Benoit Steiner
67fcf47ecb
Merged from trunk
2014-10-30 21:59:22 -07:00
Benoit Steiner
fcecafde3a
Fixed a compilation error with clang
2014-10-30 21:58:14 -07:00
Benoit Steiner
d62bfe73a9
Use the proper index type in the padding code
2014-10-30 18:15:05 -07:00
Benoit Steiner
bc99c5f7db
fixed some potential alignment issues.
2014-10-30 18:09:53 -07:00
Benoit Steiner
5e62427e22
Use the proper index type
2014-10-30 17:49:39 -07:00
Gael Guennebaud
21c0a2ce0c
Move D&C SVD to official SVD module.
2014-10-29 11:29:33 +01:00
Benoit Steiner
debc97821c
Added support for tensor references
2014-10-28 23:10:13 -07:00
Benjamin Chrétien
c426054767
BDCSVD: fix CMake install (missing separator).
2014-10-24 15:10:56 +02:00
Christoph Hertzberg
1fa793cb97
Removed weird self assignment.
2014-10-24 13:19:19 +02:00
Gael Guennebaud
fe57b2f963
bug #701 : workaround (min) and (max) blocking ADL by introducing numext::mini and numext::maxi internal functions and a EIGEN_NOT_A_MACRO macro.
2014-10-20 15:55:32 +02:00
Benoit Steiner
f786897e4b
Added access to the unerlying raw data of a tnsor slice/chip whenever possible
2014-10-17 15:33:27 -07:00
Benoit Steiner
7acd38d19e
Created some benchmarks for the tensor code
2014-10-17 09:49:03 -07:00
Gael Guennebaud
4b7c3abbea
Fix D&C SVD wrt zero matrices
2014-10-17 15:32:55 +02:00
Benoit Steiner
65af852b54
Silenced one last warning
2014-10-16 15:02:30 -07:00
Benoit Steiner
ae697b471c
Silenced a few compilation warnings
...
Generalized a TensorMap constructor
2014-10-16 14:52:50 -07:00
Benoit Steiner
94e47798f4
Fixed the return types of unary and binary expressions to properly handle the case where it is different from the input type (e.g. abs(complex<float>))
2014-10-16 10:41:07 -07:00
Benoit Steiner
d853adffdb
Avoid calling get_future() more than once on a given promise.
2014-10-16 10:10:04 -07:00
Benoit Steiner
bfdd9f3ac9
Made the blocking computation aware of the l3 cache
...
Also optimized the blocking parameters to take into account the number of threads used for a computation
2014-10-15 15:32:59 -07:00
Benoit Steiner
dba55041ab
Added support for promises
...
Started to improve multithreaded contractions
2014-10-15 11:20:36 -07:00
Gael Guennebaud
2cc41dbe83
D&C SVD: fix some numerical issues by truly skipping deflated singular values when computing them
2014-10-15 15:21:12 +02:00
Gael Guennebaud
c26e8a1af3
D&C SVD: fix deflation of repeated singular values, fix sorting of singular values, fix case of complete deflation
2014-10-15 11:59:21 +02:00
Benoit Steiner
99d75235a9
Misc improvements and cleanups
2014-10-13 17:02:09 -07:00
Benoit Steiner
4c70b0a762
Added support for patch extraction
2014-10-13 10:04:04 -07:00
Benoit Steiner
0219f8aed4
Added ability to print a tensor using an iostream.
2014-10-10 16:17:26 -07:00
Benoit Steiner
2ed1838aeb
Added support for tensor chips
2014-10-10 16:11:27 -07:00
Benoit Steiner
4b36c3591f
Fixed the tensor shuffling test
2014-10-10 15:43:21 -07:00
Benoit Steiner
a991f94c0e
Fixed the thread pool test
2014-10-10 15:20:37 -07:00
Benoit Steiner
498b7eed25
Rewrote the TensorBase::random method to support the generation of random number on gpu.
2014-10-09 15:39:13 -07:00
Benoit Steiner
767424af18
Improved the functors defined for standard reductions
...
Added a functor to encapsulate the generation of random numbers on cpu and gpu.
2014-10-09 15:36:23 -07:00
Gael Guennebaud
ccd70ba123
Various numerical fixes in D&C SVD: I cannot make it fail with double, but still need to tune for single precision, and carefully test with duplicated singular values
2014-10-09 23:29:01 +02:00
Benoit Steiner
44beee9d68
Removed dead code
2014-10-08 14:14:20 -07:00
Benoit Steiner
0a07ac574e
Added support for the *= and /* operators to TensorBase
2014-10-08 13:32:41 -07:00
Benoit Steiner
6c047d398d
Fixed a comment
2014-10-08 13:29:36 -07:00
Gael Guennebaud
118b1113d9
Workaround MSVC issue.
2014-10-07 09:53:39 +02:00
Gael Guennebaud
503c176d8e
Fix missing outer() member in DynamicSparseMatrix
2014-10-07 09:53:27 +02:00
Gael Guennebaud
dbdd8b0883
D&C SVD: add scaling to avoid overflow, fix handling of fixed size matrices
2014-10-06 19:35:57 +02:00
Benoit Steiner
152f3218ac
Improved contraction test
2014-10-03 19:33:44 -07:00
Benoit Steiner
af2e5995e2
Improved support for CUDA devices.
...
Improved contractions on GPU
2014-10-03 19:18:07 -07:00
Benoit Steiner
1269392822
Created the IndexPair type to store pair of tensor indices. CUDA doesn't support std::pair so we can't use them when targeting GPUs.
...
Improved the performance on tensor contractions
2014-10-03 10:16:59 -07:00
Benoit Steiner
b7271dffb5
Generalized the gebp apis
2014-10-02 16:51:57 -07:00
Benoit Steiner
8b2afe33a1
Fixes for the forced evaluation of tensor expressions
...
More tests
2014-10-02 10:39:36 -07:00
Benoit Steiner
5cc23199be
More tests to validate the const-correctness of the tensor code.
2014-10-02 10:30:44 -07:00
Benoit Steiner
7caaf6453b
Added support for tensor reductions and concatenations
2014-10-01 20:38:22 -07:00
Benoit Steiner
1c236f4c9a
Added tests for tensors of const values and tensors of stringswwq::
2014-10-01 20:21:42 -07:00
Christoph Hertzberg
6d26deb894
Missing outerStride in AlignedVector3 resulted in infinite recursion
2014-09-30 16:43:19 +02:00
Gael Guennebaud
842e31cf5c
Let KroneckerProduct exploits the recently introduced generic InnerIterator class.
2014-09-29 13:37:49 +02:00
Georg Drenkhahn
8a502233d8
Correcting the ReturnType in traits<KroneckerProduct<>> to include the correct Index type.
...
Fixed mixup of types Rhs::Index and Lhs:Index in various loop variables.
Added explicit type conversion for arithmetic expressions which may return a wider type.
2014-09-21 23:19:29 +02:00
Georg Drenkhahn
3bd31e21b5
Fixed compiler warning on implicit integer conversion by separating index type for matrix and permutation matrix which may not be equal.
2014-09-20 15:00:36 +02:00
Georg Drenkhahn
75e269c77b
Fixed warning on implicit integer conversion in test case code by using type VectorXd::Index instead of int.
2014-09-20 14:57:42 +02:00
Gael Guennebaud
abba11bdcf
Many improvements in Divide&Conquer SVD:
...
- Fix many numerical issues, in particular regarding deflation.
- Add heavy debugging output to help track numerical issues (there are still fews)
- Make use of Eiegn's apply-inplane-rotation feature.
2014-09-22 15:22:52 +02:00
