mirror/eigen
2
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-01-06 14:14:46 +08:00
Code Issues Packages Projects Releases Wiki Activity
8,481 Commits 11 Branches 89 Tags 168 MiB
0f56b5a6de
Commit Graph

4890 Commits

Author SHA1 Message Date
Gael Guennebaud
72744d93ef Allows the compiler to inline outer products (the change from default to dont-inline in changeset 737bed19c1 
was not motivated)
 2016-07-22 17:02:28 +02:00
Gael Guennebaud
395c835f4b Fix CUDA compilation 2016-07-22 15:30:24 +02:00
Gael Guennebaud
47afc9a365 More cleaning in half: 
- put its definition and functions in its own half_impl namespace such that the free function does not polute the Eigen namespace while still making them visible for half through ADL.
 - expose Eigen::half throguh a using statement
 - move operator<< from std to half_float namespace
 2016-07-22 14:33:28 +02:00
Gael Guennebaud
0f350a8b7e Fix CUDA compilation 2016-07-21 18:47:07 +02:00
Gael Guennebaud
bf91a44f4a Use ADL and log10 for printing matrices. 2016-07-21 15:48:24 +02:00
Gael Guennebaud
87fbda812f Add missing log10 and random generator for half. 2016-07-21 15:46:45 +02:00
Gael Guennebaud
01d12d3e82 Some cleanup in Halh: standard functions should be defined in the namespace of the class half to make ADL work, and thus the global is* functions can be removed. 2016-07-21 15:10:48 +02:00
Gael Guennebaud
7722913475 Fix ambiguous specialization with custom scalar type 2016-07-20 15:13:44 +02:00
Gael Guennebaud
fd057f86b3 Complete the coeff-wise math function table. 2016-07-20 12:14:10 +02:00
Gael Guennebaud
9e8476ef22 Add missing Eigen::rsqrt global function 2016-07-20 11:59:49 +02:00
Gael Guennebaud
4b4c296d6e Simplify ScalarBinaryOpTraits by removing the Defined enum, and extend its documentation. 2016-07-20 09:56:39 +02:00
Gael Guennebaud
e3bf874c83 Workaround MSVC 2010 compilation issue. 2016-07-18 15:17:25 +02:00
Gael Guennebaud
0f89c6d6b5 Add a summary of possible values for EIGEN_COMP_MSVC 2016-07-18 15:16:13 +02:00
Gael Guennebaud
18884f17d7 Remove static constant declaration: this enforces compiler to generate costly code for thread safety. 2016-07-18 15:05:17 +02:00
Gael Guennebaud
79574e384e Make scalar_product_op the default (instead of void) 2016-07-18 12:03:05 +02:00
Gael Guennebaud
6a3c451c1c Permits call to explicit ctor. 2016-07-18 12:02:20 +02:00
Gael Guennebaud
0c3fe4aca5 merge 2016-07-18 10:44:15 +02:00
Gael Guennebaud
db9b154193 Add missing non-const reverse method in VectorwiseOp. 2016-07-16 15:19:28 +02:00
Gael Guennebaud
461cd819c2 Workaround VS2015 bug 2016-07-13 18:46:01 +02:00
Gael Guennebaud
5ea0864c81 Fix regression in a previous commit: some diagonal entry might not be treated by the 2x2 real preconditioner. 2016-07-13 18:37:54 +02:00
Gael Guennebaud
b4343aa67e Avoid division by very small entries when extracting singularvalues, and explicitly handle the 1x1 complex case. 2016-07-12 17:22:03 +02:00
Gael Guennebaud
e2aa58b631 Consider denormals as zero in makeJacobi and 2x2 SVD. 
This also fix serious issues with x387 for which values can be much smaller than the smallest denormal!
 2016-07-12 17:21:03 +02:00
klimpel
8b3fc31b55 compile fix (SFINAE variant apparently didn't work for all compilers) for the following compiler/platform: 
gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-46)
Copyright (C) 2006 Free Software Foundation, Inc.
 2016-07-11 17:42:22 +02:00
Gael Guennebaud
a96a7ce3f7 Move CUDA's special functions to SpecialFunctions module. 2016-07-11 18:39:11 +02:00
Gael Guennebaud
fd60966310 merge 2016-07-11 18:11:47 +02:00
Gael Guennebaud
3e348fdcf9 Workaround MSVC bug 2016-07-11 15:24:52 +02:00
Konstantinos Margaritis
ef05463fcf Merged kmargar/eigen/tip into default, Altivec/VSX port should be working ok now. 2016-07-10 16:11:46 +03:00
Konstantinos Margaritis
9f7caa7e7d minor fixes for big endian altivec/vsx 2016-07-10 07:05:10 -03:00
Christoph Hertzberg
3c795c6923 bug #1119: Adjust call to ?gssvx for SuperLU 5 
Also improved corresponding cmake module to detect versions 5.x

Based on patch by Christoph Grüninger.
 2016-07-10 02:29:57 +02:00
Gael Guennebaud
2f7e2614e7 bug #1232: refactor special functions as a new SpecialFunctions module, currently in unsupported/. 2016-07-08 11:13:55 +02:00
Gael Guennebaud
66917299a9 Add debug output 2016-07-06 22:27:15 +02:00
Gael Guennebaud
c3b23d7dbf Fix support of Intel's VML 2016-07-06 14:07:32 +02:00
Gael Guennebaud
8ec4d6480d Fix compilation with recent updates of icc 2016 2016-07-06 14:07:14 +02:00
Gael Guennebaud
5b3a6f51d3 Improve numerical robustness of RealSchur: add scaling and compare sub-diag entries to largest diagonal entry instead of the 2 neighbors. 2016-07-06 13:45:30 +02:00
Gael Guennebaud
367ef66af3 Re-enable some specializations for Assignment<.,Product<>> 2016-07-05 22:58:14 +02:00
Gael Guennebaud
155d8d8603 Fix compilation with msvc 2016-07-05 14:43:42 +02:00
Gael Guennebaud
b39fd8217f Fix nesting of SolveWithGuess, and add unit test. 2016-07-04 17:47:47 +02:00
Gael Guennebaud
ec02af1047 Fix template resolution. 2016-07-04 17:37:33 +02:00
Gael Guennebaud
fbcfc2f862 Add unit test for solveWithGuess, and fix template resolution. 2016-07-04 17:19:38 +02:00
Gael Guennebaud
7f7839c12f Add documentation and exemples for inplace decomposition. 2016-07-04 17:18:26 +02:00
Gael Guennebaud
32a41ee659 bug #707: add inplace decomposition through Ref<> for Cholesky, LU and QR decompositions. 2016-07-04 15:13:35 +02:00
Gael Guennebaud
91b3039013 Change the semantic of the last template parameter of Assignment from "Scalar" to "SFINAE" only. 
The previous "Scalar" semantic was obsolete since we allow for different scalar types in the source and destination expressions.
On can still specialize on scalar types through SFINAE and/or assignment functor.
 2016-07-04 11:02:00 +02:00
Gael Guennebaud
0fa9e4a15c Fix performance regression in dgemm introduced by changeset 5d51a7f12c 2016-07-02 17:35:08 +02:00
Gael Guennebaud
672076db5d Fix performance regression introduced in changeset e56aabf205 
.
Register blocking sizes are better handled by the cache size heuristics.
The current code introduced very small blocks, for instance for 9x9 matrix,
thus killing performance.
 2016-07-02 15:40:56 +02:00
Justin Carpentier
6126886a67 Use complete nested namespace Eigen::internal 2016-06-28 20:09:25 +02:00
Benoit Jacob
328c5d876a Undo changes in AltiVec --- I don't have any way to test there. 2016-06-28 11:15:25 -04:00
Benoit Jacob
38fb606052 Avoid global variables with static constructors in NEON/Complex.h 2016-06-28 11:12:49 -04:00
Gael Guennebaud
d476cadbb8 bug #1247: fix regression in compilation of pow(integer,integer), and add respective unit tests. 2016-06-25 10:12:06 +02:00
Gael Guennebaud
c50c73cae2 Fix missing specialization. 2016-06-24 23:10:39 +02:00
Gael Guennebaud
cd577a275c Relax promote_scalar_arg logic to enable promotion to Expr::Scalar if conversion to Expr::Literal fails. 
This is useful to cancel expression template at the scalar level, e.g. with AutoDiff<AutoDiff<>>.
This patch also defers calls to NumTraits in cases for which types are not directly compatible.
 2016-06-24 11:28:54 +02:00
Gael Guennebaud
deb45ad4bc bug #1245: fix compilation with msvc 2016-06-24 09:52:25 +02:00
Gael Guennebaud
55fc04e8b5 Fix operator priority 2016-06-23 15:36:42 +02:00
Gael Guennebaud
bf2d5edecc Fix warning. 2016-06-23 15:35:17 +02:00
Gael Guennebaud
7c6561485a merge PR 194 2016-06-23 15:29:57 +02:00
Konstantinos Margaritis
be107e387b fix compilation with clang 3.9, fix performance with pset1, use vector operators instead of intrinsics in some cases 2016-06-23 10:19:05 -03:00
Gael Guennebaud
76faf4a965 Introduce a NumTraits<T>::Literal type to be used for literals, and 
improve mixing type support in operations between arrays and scalars:
 - 2 * ArrayXcf is now optimized in the sense that the integer 2 is properly promoted to a float instead of a complex<float> (fix a regression)
 - 2.1 * ArrayXi is now forbiden (previously, 2.1 was converted to 2)
 - This mechanism should be applicable to any custom scalar type, assuming NumTraits<T>::Literal is properly defined (it defaults to T)
 2016-06-23 14:27:20 +02:00
Gael Guennebaud
a3f7edf7e7 Biug 1242: fix comma init with empty matrices. 2016-06-23 10:25:04 +02:00
Konstantinos Margaritis
8c34b5a0e3 mostly cleanups and modernizing code 2016-06-19 16:13:17 -03:00
Konstantinos Margaritis
b410d46482 mostly cleanups and modernizing code 2016-06-19 16:12:52 -03:00
Konstantinos Margaritis
b80379bda0 fixed pexp<Packet2d>, was failing tests 2016-06-19 16:11:58 -03:00
Benoit Steiner
b055590e91 Made log1p_impl usable inside a GPU kernel 2016-06-16 11:37:40 -07:00
Gael Guennebaud
67c12531e5 Fix warnings with gcc 2016-06-15 18:11:33 +02:00
Gael Guennebaud
eb91345d64 Move scalar/expr to ArrayBase and fix documentation 2016-06-15 15:22:03 +02:00
Gael Guennebaud
4794834397 Propagate functor to ScalarBinaryOpTraits 2016-06-15 09:58:49 +02:00
Gael Guennebaud
c55035b9c0 Include the cost of stores in unrolling of triangular expressions. 2016-06-15 09:57:33 +02:00
Gael Guennebaud
4e7c3af874 Cleanup useless helper: internal::product_result_scalar 2016-06-15 00:04:10 +02:00
Gael Guennebaud
101ea26f5e Include the cost of stores in unrolling (also fix infinite unrolling with expression costing 0 like Constant) 2016-06-15 00:01:16 +02:00
Gael Guennebaud
76236cdea4 merge 2016-06-14 15:33:47 +02:00
Gael Guennebaud
1004c4df99 Cleanup unused functors. 2016-06-14 15:27:28 +02:00
Gael Guennebaud
70dad84b73 Generalize expr/expr and scalar/expr wrt scalar types. 2016-06-14 15:26:37 +02:00
Gael Guennebaud
62134082aa Update AutoDiffScalar wrt to scalar-multiple. 2016-06-14 15:06:35 +02:00
Gael Guennebaud
396d9cfb6e Generalize expr.pow(scalar), pow(expr,scalar) and pow(scalar,expr). 
Internal: scalar_pow_op (unary) is removed, and scalar_binary_pow_op is renamed scalar_pow_op.
 2016-06-14 14:10:07 +02:00
Gael Guennebaud
a8c08e8b8e Implement expr+scalar, scalar+expr, expr-scalar, and scalar-expr as binary expressions, and generalize supported scalar types. 
The following functors are now deprecated: scalar_add_op, scalar_sub_op, and scalar_rsub_op.
 2016-06-14 12:06:10 +02:00
Gael Guennebaud
756ac4a93d Fix doc. 2016-06-14 12:03:39 +02:00
Gael Guennebaud
bcc0f38f98 Add unittesting plugins to scalar_product_op and scalar_quotient_op to help chaking that types are properly propagated. 2016-06-14 11:31:27 +02:00
Gael Guennebaud
f57fd78e30 Generalize coeff-wise sparse products to support different scalar types 2016-06-14 11:29:54 +02:00
Gael Guennebaud
f5b1c73945 Set cost of constant expression to 0 (the cost should be amortized through the expression) 2016-06-14 11:29:06 +02:00
Gael Guennebaud
deb8306e60 Move MatrixBase::operaotr*(UniformScaling) as a free function in Scaling.h, and fix return type. 2016-06-14 11:28:03 +02:00
Gael Guennebaud
64fcfd314f Implement scalar multiples and division by a scalar as a binary-expression with a constant expression. 
This slightly complexifies the type of the expressions and implies that we now have to distinguish between scalar*expr and expr*scalar to catch scalar-multiple expression (e.g., see BlasUtil.h), but this brings several advantages:
- it makes it clear on each side the scalar is applied,
- it clearly reflects that we are dealing with a binary-expression,
- the complexity of the type is hidden through macros defined at the end of Macros.h,
- distinguishing between "scalar op expr" and "expr op scalar" is important to support non commutative fields (like quaternions)
- "scalar op expr" is now fully equivalent to "ConstantExpr(scalar) op expr"
- scalar_multiple_op, scalar_quotient1_op and scalar_quotient2_op are not used anymore in officially supported modules (still used in Tensor)
 2016-06-14 11:26:57 +02:00
Gael Guennebaud
3c12e24164 Add bind1st_op and bind2nd_op helpers to turn binary functors into unary ones, and implement scalar_multiple2 and scalar_quotient2 on top of them. 2016-06-13 16:18:59 +02:00
Gael Guennebaud
7a9ef7bbb4 Add default template parameters for the second scalar type of binary functors. 
This enhences backward compatibility.
 2016-06-13 16:17:23 +02:00
Gael Guennebaud
4c61f00838 Add missing explicit scalar conversion 2016-06-12 22:42:13 +02:00
Gael Guennebaud
83904a21c1 Make sure T(i+1,i)==0 when diagonalizing T(i:i+1,i:i+1) 2016-06-11 14:41:36 +02:00
Gael Guennebaud
fabae6c9a1 Cleanup 2016-06-10 15:58:33 +02:00
Gael Guennebaud
5fdd703629 Enable mixing types in numext::pow 2016-06-10 15:58:04 +02:00
Gael Guennebaud
2e238bafb6 Big 279: enable mixing types for comparisons, min, and max. 2016-06-10 15:05:43 +02:00
Gael Guennebaud
0028049380 bug #1240: Remove any assumption on NEON vector types. 2016-06-09 23:08:11 +02:00
Gael Guennebaud
2c462f4201 Clean handling for void type in EIGEN_CHECK_BINARY_COMPATIBILIY 2016-06-06 23:11:38 +02:00
Gael Guennebaud
3d71d3918e Disable shortcuts for res ?= prod when the scalar types do not match exactly. 2016-06-06 23:10:55 +02:00
Gael Guennebaud
66e99ab6a1 Relax mixing-type constraints for binary coefficient-wise operators: 
- Replace internal::scalar_product_traits<A,B> by Eigen::ScalarBinaryOpTraits<A,B,OP>
- Remove the "functor_is_product_like" helper (was pretty ugly)
- Currently, OP is not used, but it is available to the user for fine grained tuning
- Currently, only the following operators have been generalized: *,/,+,-,=,*=,/=,+=,-=
- TODO: generalize all other binray operators (comparisons,pow,etc.)
- TODO: handle "scalar op array" operators (currently only * is handled)
- TODO: move the handling of the "void" scalar type to ScalarBinaryOpTraits
 2016-06-06 15:11:41 +02:00
Benoit Steiner
1f1e0b9e30 Silenced compilation warning 2016-06-05 12:59:11 -07:00
Benoit Steiner
5b95b4daf9 Moved static assertions into the class constructor to make the code more portable 2016-06-05 12:57:48 -07:00
Sean Templeton
bd21243821 Fix compile errors initializing packets on ARM DS-5 5.20 
The ARM DS-5 5.20 compiler fails compiling with the following errors:

"src/Core/arch/NEON/PacketMath.h", line 113: Error:  #146: too many initializer values
    Packet4f countdown = EIGEN_INIT_NEON_PACKET4(0, 1, 2, 3);
                         ^
"src/Core/arch/NEON/PacketMath.h", line 118: Error:  #146: too many initializer values
    Packet4i countdown = EIGEN_INIT_NEON_PACKET4(0, 1, 2, 3);
                         ^
"src/Core/arch/NEON/Complex.h", line 30: Error:  #146: too many initializer values
  static uint32x4_t p4ui_CONJ_XOR = EIGEN_INIT_NEON_PACKET4(0x00000000, 0x80000000, 0x00000000, 0x80000000);
                                    ^
"src/Core/arch/NEON/Complex.h", line 31: Error:  #146: too many initializer values
  static uint32x2_t p2ui_CONJ_XOR = EIGEN_INIT_NEON_PACKET2(0x00000000, 0x80000000);
                                    ^

The vectors are implemented as two doubles, hence the too many initializer values error.
Changed the code to use intrinsic load functions which all compilers
implementing NEON should have.
 2016-06-03 10:51:35 -05:00
Gael Guennebaud
1fc2746417 Make Arrays's ctor/assignment noexcept 2016-06-09 22:52:37 +02:00
Gael Guennebaud
2bd59b0e0d Take advantage that T is already diagonal in the extraction of generalized complex eigenvalues. 2016-06-09 17:12:03 +02:00
Gael Guennebaud
c1f9ca9254 Update RealQZ to reduce 2x2 diagonal block of T corresponding to non reduced diagonal block of S to positive diagonal form. 
This step involve a real 2x2 SVD problem. The respective routine is thus in src/misc/ to be shared by both EVD and AVD modules.
 2016-06-09 17:11:03 +02:00
Gael Guennebaud
a20d2ec1c0 Fix shadow variable, and indexing. 2016-06-09 16:16:22 +02:00
Abhijit Kundu
0beabb4776 Fixed type conversion from int 2016-06-08 16:12:04 -04:00
Gael Guennebaud
df095cab10 Fixes for PARDISO: warnings, and defaults to metis+ in-core mode. 2016-06-08 18:31:19 +02:00
Gael Guennebaud
9fc8379328 Fix extraction of complex eigenvalue pairs in real generalized eigenvalue problems. 2016-06-08 16:39:11 +02:00