[ Thread Index | More lists.tuxfamily.org/eigen-commits Archives ]

156 message(s), the most recent arriving on Wed 01 June at 01:47 GMT.

• [eigen-commits] commit/eigen: ggael: Fix performance regression: with AVX, unaligned stores were emitted instead of aligned ones for fixed size assignement. Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: Made a cast explicit Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Added a test to validate the computation of exp and log on 16bit floats Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1214: consider denormals as zero in D&C SVD. This also workaround infinite binary search when compiling with ICC's unsafe optimizations. Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: Use numext::isfinite instead of std::isfinite Bitbucket
• [eigen-commits] commit/eigen: ggael: Improve documentation of BDCSVD Bitbucket
• [eigen-commits] commit/eigen: Ola Røer Thorsen: fix double-promotion/float-conversion in Core/SpecialFunctions.h Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Removed extraneous 'explicit' keywords Bitbucket

• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: 3 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed some incorrect assertions Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Added a benchmark to measure the performance of full reductions of 16 bit floats Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Relaxed an assertion that was tighter that necessary. Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: Avoid double promotion Bitbucket

• [eigen-commits] commit/eigen: 3 new changesets Bitbucket
• [eigen-commits] commit/eigen: 3 new changesets Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: Optimized the non blocking thread pool: Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Added a new parallelFor api to the thread pool device. Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: Added the ability to use a scratch buffer in cuda kernels Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Improved the performance of full reductions on GPU: Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed compilation warning Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Simplified the reduction code a little. Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Small improvement to the full reduction of fp16 Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Extended the tests for ptanh Bitbucket

• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: chtz: Workaround maybe-uninitialized warning Bitbucket
• [eigen-commits] commit/eigen: chtz: Closing branch: default Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Explicitely initialize all the atomic variables. Bitbucket
• [eigen-commits] commit/eigen: 3 new changesets Bitbucket
• [eigen-commits] commit/eigen: chtz: Bug 1213: Give names to anonymous enums Bitbucket
• [eigen-commits] commit/eigen: 4 new changesets Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: chtz: Do not copy for trivial 1x1 case. This also avoids a "maybe-uninitialized" warning in some situations. Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: Made predux_min and predux_max on fp16 less noisy Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Misc fixes for fp16 Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Avoid unnecessary conversions between floats and doubles Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed a couple of bugs related to the Pascalfamily of GPUs Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Improved the portability of the tensor code Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Worked around compilation errors with older versions of gcc Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Replace implicit cast with an explicit one Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed potential race condition in the non blocking thread pool Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Worked around a compilation error triggered by nvcc when compiling a tensor concatenation kernel. Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: Fixed compilation errors triggered by old versions of gcc Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Removed unnecessary thread synchronization Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Added benchmarks for contraction on CPU. Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: New multithreaded contraction that doesn't rely on the thread pool to run the closure in the order in which they are enqueued. This is needed in order to switch to the new non blocking thread pool since this new thread pool can execute the closure in any order. Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: Turnon the new thread pool by default since it scales much better over multiple cores. It is still possible to revert to the old thread pool by compiling with the EIGEN_USE_SIMPLE_THREAD_POOL define. Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed syntax error Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Turn on the cost model by default. This results in some significant speedups for smaller tensors. For example, below are the results for the various tensor reductions. Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Added missing costPerCoeff method Bitbucket

• [eigen-commits] commit/eigen: 3 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Allow vectorized padding on GPU. This helps speed things up a little Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Enable the use of the packet api to evaluate tensor broadcasts. This speed things up quite a bit: Bitbucket
• [eigen-commits] commit/eigen: 6 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed compilation error in the tensor thread pool Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed compilation error Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: #if defined(EIGEN_USE_NONBLOCKING_THREAD_POOL) is now #if !defined(EIGEN_USE_SIMPLE_THREAD_POOL): the non blocking thread pool is the default since it's more scalable, and one needs to request the old thread pool explicitly. Bitbucket

• [eigen-commits] commit/eigen: ggael: Bug 1213: rename some enums type for consistency. Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1213: backport "Give names to anonymous enums" to workaround gcc linking issues. Bitbucket
• [eigen-commits] commit/eigen: ggael: Big 1213: add regression unit test. Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1222: fix compilation in AutoDiffScalar and add respective unit test Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1222: fix compilation in AutoDiffScalar and add respective unit test Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1229: bypass usage of Derived::Options which is available for plain matrix types only. Better use column-major storage anyway. Bitbucket
• [eigen-commits] commit/eigen: 3 new changesets Bitbucket
• [eigen-commits] commit/eigen: ggael: Use coeff(i,j) instead of operator(). Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1231: fix compilation regression regarding complex_array/=real_array and add respective unit tests Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Advertize the packet api of the tensor reducers iff the corresponding packet primitives are available. Bitbucket

• [eigen-commits] commit/eigen: ggael: Bug 1230: add support for SelfadjointView::triangularView. Bitbucket
• [eigen-commits] commit/eigen: ggael: Fix SelfAdjointView::triangularView for complexes. Bitbucket
• [eigen-commits] commit/eigen: 3 new changesets Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1221: disable gcc 6 warning: ignoring attributes on template argument Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1221: disable gcc 6 warning: ignoring attributes on template argument Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1201: optimize affine*vector products Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: ggael: Fix unit test. Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket

• [eigen-commits] commit/eigen: 3 new changesets Bitbucket
• [eigen-commits] commit/eigen: ggael: Improve doc of special math functions Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: 7 new changesets Bitbucket

• [eigen-commits] commit/eigen: chtz: Identify clang++ even if it is not named llvm-clang++ Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: chtz: Silenced several double-promotion warnings Bitbucket

• [eigen-commits] commit/eigen: chtz: Replace multiple constructors of half-type by a generic/templated constructor. This fixes an incompatibility with long double, exposed by the previous commit. Bitbucket
• [eigen-commits] commit/eigen: 5 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Merged in rmlarsen/eigen (pull request #188) Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed a typo in the array.cpp test Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket

• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: ptranspose is not a template. Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Include mmintrin.h to make it possible to use mmx instructions when needed. For example, this will enable the definition of a half packet for the Packet4f type. Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Added missing EIGEN_DEVICE_FUNC qualifier Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Worked around missing clang intrinsics Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Use the generic ploadquad intrinsics since it does the job Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Worked around missing clang intrinsic Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Don't attempt to use MMX instructions with visualstudio since they're only partially supported. Bitbucket
• [eigen-commits] commit/eigen: Benoit Jacob: Remove the rotating kernel. It was only useful on some ARM CPUs (Qualcomm Krait) that are not as ubiquitous today as they were when I introduced it. Bitbucket
• [eigen-commits] commit/eigen: Benoit Jacob: Remove now-unused protate PacketMath func Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 256: enable vectorization with unaligned loads/stores. Bitbucket
• [eigen-commits] commit/eigen: ggael: Workaround clang/llvm bug in code generation. Bitbucket
• [eigen-commits] commit/eigen: ggael: Document EIGEN_UNALIGNED_VECTORIZE. Bitbucket
• [eigen-commits] commit/eigen: ggael: update doc wrt to unaligned vectorization Bitbucket

• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Cleaned up the fp16 code a little more Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed compilation warning Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: There is no need to make the fp16 full reduction kernel a static function. Bitbucket
• [eigen-commits] commit/eigen: 2 new changesets Bitbucket
• [eigen-commits] commit/eigen: ggael: Fix warning. Bitbucket
• [eigen-commits] commit/eigen: ggael: Fix typo in dont_over_optimize Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Use numext:: instead of std:: functions. Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Disable the use of MMX instructions on x86_64 since too many compilers only support them in 32bit mode Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Added support for statically known lists of pairs of indices Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Improved the performance of tensor padding Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Made the index pair list code more portable accross various compilers Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Made the IndexPair code compile in non cxx11 mode Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Explicitly specify the rounding mode when converting floats to fp16 Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Updated the README file for the tensor benchmarks Bitbucket

• [eigen-commits] commit/eigen: benoitsteiner: Specify the rounding mode in the correct location Bitbucket
• [eigen-commits] commit/eigen: 5 new changesets Bitbucket
• [eigen-commits] commit/eigen: ggael: Add missing inclusion of mmintrin.h Bitbucket
• [eigen-commits] commit/eigen: ggael: Turn off ICC's conversion warning in is_convertible implementation Bitbucket
• [eigen-commits] commit/eigen: ggael: Introduce internal's UIntPtr and IntPtr types for pointer to integer conversions. Bitbucket
• [eigen-commits] commit/eigen: ggael: Remove debuging code. Bitbucket
• [eigen-commits] commit/eigen: chtz: Temporary workaround for Bug 1237. The snippet (expectedly) failed with enabled assertions. Bitbucket
• [eigen-commits] commit/eigen: 11 new changesets Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1236: fix possible integer overflow in density estimation. Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1236: fix possible integer overflow in density estimation. Bitbucket
• [eigen-commits] commit/eigen: ggael: Disable usage of MMX with msvc. Bitbucket
• [eigen-commits] commit/eigen: chtz: Bug 1237: Redefine eigen_assert instead of disabling assertions for documentation snippets Bitbucket
• [eigen-commits] commit/eigen: 7 new changesets Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Resolved merge conflicts Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Preserve the ability to vectorize the evaluation of an expression even when it involves a cast that isn't vectorized (e.g fp16 to float) Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Deleted extra namespace Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed some compilation warnings Bitbucket

• [eigen-commits] commit/eigen: ggael: Fix compilation with old ICC version (use C99 types instead of C++11 ones) Bitbucket
• [eigen-commits] commit/eigen: ggael: Fix compilation when defaulting to row-major Bitbucket
• [eigen-commits] commit/eigen: chtz: Cleaner implementation of dont_over_optimize. Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Disable the use of MMX instructions since the code is broken on many platforms Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Fixed option '--relaxed-constexpr' has been deprecated and replaced by option '--expt-relaxed-constexpr' warning generated by nvcc 7.5 Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Added a new operation to enable more powerful tensorindexing. Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Use NULL instead of nullptr to preserve the compatibility with cxx03 Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Reimplement clamp as a static function. Bitbucket
• [eigen-commits] commit/eigen: benoitsteiner: Silenced some compilation warnings triggered by nvcc 8.0 Bitbucket

• [eigen-commits] commit/eigen: ggael: Bug 1238: fix SparseMatrix::sum() overload for un-compressed mode. Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1238: fix SparseMatrix::sum() overload for un-compressed mode. Bitbucket
• [eigen-commits] commit/eigen: ggael: Fix compilation with old icc Bitbucket
• [eigen-commits] commit/eigen: ggael: Bug 1181: help MSVC inlining. Bitbucket
• [eigen-commits] commit/eigen: 3 new changesets Bitbucket