tree bf6d7743bc595eda206f1b0ffaa9f8c8e1f883c1
parent e11e9aa6b6994c3cdad5739ae7e57fc0cfac07df
author Rasmus Munk Larsen <rmlarsen@google.com> 1683738231 -0700
committer Copybara-Service <copybara-worker@google.com> 1683738266 -0700

Update Eigen to commit:2709f4c8fbbe71a5383de1ba27e9833f218a642d

CHANGELOG
=========
2709f4c8f - Use relative path to include EmulateArray.h in CXX11Meta.h, and get rid of redundant meta-programming code, which was moved to Core.
9a02c977e - Use relative paths to include Meta.h and MaxSizeVector.h in Tensor
96c42771d - Make it possible to override the synchonization primitives used by the threadpool using macros.
1321821e8 - Add missing braces in Umeyama.h
524c329ab - Work around compiler bug in Umeyama.h.
fbf7189bd - Fix cuda compilation
062379193 - [SYCL-2020] Enabling USM support for SYCL. SYCL-1.2.1 did not have support for USM.
1698c367a - Use std::shared_ptr for FFTW/IMKL FFT plan implementation; Fixes #2651
1f79a6078 - Return NaN in ndtri for values outside valid input range.
94f57867f - Thread pool
9eb8e2afb - Change array_cwise test name
725c11719 - Visitor: fix modulo by zero compiler warning
b8208b363 - Specialized loadColData correctly - fix previous BF16 GEMV MR
2af03fb68 - clean up array_cwise test
fda1373a1 - Fix ColMajor BF16 GEMV for when vector is RowMajor
fdc749de2 - JacobiSVD: set m_nonzeroSingularValues to zero if not finite
6418ac028 - Unroll F32 to BF16 loop - 1.8X faster conversions for LLVM.  Use vector pairs for GCC.
874f5947f - Add half-`Packet` operations to `StridedLinearBufferCopy`.
c9a14f48d - SSE Packet4ui has pcmp, pmin, pmax

PiperOrigin-RevId: 530938227
Change-Id: Ibb0f0eb4bf97b5caf9db59b1fdeca354be07b728