
Gael Guennebaud
509a5fa77f
Fix IsRelocatable without C++11
2018-07-19 18:47:38 +02:00
Gael Guennebaud
2ca2592009
Fix determination of EIGEN_HAS_TYPE_TRAITS
2018-07-19 18:47:18 +02:00
Gael Guennebaud
5e5987996f
Fix stupid error in Quaternion move ctor
2018-07-19 18:33:53 +02:00
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
Gael Guennebaud
40797dbea3
bug #1572 : use c++11 atomic instead of volatile if c++11 is available, and disable multi-threaded GEMM on non-x86 without c++11.
2018-07-17 00:11:20 +02:00
Gael Guennebaud
add5757488
Simplify handling and non-splitted tests and include split_test_helper.h instead of re-generating it. This also allows us to modify it without breaking existing build folder.
2018-07-16 18:55:40 +02:00
Gael Guennebaud
901c7d31f0
Fix usage of EIGEN_SPLIT_LARGE_TESTS=ON: some unit tests, such as indexed_view have to be split unconditionally.
2018-07-16 18:35:05 +02:00
Gael Guennebaud
f2b52f9946
Add the cmake option "EIGEN_DASHBOARD_BUILD_TARGET" to control the build target in dashboard mode (e.g., ctest -D Experimental)
2018-07-16 17:59:30 +02:00
Gael Guennebaud
23d82c1ac5
Merged in rmlarsen/eigen2 (pull request PR-422)
...
Optimize the case where broadcasting is a no-op.
2018-07-14 11:42:58 +00:00
Gael Guennebaud
a87cff20df
Fix GeneralizedEigenSolver when requesting for eigenvalues only.
2018-07-14 09:38:49 +02:00
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
Rasmus Munk Larsen
4a3952fd55
Relax the condition to not only work on Android.
2018-07-13 11:24:07 -07:00
Rasmus Munk Larsen
02a9443db9
Clang produces incorrect Thumb2 assembler when using alloca.
...
Don't define EIGEN_ALLOCA when generating Thumb with clang.
2018-07-13 11:03:04 -07:00
Gael Guennebaud
20991c3203
bug #1571 : fix is_convertible<from,to> with "from" a reference.
2018-07-13 17:47:28 +02:00
Gael Guennebaud
1920129d71
Remove clang warning
2018-07-13 16:05:35 +02:00
Gael Guennebaud
195c9c054b
Print more debug info in gpu_basic
2018-07-13 16:05:07 +02:00
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
Gael Guennebaud
5fd03ddbfb
Make EIGEN_TEST_CUDA_CLANG more friendly with OSX
2018-07-13 16:03:14 +02:00
Gael Guennebaud
86d9c0255c
Forward declaring std::array does not work with all std libs, so let's just include <array>
2018-07-13 13:06:44 +02:00
David Hyde
d908afe35f
bug #1558 : fix a corner case in MINRES when both v_new and w_new vanish.
2018-07-08 22:06:38 -07:00
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
Alexey Frunze
1f523e7304
Add MIPS changes missing from previous merge.
2018-07-18 12:27:50 -07:00
Eugene Zhulenev
e3c2d61739
Assert that no output kernel is defined for GPU contraction
2018-07-18 14:34:22 -07:00
Eugene Zhulenev
086ded5c85
Disable type traits for GCC < 5.1.0
2018-07-18 16:32:55 -07:00