devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • e24a1f57e3 [SYCL Function pointer Issue]: SYCL does not support function pointer inside the kernel, due to the portability issue of a function pointer and memory address space among host and accelerators. To fix the issue, function pointers have been replaced by function objects. mehdi-goli 2020-11-02 11:13:37 +00:00
  • 6961468915 Address issues with openglsupport test. Antonio Sanchez 2020-11-10 14:57:55 -08:00
  • 348a48682e Fix erroneous forward declaration of boost nvp. Everton Constantino 2020-11-02 18:50:17 +00:00
  • 3620371c5c Bug #2036 make sure find_standard_math_library_test_program actually compiles (and is guaranteed to call math functions) Christoph Hertzberg 2020-10-24 15:22:21 +02:00
  • 5dda502f84 Rename test/array.cpp to test/array_cwise.cpp David Tellenbach 2020-11-04 13:01:17 +01:00
  • 82fe059f35 Fix issue2045 which get a error case _mm256_set_m128d op not supported by gcc 7.x guoqiangqi 2020-11-04 09:21:39 +08:00
  • 9d11e2c03e CMakefile update for ROCm 4.0 Deven Desai 2020-10-29 17:00:04 +00:00
  • 39a038f2e4 Fix for ROCm (and CUDA?) breakage - 201029 Deven Desai 2020-10-29 15:34:05 +00:00
  • f895755c0e Remove unused functions in Half.h. David Tellenbach 2020-10-29 07:37:52 +01:00
  • 09f015852b Replace numext::as_uint with numext::bit_cast<numext::uint32_t> David Tellenbach 2020-10-29 07:28:28 +01:00
  • e265f7ed8e Add support for Armv8.2-a __fp16 David Tellenbach 2020-10-28 20:15:09 +00:00
  • 590aec8fab check two ctors Gael Guennebaud 2018-11-23 15:37:09 +01:00
  • a725a3233c [SYCL clean up the code] : removing exrta #pragma unroll in SYCL which was causing issues in embeded systems mehdi-goli 2020-10-27 16:31:33 +00:00
  • b9ff791fed [Missing SYCL math op]: Addin the missing LDEXP Function for SYCL. mehdi-goli 2020-10-27 16:26:38 +00:00
  • 61461d682a [Fixing expf issue]: Eigen uses the packet type operation for scaler type float on Sigmoid function(https://gitlab.com/libeigen/eigen/-/blob/master/Eigen/src/Core/functors/UnaryFunctors.h#L990). As a result SYCL backend breaks since SYCL backend only supports packet operation for vectorized type float4 and double2. The issue has been fixed by adding scalar type float to packet operation pexp for SYCL backend. mehdi-goli 2020-10-27 16:22:26 +00:00
  • ecb7bc9514 Bug #2036 make sure find_standard_math_library_test_program actually compiles (and is guaranteed to call math functions) Christoph Hertzberg 2020-10-24 15:22:21 +02:00
  • 09f595a269 Make sure compiler does not optimize away calls to math functions Susi Lehtola 2020-10-24 06:16:50 +00:00
  • 28aef8e816 Improve polynomial evaluation with instruction-level parallelism for pexp_float and pexp<Packet16f> guoqiangqi 2020-10-20 11:37:09 +08:00
  • 4a77eda1fd remove unnecessary specialize template of pexp for scale float/double guoqiangqi 2020-10-15 11:34:17 +08:00
  • d9f0d9eb76 Fix missing pfirst<Packet16b> for MSVC. Antonio Sanchez 2020-10-16 16:20:32 -07:00
  • 21edea5edd Fix the specialization of pfrexp for AVX to be faster when AVX2/AVX512DQ is not available, and avoid undefined behavior in C++. Also mask off the sign bit when extracting the exponent. Rasmus Munk Larsen 2020-10-15 18:39:58 -07:00
  • 011e0db31d Fix for ROCm/HIP breakage - 201013 Deven Desai 2020-10-14 00:53:20 +00:00
  • 6ea8091705 Revert change from 4e4d3f32d1 that broke BFloat16.h build with older compilers. Rasmus Munk Larsen 2020-10-15 01:20:08 +00:00
  • 4700713faf Add AVX plog<Packet4d> and AVX512 plog<Packet8d> ops,also unified AVX512 plog<Packet16f> op with generic api Guoqiang QI 2020-10-15 00:54:45 +00:00
  • af6f43d7ff Add specializations for pmin/pmax with prescribed NaN propagation semantics for SSE/AVX/AVX512. Rasmus Munk Larsen 2020-10-13 18:22:41 -07:00
  • 274ef12b61 Remove leftover debug print statement in cxx11_tensor_expr.cpp Rasmus Munk Larsen 2020-10-14 22:59:51 +00:00
  • 208b3626d1 Revert generic implementation of predux, since it break compilation of predux_any with MSVC. Rasmus Munk Larsen 2020-10-14 21:41:28 +00:00
  • e3e2cf9d24 Add MatrixBase::cwiseArg() David Tellenbach 2020-10-14 01:56:42 +00:00
  • 61fc78bbda Get rid of nested template specialization in TensorReductionGpu.h, which was broken by c6953f799b. Rasmus Munk Larsen 2020-10-13 23:53:11 +00:00
  • c6953f799b Add packet generic ops predux_fmin, predux_fmin_nan, predux_fmax, and predux_fmax_nan that implement reductions with PropagateNaN, and PropagateNumbers semantics. Add (slow) generic implementations for most reductions. Rasmus Munk Larsen 2020-10-13 21:48:31 +00:00
  • 807e51528d undefine EIGEN_CONSTEXPR before redefinition acxz 2020-10-12 20:24:15 -04:00
  • 9a4d04c05f Make bitwise_helper a device function to unbreak GPU builds. Rasmus Munk Larsen 2020-10-10 01:45:20 +00:00
  • 4e4d3f32d1 Clean up packetmath tests and fix various bugs to make bfloat16 pass (almost) all packetmath tests with SSE, AVX, and AVX512. Rasmus Munk Larsen 2020-10-09 20:05:49 +00:00
  • 7a8d3d5b81 Disable test exceptions when using OpenMP. David Tellenbach 2020-10-09 17:49:07 +02:00
  • 75f8b06e50 Mention problems when using potentially throwing scalars and OpenMP David Tellenbach 2020-10-09 17:04:25 +02:00
  • 9022f5aa8a Mention problems when using potentially throwing scalars and OpenMP David Tellenbach 2020-10-09 17:04:25 +02:00
  • e91e5d8c87 Fix typo in Tutorial_BlockOperations_block_assignment.cpp Karl Ljungkvist 2020-10-09 07:51:36 +00:00
  • d199c17b14 Fix typo in Tutorial_BlockOperations_block_assignment.cpp Karl Ljungkvist 2020-10-09 07:51:36 +00:00
  • ef3cc72cb6 Remove error counting in OpenMP parallelize_gemm Luke Peterson 2020-10-08 12:16:53 -07:00
  • 4091f6b25c Drop EIGEN_USING_STD_MATH in favour of EIGEN_USING_STD David Tellenbach 2020-10-09 02:05:05 +02:00
  • 7a0a2a5001 Define coeff-wise binary array operators for base class David Tellenbach 2020-10-09 00:53:34 +02:00
  • 183a208212 Implement generic bitwise logical packet ops that work for all types. Rasmus Munk Larsen 2020-10-08 22:45:20 +00:00
  • bfdd4a9903 Fix Paradiso. szczepaniak bartek 2020-10-08 19:38:35 +00:00
  • 8f8d77b516 Add EIGEN prefix for HAS_LGAMMA_R David Tellenbach 2020-10-08 18:32:19 +02:00
  • 2279f2c62f Use lgamma_r if it is available (update check for glibc 2.19+) Eugene Zhulenev 2020-10-08 00:26:45 +00:00
  • b431024404 Don't make assumptions about NaN-propagation for pmin/pmax - it various across platforms. Change test to only test for NaN-propagation for pfmin/pfmax. Rasmus Munk Larsen 2020-10-07 19:05:18 +00:00
  • dc252fbf00 Bump to 3.3.8 3.3.8 David Tellenbach 2020-10-05 13:24:15 +02:00
  • f66f3393e3 Use reinterpret_cast instead of C-style cast in Inverse_NEON.h David Tellenbach 2020-10-04 00:26:40 +02:00
  • 22c971a225 Don't cast away const in Inverse_NEON.h. Rasmus Munk Larsen 2020-10-02 15:06:34 -07:00
  • f93841b53e Use EIGEN_USING_STD to fix CUDA compilation error on BFloat16.h. Rasmus Munk Larsen 2020-10-02 14:47:15 -07:00
  • ee714f79f7 Fix CUDA build breakage and incorrect result for absdiff on HIP with long double arguments. Rasmus Munk Larsen 2020-10-02 21:05:35 +00:00
  • f7b185a8b1 dont use =* might not return a Scalar janos 2020-10-02 14:36:51 +02:00
  • 9f202c6f1e Fix undefined behaviour caused by uncaught exceptions in OMP section David Tellenbach 2020-10-02 00:41:01 +02:00
  • 9078f47cd6 Fix build breakage with MSVC 2019, which does not support MMX intrinsics for 64 bit builds, see: https://stackoverflow.com/questions/60933486/mmx-intrinsics-like-mm-cvtpd-pi32-not-found-with-msvc-2019-for-64bit-targets-c Rasmus Munk Larsen 2020-10-01 12:37:55 -07:00
  • 3b445d9bf2 Add a generic packet ops corresponding to {std}::fmin and {std}::fmax. The non-sensical NaN-propagation rules for std::min std::max implemented by pmin and pmax in Eigen is a longstanding source og confusion and bug report. This change is a first step towards addressing it, as discussing in issue #564. Rasmus Munk Larsen 2020-10-01 16:54:31 +00:00
  • 44b9d4e412 Specialize pldexp_double and pfdexp_double and get rid of Packet2l definition for SSE. SSE does not support conversion between 64 bit integers and double and the existing implementation of casting between Packet2d and Packer2l results in undefined behavior when casting NaN to int. Since pldexp and pfdexp only manipulate exponent fields that fit in 32 bit, this change provides specializations that use existing instructions _mm_cvtpd_pi32 and _mm_cvtsi32_pd instead. Rasmus Munk Larsen 2020-09-30 13:33:44 -07:00
  • d5a0d89491 Fix alignedbox 32-bit precision test failure. Antonio Sanchez 2020-09-30 08:21:20 -07:00
  • b933946d63 Fix failure in GEBP kernel when compiling with OpenMP and FMA David Tellenbach 2020-09-30 01:26:07 +02:00
  • 30960d485e Fix failure in GEBP kernel when compiling with OpenMP and FMA David Tellenbach 2020-09-30 01:26:07 +02:00
  • 1b080fdcb9 Make relative path variables of type STRING Alexander Grund 2020-09-28 16:39:48 +00:00
  • f9d1500f74 Revert !182. Rasmus Munk Larsen 2020-09-29 13:56:17 -07:00
  • 068121ec02 Add missing newline at the end of Inverse_NEON.h Rasmus Munk Larsen 2020-09-29 15:32:52 +00:00
  • 74ff5719b3 Fix compilation of 64 bit constant arguments to pset1frombits in TypeCasting.h on platforms where uint64_t != unsigned long. Rasmus Munk Larsen 2020-09-28 22:47:11 +00:00
  • 3a0b23e473 Fix compilation of pset1frombits calls on iOS. Rasmus Munk Larsen 2020-09-28 22:30:36 +00:00
  • 6b0c0b587e Provide a more efficient Packet2l->Packet2d cast method Christoph Hertzberg 2020-09-27 23:02:34 +02:00
  • 6425e875a1 Added AlignedBox::transform(AffineTransform). Martin Pecka 2020-09-28 18:06:23 +00:00
  • a967fadb21 Make relative path variables of type STRING Alexander Grund 2020-09-28 16:39:48 +00:00
  • e4b24e7fb2 Fix Eigen::ThreadPool::CurrentThreadId returning wrong thread id when EIGEN_AVOID_THREAD_LOCAL and NDEBUG are defined Zhuyie 2020-09-25 09:36:43 +00:00
  • ce5c59729d Fix for ROCm/HIP breakage - 200921 Deven Desai 2020-09-22 20:28:08 +00:00
  • b8a13f13ca Add CI configuration for ppc64le David Tellenbach 2020-09-22 00:26:23 +00:00
  • 821702e771 Fix the #issue1997 and #issue1991 bug triggered by unsupport a[index](type a: __i28d) ops with MSVC compiler Guoqiang QI 2020-09-21 15:49:00 +00:00
  • 493a7c773c Remove EIGEN_CONSTEXPR from NumTraits<boost::multiprecision::number<...>> David Tellenbach 2020-09-21 12:43:41 +02:00
  • 38e4a67394 Fix using FindStandardMathLibrary.cmake with -Wall (-Wunused-value) added to CMAKE_CXX_FLAG Павел Мацула 2020-09-19 16:13:16 +00:00
  • c4b99f78c7 Fix breakage in pcast<Packet2l, Packet2d> due to _mm_cvtsi128_si64 not being available on 32 bit x86. If SSE 4.1 is available use the faster _mm_extract_epi64 intrinsic. Rasmus Munk Larsen 2020-09-18 18:13:20 -07:00
  • 9aad16b443 Fix undefined reference to pset1frombits bug on different platforms guoqiangqi 2020-09-18 16:47:17 +08:00
  • c4aa8e0db2 Rename variable to avoid shadowing of a previously declared one David Tellenbach 2020-09-18 22:49:54 +02:00
  • e55182ac09 Get rid of initialization logic for blueNorm by making the computed constants static const or constexpr. Move macro definition EIGEN_CONSTEXPR to Core and make all methods in NumTraits constexpr when EIGEN_HASH_CONSTEXPR is 1. Rasmus Munk Larsen 2020-09-18 17:38:58 +00:00
  • 14022f5eb5 Fix more mildly embarrassing typos in ARM intrinsics in PacketMath.h. 'vmvnq_u64' does not exist for some reason. Rasmus Munk Larsen 2020-09-18 04:14:13 +00:00
  • a5b226920f Fix typo in PacketMath.h Rasmus Munk Larsen 2020-09-18 01:22:23 +00:00
  • 3af744b023 Add missing packet op pcmp_lt_or_nan for Packet2d on ARM. Rasmus Munk Larsen 2020-09-18 01:07:01 +00:00
  • 31a6b88ff3 Disable double version of compute_inverse_size4 on Inverse_NEON.h if Packet2d is not supported. Rasmus Munk Larsen 2020-09-17 23:51:06 +00:00
  • a796be81a4 relax number of iterations checks to avoid false negatives Gael Guennebaud 2018-10-15 10:23:32 +02:00
  • 880fa43b2b Add support for CastXML on ARM aarch64 Brad King 2020-09-16 13:37:27 -04:00
  • 6f0f6f792e Fix compiler error due to c++20 operator== generation rules daravi 2020-09-16 02:06:53 +00:00
  • cc0c38ace8 Remove old Clang compiler bug work-arounds. The two LLVM bugs referenced in the comments here have long been fixed. The workarounds were now detrimental because (1) they prevented using fused mul-add on Clang/ARM32 and (2) the unnecessary 'volatile' in 'asm volatile' prevented legitimate reordering by the compiler. Benoit Jacob 2020-09-15 11:07:57 -04:00
  • bb56a62582 Make bfloat16(float(-nan)) produce -nan, not nan. Tim Shen 2020-09-14 17:32:56 -07:00
  • 3012e755e9 Add plog ops support packet2d for NEON Guoqiang QI 2020-09-15 17:10:35 +00:00
  • e4fb0ddf78 Add EIGEN_UNUSED_VARIABLE to unused variable in Memory.h Rasmus Munk Larsen 2020-09-15 01:18:55 +00:00
  • 65e400896b Fix bfloat16 round on gcc 4.8 Pedro Caldeira 2020-09-11 09:14:33 -05:00
  • 76f0650563 Bump to 3.3.8-rc1 3.3.8-rc1 David Tellenbach 2020-09-14 13:43:55 +02:00
  • 5636f80d11 Fix issue #1968. Don't discard return value from "new" in C++17. Rasmus Munk Larsen 2020-09-11 18:17:08 -07:00
  • 7c5d48f313 Unified sse pldexp_double api Guoqiang QI 2020-09-12 10:56:55 +00:00
  • 71e08c702b Make blueNorm threadsafe if C++11 atomics are available. Rasmus Munk Larsen 2020-09-12 01:23:29 +00:00
  • adc861cabd New CI infrastructure, including AArch64 runners David Tellenbach 2020-09-11 18:11:49 +00:00
  • 5328c9be43 Fix half_impl::float_to_half_rtne(float) warning: '<<' causes overflow Niels Dekker 2020-09-10 16:22:28 +02:00
  • 35d149e34c Add missing functions for Packet8bf in Altivec architecture. Including new tests for bfloat16 Packets. Fix prsqrt on GenericPacketMath. Pedro Caldeira 2020-08-21 17:52:34 -05:00
  • 85428a3440 Add Neon psqrt<Packet2d> and pexp<Packet2d> Guoqiang QI 2020-09-08 09:04:03 +00:00
  • 5272106826 remove semi triggering -Wextra-semi-stmt Alexander Neumann 2020-09-07 11:42:30 +02:00
  • 4707c3aa86 Fix incorrect integer cast in predux<half2>(). nluehr 2017-11-21 10:47:00 -08:00
  • 5f25bcf7d6 Add Inverse_NEON.h Stephen Zheng 2020-09-04 10:55:47 +00:00