| [eigen-commits] commit/eigen: 85 new changesets |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen-commits Archives
]
85 new commits in eigen:
https://bitbucket.org/eigen/eigen/commits/d8652709345d/
Changeset: d8652709345d
User: benoitsteiner
Date: 2014-01-29 20:43:05
Summary: Added support for AVX to Eigen.
Affected #: 17 files
https://bitbucket.org/eigen/eigen/commits/399de6bd1e33/
Changeset: 399de6bd1e33
User: benoitsteiner
Date: 2014-02-19 03:06:44
Summary: Reverted the definition of the EIGEN_ALIGN to its former meaning (i.e. a boolean)
Created a new EIGEN_ALIGN_BYTES define to encode how the data should be aligned
Fixed a few remaining alignment issues exposed when the Eigen code is compiled with avx enabled.
Created a new EIGEN_ALIGN_DEFAULT define, which is set to the minimum alignment value required for the chosen instruction set. Use this value instead of EIGEN_ALIGN32 to preserve the existing alignment on SSE/Altivec/Neon.
Affected #: 13 files
https://bitbucket.org/eigen/eigen/commits/24b4dc92c6d3/
Changeset: 24b4dc92c6d3
User: benoitsteiner
Date: 2014-02-19 03:51:24
Summary: Merged the latest version of the code from eigen/eigen
Affected #: 37 files
https://bitbucket.org/eigen/eigen/commits/b64f87faf976/
Changeset: b64f87faf976
User: benoitsteiner
Date: 2014-02-24 22:54:07
Summary: Merged eigen/eigen into default
Affected #: 7 files
https://bitbucket.org/eigen/eigen/commits/997c2ef9fc8b/
Changeset: 997c2ef9fc8b
User: benoitsteiner
Date: 2014-02-24 22:45:32
Summary: Added support for FMA instructions
Affected #: 4 files
https://bitbucket.org/eigen/eigen/commits/dc71817e1afa/
Changeset: dc71817e1afa
User: benoitsteiner
Date: 2014-02-24 22:56:01
Summary: Pulled latest changes from the Eigen main trunk
Affected #: 37 files
https://bitbucket.org/eigen/eigen/commits/6c16eab171f9/
Changeset: 6c16eab171f9
User: benoitsteiner
Date: 2014-02-24 22:59:43
Summary: Merged latest changes from the main trunk
Affected #: 7 files
https://bitbucket.org/eigen/eigen/commits/b29a4a0fcc66/
Changeset: b29a4a0fcc66
User: benoitsteiner
Date: 2014-03-18 20:58:08
Summary: Merged latest changes from the parent
Affected #: 36 files
https://bitbucket.org/eigen/eigen/commits/ef2da9856146/
Changeset: ef2da9856146
User: benoitsteiner
Date: 2014-03-21 19:40:29
Summary: Made sure that EIGEN_ALIGN is defined when EIGEN_DONT_VECTORIZE is set to true to prevent build failures when vectorization is disabled.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/35c66599df0b/
Changeset: 35c66599df0b
User: benoitsteiner
Date: 2014-03-24 17:52:45
Summary: Added proper support for AVX and FMA in the makefiles.
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/08f2ed800a1e/
Changeset: 08f2ed800a1e
User: benoitsteiner
Date: 2014-03-24 21:33:40
Summary: Used AVX instructions to vectorize the predux_min<Packet8f>, predux_min<Packet4d>, predux_max<Packet8f>, and predux_max<Packet4d> packet primitives.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/34451426eb2c/
Changeset: 34451426eb2c
User: benoitsteiner
Date: 2014-03-25 17:00:43
Summary: Use AVX instructions to vectorize pset1<Packet2cd>, pset1<Packet4cf>, preverse<Packet2cd>, and preverse<Packet4cf>
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/a9f851a849d0/
Changeset: a9f851a849d0
User: benoitsteiner
Date: 2014-03-26 20:03:31
Summary: Used AVX instructions to vectorize the complex version of the pfirst and ploaddup packet primitives.
Silenced a few compilation warnings.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/e1eafd14eaa1/
Changeset: e1eafd14eaa1
User: benoitsteiner
Date: 2014-03-26 23:11:18
Summary: Vectorized the multiplication and division of complex numbers using AVX instructions.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/7d076a91bebd/
Changeset: 7d076a91bebd
User: benoitsteiner
Date: 2014-03-26 23:23:59
Summary: Merged latest updates from the parent branch
Affected #: 11 files
https://bitbucket.org/eigen/eigen/commits/579599cae16d/
Changeset: 579599cae16d
User: benoitsteiner
Date: 2014-03-26 23:58:13
Summary: Specialized the pload1 packet primitive for Packet8f and Packet4d in order to take advantage of the vbroadcastss and vbroadcastsd instructions whenever possible.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/46bae385c39e/
Changeset: 46bae385c39e
User: benoitsteiner
Date: 2014-03-27 01:35:18
Summary: Made sure that the version of gemm_pack_rhs specialized for row major matrices is vectorized when nr == 2*PacketSize (which is the case for SSE when compiling in 64bit mode).
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/f8ee3c721251/
Changeset: f8ee3c721251
User: benoitsteiner
Date: 2014-03-27 03:03:07
Summary: Created the ptranspose packet primitive that can transpose an array of N packets, where N is the number of words in each packet. This primitive will be used to complete the vectorization of the gemm_pack_lhs and gemm_pack_rhs functions.
Implemented the primitive using SSE instructions.
Affected #: 4 files
https://bitbucket.org/eigen/eigen/commits/041029795e16/
Changeset: 041029795e16
User: ggael
Date: 2014-03-27 14:47:00
Summary: Implement pcplflip, palign, predux and the likes from AVC/complexes
Affected #: 3 files
https://bitbucket.org/eigen/eigen/commits/3f517d698f50/
Changeset: 3f517d698f50
User: ggael
Date: 2014-03-27 15:29:56
Summary: Fix geo_* unit tests with respect to AVX
Affected #: 4 files
https://bitbucket.org/eigen/eigen/commits/bcd0938c48be/
Changeset: bcd0938c48be
User: ggael
Date: 2014-03-27 15:38:40
Summary: Change abi version when enabling AVX with GCC
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/9a812f3ddb68/
Changeset: 9a812f3ddb68
User: benoitsteiner
Date: 2014-03-27 17:34:51
Summary: Implemented the AVX version of the ptranspose packet primitive.
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/beac1d93fcc5/
Changeset: beac1d93fcc5
User: benoitsteiner
Date: 2014-03-27 18:02:24
Summary: Vectorized the packing of a row-major matrix used as the left hand side argument in a matrix-matrix product.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/e4946ef3960d/
Changeset: e4946ef3960d
User: benoitsteiner
Date: 2014-03-27 18:38:41
Summary: Vectorized the packing of a col-major matrix used as the right hand side argument in a matrix-matrix product when AVX instructions are used. No vectorization takes place when SSE instructions are used, however this doesn't seem to impact performance.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/e45243c0c8ba/
Changeset: e45243c0c8ba
User: benoitsteiner
Date: 2014-03-27 19:00:47
Summary: Silenced "unused variable" warnings when compiling with FMA.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/85d938c994fd/
Changeset: 85d938c994fd
User: benoitsteiner
Date: 2014-03-27 19:20:41
Summary: Fixed compilation error when FMA instructions are enabled.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/bb6b5be85379/
Changeset: bb6b5be85379
User: ggael
Date: 2014-03-27 23:38:50
Summary: enforce the use of vfmadd231ps for pmadd (gcc and clang stupidely generates the other fmadd variants plus some register moves...)
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/a8d27ff2df80/
Changeset: a8d27ff2df80
User: benoitsteiner
Date: 2014-03-28 00:03:03
Summary: Introduced pscatter/pgather packet primitives. They will be used to optimize the loop peeling code of the block-panel matrix multiplication kernel.
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/93dfaf66e00f/
Changeset: 93dfaf66e00f
User: benoitsteiner
Date: 2014-03-28 01:42:25
Summary: Implemented the AVX version of the gather and scatter packet primitives.
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/b95b1a9748cd/
Changeset: b95b1a9748cd
User: benoitsteiner
Date: 2014-03-28 02:29:01
Summary: Implemented the SSE version of the gather and scatter packet primitives.
Affected #: 4 files
https://bitbucket.org/eigen/eigen/commits/ca808bb456b0/
Changeset: ca808bb456b0
User: benoitsteiner
Date: 2014-03-28 02:32:15
Summary: Merged latest changes from parent.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/d7473a445544/
Changeset: d7473a445544
User: ggael
Date: 2014-03-28 09:11:06
Summary: Enable vectorization of gemv for PacketSize>4 through unaligned loads (still better than no vectorization)
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/11ea98087b56/
Changeset: 11ea98087b56
User: ggael
Date: 2014-03-28 09:24:18
Summary: merge with default branch
Affected #: 6 files
https://bitbucket.org/eigen/eigen/commits/a08b9f90102a/
Changeset: a08b9f90102a
User: ggael
Date: 2014-03-28 10:18:04
Summary: Add a mechanism to recursively access to half-size packet types
Affected #: 9 files
https://bitbucket.org/eigen/eigen/commits/9a29eb7d5ffd/
Changeset: 9a29eb7d5ffd
User: benoitsteiner
Date: 2014-03-28 20:00:08
Summary: Properly align the input data to prevent false failures of the packetmath.cpp test.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/39593eacedb8/
Changeset: 39593eacedb8
User: benoitsteiner
Date: 2014-03-28 20:11:23
Summary: Vectorized the loop peeling of the inner loop of the block-panel matrix multiplication code. This speeds up the multiplication of matrices which size is not a multiple of the packet size.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/8b1001f9e3ac/
Changeset: 8b1001f9e3ac
User: ggael
Date: 2014-03-30 21:57:05
Summary: Optimize gebp kernel:
1 - increase peeling level along the depth dimention (+5% for large matrices, i.e., >1000)
2 - improve pipelining when dealing with latest rows of the lhs
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/a060ac3bf4d3/
Changeset: a060ac3bf4d3
User: ggael
Date: 2014-03-30 22:43:47
Summary: Workaround alignment warnings
Affected #: 3 files
https://bitbucket.org/eigen/eigen/commits/21e0f00bb7d0/
Changeset: 21e0f00bb7d0
User: ggael
Date: 2014-03-31 10:41:40
Summary: Enable repetition in mixing type unit test
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/43417425d7fe/
Changeset: 43417425d7fe
User: ggael
Date: 2014-03-31 10:42:19
Summary: Finally, prefetching seems to help getting more stable performance
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/d7720f0d7f5b/
Changeset: d7720f0d7f5b
User: ggael
Date: 2014-03-31 10:58:30
Summary: BTL: fix warnings and extend to 5k matrices, update GotoBlas to OpenBlas, etc.
Affected #: 19 files
https://bitbucket.org/eigen/eigen/commits/86390b882145/
Changeset: 86390b882145
User: ggael
Date: 2014-03-31 10:59:55
Summary: BTL: add blaze
Affected #: 4 files
https://bitbucket.org/eigen/eigen/commits/3021ca643e4a/
Changeset: 3021ca643e4a
User: benoitsteiner
Date: 2014-04-02 07:07:05
Summary: Pulled the latest updates from the eigen trunk.
Affected #: 15 files
https://bitbucket.org/eigen/eigen/commits/9c021a628489/
Changeset: 9c021a628489
User: benoitsteiner
Date: 2014-04-04 23:12:24
Summary: Deleted some dead code.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/500a43111423/
Changeset: 500a43111423
User: benoitsteiner
Date: 2014-04-05 02:08:47
Summary: Updated the geo_parametrizedline_2 test for AVX.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/8e1d7973fdb1/
Changeset: 8e1d7973fdb1
User: benoitsteiner
Date: 2014-04-14 19:44:17
Summary: Optimized SSE unaligned loads and stores when compiling a 64bit target with a recent version of gcc (ie gcc 4.8).
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/5a4ca1ad8c53/
Changeset: 5a4ca1ad8c53
User: ggael
Date: 2014-04-16 17:05:11
Summary: New gebp kernel handling up to 3 packets x 4 register-level blocks. Huge speeup on Haswell.
This changeset also introduce new vector functions: ploadquad and predux4.
Affected #: 12 files
https://bitbucket.org/eigen/eigen/commits/95f2d2d3cc88/
Changeset: 95f2d2d3cc88
User: ggael
Date: 2014-04-16 23:14:58
Summary: Enable alloca on MAC OSX
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/2dd06ce4a737/
Changeset: 2dd06ce4a737
User: ggael
Date: 2014-04-16 23:15:42
Summary: Fallback to lazy products for very small ones.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/f85b0373bbed/
Changeset: f85b0373bbed
User: ggael
Date: 2014-04-17 16:02:27
Summary: Optimize ploaddup for AVX
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/5ad5255ea252/
Changeset: 5ad5255ea252
User: ggael
Date: 2014-04-17 16:04:30
Summary: Fix and optimize mixed products
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/12bda05bd71e/
Changeset: 12bda05bd71e
User: ggael
Date: 2014-04-17 16:26:35
Summary: Extend mixingtype unit test to check transposed cases.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/fd4b16d27189/
Changeset: fd4b16d27189
User: ggael
Date: 2014-04-17 16:27:22
Summary: add unit tests for ploadquad and predux4, and split packetmath unit test wrt real/complex
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/95bd5cc424a1/
Changeset: 95bd5cc424a1
User: ggael
Date: 2014-04-17 16:27:58
Summary: Reduce block sizes in unit tests.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/9e3787a8bb8f/
Changeset: 9e3787a8bb8f
User: ggael
Date: 2014-04-17 20:51:04
Summary: Optimize AVX pset1 for complexes and ploaddup
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/27193eecd032/
Changeset: 27193eecd032
User: ggael
Date: 2014-04-17 21:01:45
Summary: Various minor fixes in BTL
Affected #: 9 files
https://bitbucket.org/eigen/eigen/commits/6e35ed5c26de/
Changeset: 6e35ed5c26de
User: ggael
Date: 2014-04-17 21:03:26
Summary: Make our gemm bench a little more powerful.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/56d978fb7e4a/
Changeset: 56d978fb7e4a
User: benoitsteiner
Date: 2014-04-17 21:28:01
Summary: Implemented the pgather/pscatter packet primitives for the arm/NEON architecture
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/f3488f4e45b2/
Changeset: f3488f4e45b2
User: ggael
Date: 2014-04-17 22:50:49
Summary: merge
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/67e54d295a2a/
Changeset: 67e54d295a2a
User: ggael
Date: 2014-04-17 23:26:34
Summary: Fixes for fixed sizes and non vectorizable types
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/970dbd9a054f/
Changeset: 970dbd9a054f
User: ggael
Date: 2014-04-18 11:43:13
Summary: Fix typo (was working with clang\!)
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/bb9fc0721496/
Changeset: bb9fc0721496
User: ggael
Date: 2014-04-18 15:35:34
Summary: Smarter block size computation
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/f8c8eba11211/
Changeset: f8c8eba11211
User: ggael
Date: 2014-04-18 17:06:03
Summary: Fix calls to lazy products (lazy product does not like matrices with 0 length)
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/9d05fce38984/
Changeset: 9d05fce38984
User: ggael
Date: 2014-04-18 17:06:31
Summary: Fix alignment assertion.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/466992f93aba/
Changeset: 466992f93aba
User: ggael
Date: 2014-04-18 21:14:40
Summary: Fix 128bit packet size assumptions in unit tests.
Affected #: 5 files
https://bitbucket.org/eigen/eigen/commits/5b40cf467280/
Changeset: 5b40cf467280
User: ggael
Date: 2014-04-22 16:03:19
Summary: Workaround gcc's default ABI not being able to distinghish between vector types of different sizes.
Affected #: 4 files
https://bitbucket.org/eigen/eigen/commits/45b435601179/
Changeset: 45b435601179
User: ggael
Date: 2014-04-22 17:00:38
Summary: merge with default branch
Affected #: 19 files
https://bitbucket.org/eigen/eigen/commits/e036b77e4c25/
Changeset: e036b77e4c25
User: ggael
Date: 2014-04-22 17:03:57
Summary: Fix EIGEN_MAKE_UNALIGNED_ARRAY_ASSERT macro
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/66c7faae20df/
Changeset: 66c7faae20df
User: ggael
Date: 2014-04-23 13:09:26
Summary: Add Altivec implementation of pgather/pscatter (not tested)
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/e96c884af896/
Changeset: e96c884af896
User: benoitsteiner
Date: 2014-04-24 03:22:10
Summary: Created a NEON version of the ptranspose packet primitives
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/7717130381cd/
Changeset: 7717130381cd
User: benoitsteiner
Date: 2014-04-24 03:23:07
Summary: Fixed the NEON implementation of predux_max<Packet4i>.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/663a49ea2648/
Changeset: 663a49ea2648
User: ggael
Date: 2014-04-24 14:47:53
Summary: Implement ptranspose on altivec and fix pgather/pscatter
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/fdeb126954b9/
Changeset: fdeb126954b9
User: ggael
Date: 2014-04-24 23:17:18
Summary: Enable fused madd for Altivec
Affected #: 4 files
https://bitbucket.org/eigen/eigen/commits/a43fac4ae421/
Changeset: a43fac4ae421
User: ggael
Date: 2014-04-25 10:56:18
Summary: Enable vectorization of pack_rhs with a column-major RHS.
Rename and generalize Kernel<*> to PacketBlock<*,N>.
Affected #: 11 files
https://bitbucket.org/eigen/eigen/commits/1713099c5de7/
Changeset: 1713099c5de7
User: ggael
Date: 2014-04-25 11:04:02
Summary: Avoid blocking-size mismatch in unit tests calling Eigen's blas interface.
Affected #: 6 files
https://bitbucket.org/eigen/eigen/commits/9f1c447ac813/
Changeset: 9f1c447ac813
User: ggael
Date: 2014-04-25 11:06:03
Summary: Minor optimizations in product kernel:
- use pbroadcast4 (helpful when AVX is not available)
- process all remaining rows at once (significant speedup for small matrices)
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/34514f9e86a6/
Changeset: 34514f9e86a6
User: ggael
Date: 2014-04-25 11:15:13
Summary: Fix ptranspose overload prototypes for NEON
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/98d43597bfd6/
Changeset: 98d43597bfd6
User: ggael
Date: 2014-04-25 11:21:18
Summary: Add unit test for pbroadcast4/2
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/e09031dccfd9/
Changeset: e09031dccfd9
User: ggael
Date: 2014-04-25 11:48:22
Summary: Disable 3pX4 kernel on Altivec: despite this platform has 32 registers, this version seems significantly slower.
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/ea095da2cda6/
Changeset: ea095da2cda6
User: ggael
Date: 2014-04-25 11:46:22
Summary: pbroadcast4/2 assume aligned memory
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/92f6fcf45f2f/
Changeset: 92f6fcf45f2f
User: ggael
Date: 2014-04-25 11:46:57
Summary: Implement pbroadcast4 on altivec
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/783456c2d916/
Changeset: 783456c2d916
User: ggael
Date: 2014-04-25 13:22:34
Summary: Fix for mixed products
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/bb5656aff10d/
Changeset: bb5656aff10d
User: ggael
Date: 2014-04-25 14:05:54
Summary: Fix sizeof unit test
Affected #: 2 files
https://bitbucket.org/eigen/eigen/commits/c9afd13d4de9/
Changeset: c9afd13d4de9
User: ggael
Date: 2014-04-25 16:54:30
Summary: Product kernel: skip loop on columns if there is no remaining rows
Affected #: 1 file
https://bitbucket.org/eigen/eigen/commits/2170940e93fc/
Changeset: 2170940e93fc
User: ggael
Date: 2014-04-25 22:25:48
Summary: Make sure that calls to broadcast4 are 16 bytes aligned
Affected #: 2 files
Repository URL: https://bitbucket.org/eigen/eigen/
--
This is a commit notification from bitbucket.org. You are receiving
this because you have the service enabled, addressing the recipient of
this email.