Commit Graph

9662 Commits

Author SHA1 Message Date
Gael Guennebaud
6190aa5632 bug #1567: add optimized path for tensor broadcasting and 'Channel First' shape 2018-07-09 11:23:16 +02:00
Gael Guennebaud
ec323b7e66 Skip null numerators in triangular-vector-solve (as in BLAS TRSV). 2018-07-09 11:13:19 +02:00
Gael Guennebaud
359dd77ec3 Fix legitimate "declaration shadows a typedef" warning 2018-07-09 11:03:39 +02:00
Mark D Ryan
90a53ca6fd Fix the Packet16h version of ptranspose
The AVX512 version of ptranpose for PacketBlock<Packet16h,16> was
reordering the PacketBlock argument incorrectly.  This lead to errors in
the multiplication of matrices composed of 16 bit floats on AVX512
machines, if at least of the matrices was using RowMajor order.  This
error is responsible for one tensorflow unit test failure on AVX512
machines:

//tensorflow/python/kernel_tests:batch_matmul_op_test
2018-06-16 15:13:06 -07:00
Gael Guennebaud
1f54164eca Fix a few issues with Packet16h 2018-07-07 00:15:07 +02:00
Gael Guennebaud
f2dc048df9 complete implementation of Packet16h (AVX512) 2018-07-06 17:43:11 +02:00
Gael Guennebaud
a937c50208 palign is not used anymore, so let's relax the unit test 2018-07-06 17:41:52 +02:00
Gael Guennebaud
56a33ae57d test product kernel with half-floats. 2018-07-06 17:14:04 +02:00
Gael Guennebaud
f4d623ffa7 Complete Packet8h implementation and test it in packetmath unit test 2018-07-06 17:13:36 +02:00
Gael Guennebaud
a8ab6060df Add unitests for inverse and selfadjoint-eigenvalues on CUDA 2018-07-06 09:58:45 +02:00
Andrea Bocci
f7124b3e46 Extend CUDA support to matrix inversion and selfadjointeigensolver 2018-06-11 18:33:24 +02:00
Gael Guennebaud
0537123953 bug #1565: help MSVC to generatenot too bad ASM in reductions. 2018-07-05 09:21:26 +02:00
Gael Guennebaud
6a241bd8ee Implement custom inplace triangular product to avoid a temporary 2018-07-03 14:02:46 +02:00
Gael Guennebaud
3ae2083e23 Make is_same_dense compatible with different scalar types. 2018-07-03 13:21:43 +02:00
Gael Guennebaud
67ec37f7b0 Activate dgmres unit test 2018-07-02 12:54:14 +02:00
Gael Guennebaud
047677a08d Fix regression in changeset f05dea6b23
: computeFromHessenberg can take any expression for matrixQ, not only an HouseholderSequence.
2018-07-02 12:18:25 +02:00
Gael Guennebaud
d625564936 Simplify redux_evaluator using inheritance, and properly rename parameters in reducers. 2018-07-02 11:50:41 +02:00
Gael Guennebaud
d428a199ab bug #1562: optimize evaluation of small products of the form s*A*B by rewriting them as: s*(A.lazyProduct(B)) to save a costly temporary. Measured speedup from 2x to 5x... 2018-07-02 11:41:09 +02:00
Gael Guennebaud
a7b313a16c Fix unit test 2018-07-01 22:45:47 +02:00
Gael Guennebaud
0cdacf3fa4 update comment 2018-06-29 11:28:36 +02:00
Gael Guennebaud
54f6eeda90 Merged in net147/eigen (pull request PR-411)
Use std::complex constructor instead of assignment from scalar
2018-06-28 21:01:04 +00:00
Gael Guennebaud
9a81de1d35 Fix order of EIGEN_DEVICE_FUNC and returned type 2018-06-28 00:20:59 +02:00
Jonathan Liu
b7689bded9 Use std::complex constructor instead of assignment from scalar
Fixes GCC conversion to non-scalar type requested compile error when
using boost::multiprecision::cpp_dec_float_50 as scalar type.
2018-06-28 00:32:37 +10:00
Gael Guennebaud
f9d337780d First step towards a generic vectorised quaternion product 2018-06-25 14:26:51 +02:00
Gael Guennebaud
ee5864f72e bug #1560 fix product with a 1x1 diagonal matrix 2018-06-25 10:30:12 +02:00
Rasmus Munk Larsen
2f62cc68cd merge 2018-06-22 15:09:44 -07:00
Rasmus Munk Larsen
bda71ad394 Fix typo in pbend for AltiVec. 2018-06-22 15:04:35 -07:00
Benoit Steiner
b6ffcd22e3 Merged in rmlarsen/eigen2 (pull request PR-409)
Fix oversharding bug in parallelFor.
2018-06-21 18:34:57 +00:00
Gael Guennebaud
4cc32d80fd bug #1555: compilation fix with XLC 2018-06-21 10:28:38 +02:00
Rasmus Munk Larsen
5418154a45 Fix oversharding bug in parallelFor. 2018-06-20 17:51:48 -07:00
Gael Guennebaud
b8271bb368 fix md5sum of lapack_addons 2018-06-15 14:21:29 +02:00
Benoit Steiner
d3a380af4d Merged in mfigurnov/eigen/gamma-der-a (pull request PR-403)
Derivative of the incomplete Gamma function and the sample of a Gamma random variable

Approved-by: Benoit Steiner <benoit.steiner.goog@gmail.com>
2018-06-11 17:57:47 +00:00
Gael Guennebaud
cb4c9a6a94 bug #1531: make dedicatd unit testing for NumDimensions 2018-06-08 17:11:45 +02:00
Gael Guennebaud
d6813fb1c5 bug #1531: expose NumDimensions for solve and sparse expressions. 2018-06-08 16:55:10 +02:00
Gael Guennebaud
89d65bb9d6 bug #1531: expose NumDimensions for compatibility with Tensor 2018-06-08 16:50:17 +02:00
Gael Guennebaud
f05dea6b23 bug #1550: prevent avoidable memory allocation in RealSchur 2018-06-08 10:14:57 +02:00
Gael Guennebaud
7933267c67 fix prototype 2018-06-08 09:56:01 +02:00
Gael Guennebaud
f4d1461874 Fix the way matrix folder is passed to the tests. 2018-06-08 09:55:46 +02:00
Benoit Steiner
522d3ca54d Don't use std::equal_to inside cuda kernels since it's not supported. 2018-06-07 13:02:07 -07:00
Christoph Hertzberg
7d7bb91537 Missing line during manual rebase of PR-374 2018-06-07 20:30:09 +02:00
Michael Figurnov
30fa3d0454 Merge from eigen/eigen 2018-06-07 17:57:56 +01:00
Benoit Steiner
d2b0a4a59b Merged in mfigurnov/eigen/fix-bessel (pull request PR-404)
Fix compilation of special functions without C99 math.
2018-06-07 16:12:42 +00:00
Michael Figurnov
6c71c7d360 Merge from eigen/eigen. 2018-06-07 15:54:18 +01:00
Gael Guennebaud
c25034710e Fiw some warnings in dox examples 2018-06-07 16:09:22 +02:00
Gael Guennebaud
37348d03ae Fix int versus Index 2018-06-07 15:56:43 +02:00
Gael Guennebaud
c723ffd763 Fix warning 2018-06-07 15:56:20 +02:00
Gael Guennebaud
af7c83b9a2 Fix warning 2018-06-07 15:45:24 +02:00
Gael Guennebaud
7fe29aceeb Fix MSVC warning C4290: C++ exception specification ignored except to indicate a function is not __declspec(nothrow) 2018-06-07 15:36:20 +02:00
Michael Figurnov
aa813d417b Fix compilation of special functions without C99 math.
The commit with Bessel functions i0e and i1e placed the ifdef/endif incorrectly,
causing i0e/i1e to be undefined when EIGEN_HAS_C99_MATH=0. These functions do not
actually require C99 math, so now they are always available.
2018-06-07 14:35:07 +01:00
Gael Guennebaud
55774b48e4 Fix short vs long 2018-06-07 15:26:25 +02:00