PiperOrigin-RevId: 481142604

	update of g3doc/_includes/tf_passes.md
	update of g3doc/hlo_ops.md

PiperOrigin-RevId: 481140074

PiperOrigin-RevId: 481135234

PiperOrigin-RevId: 481126296

The former name conflicts with the regular index_vector_dim from other scatters, which is always 1 here, so it's a bit confusing.

PiperOrigin-RevId: 481111212

hlo_module_importer is used in many places in ways we don't fully
understand. The assumptions they make about layouts seem to be
incompatible with the ones we make in the MLIR CPU pipeline.

Verified that tensorflow/compiler/xla/tests:pad_test_cpu
and tensorflow/compiler/xla/tests:client_test_cpu pass
with XLA_FLAGS=--xla_cpu_enable_mlir_lowering=true. This also
fixes tensorflow/compiler/xla/tests:slice_test_cpu.

PiperOrigin-RevId: 481108763

http://github.com/tensorflow/runtime/commit/a2cf754d51e0db0aeab14285c0377308959bf28d.

PiperOrigin-RevId: 481105622

PiperOrigin-RevId: 481105084

Prepare to change kDynamicSize to int64_t::min.

PiperOrigin-RevId: 481104488

Updates LLVM usage to match
[2f357054d705](https://github.com/llvm/llvm-project/commit/2f357054d705)

PiperOrigin-RevId: 481103846

PiperOrigin-RevId: 481101180

PiperOrigin-RevId: 481095035

TFLite CompatibilityList: Ensure native code is available by calling GpuDelegateNative::init(). This allows for a different library name and external loading by client.

PiperOrigin-RevId: 481093953

PiperOrigin-RevId: 481093435

PiperOrigin-RevId: 481092113

PiperOrigin-RevId: 481092096

PiperOrigin-RevId: 481089797

PiperOrigin-RevId: 481088303

Due to security issue, zlib 1.2.12 is yanked, just like 1.2.11 was before. So builds not using TF mirror will break.

See https://www.zlib.net

PiperOrigin-RevId: 481064905

Fix speedup calculation. Document that the numbers presented are running times (e.g., not flops rate).

PiperOrigin-RevId: 481052613

This changes the default value to the original intent: exposing headers only
is a "hack" in the build graph and should be an "opt-in". Tensorflow does that
when building its shared library, but that shouldn't be the majority of the
clients (hopefully).
This also aligns with the default value for enabling the shared library mode
in tensorflow: it's off by default (even though TensorFlow changes the
default in its bazelrc!)

PiperOrigin-RevId: 481050893

http://github.com/tensorflow/runtime/commit/c46ebbdbc7fb252b23de9b36c15045fab5df6c4e.

PiperOrigin-RevId: 481046722

CHANGELOG
=========
3bb6a48d8 - Fix bug atan2
14c847dc0 - Refactor special values test for pow, and add a similar test for atan2
462758e8a - Don'\''t use generic sign function for sign(complex) unless it is vectorizable
c0d6a7261 - Use pnegate(pzero(x)) as a generic way to generate -0.0. Some compiler do not handle the literal -0.0 properly in fastmath mode.
7846c7387 - Eigen/Sparse: fix warnings -Wunused-but-set-variable
316754487 - Handle NaN inputs to atan2.
72db3f0fa - Remove references to M_PI_2 and M_PI_4.
d6bc06259 - Remove reference to EIGEN_HAS_CXX11_MATH.
5ceed0d57 - Guard GCC-specific pragmas with "#ifdef EIGEN_COMP_GNUC"
528b68674 - [clang-format] Add a few macros to AttributeMacros
e95c4a837 - Simpler range reduction strategy for atan<float>().
80efbfded - Unconditionally enable CXX11 math.
e5794873c - Replace assert with eigen_assert.
7d6a9925c - Fix 4x4 inverse when compiling with -Ofast.
1414a76fa - Only vectorize atan<double> for Altivec if VSX is available.
c475228b2 - Vectorize atan() for double.
1e1848fdb - Add a vectorized implementation of atan2 to Eigen.

* Switch TensorFlow to using the new fast atan2 in Eigen.
* Get rid of local implementations since Eigen is now guaranteed to support C++11 math.

PiperOrigin-RevId: 481044760

http://github.com/tensorflow/runtime/commit/97d12a118984d820c14bdad0971704d95d99f411.

PiperOrigin-RevId: 481043847

Updates LLVM usage to match
[1fda6f6859aa](https://github.com/llvm/llvm-project/commit/1fda6f6859aa)

PiperOrigin-RevId: 481042584

PiperOrigin-RevId: 481042256

For tf.DivNoNan operators, the calibrator should NaN values since the lowered
form of TFLite will encounter NaN value by design.

PiperOrigin-RevId: 481039479

PiperOrigin-RevId: 481035658

PiperOrigin-RevId: 481034607

PiperOrigin-RevId: 481030210

Updating determinism check in svd op. Now the op works deterministically when enabled, except in the case that the input matrix has column size 1.

PiperOrigin-RevId: 481018683

PiperOrigin-RevId: 481016434

This CL introduces a new legalization pattern that rewrites sigmoid_grad
operation as a `grad = dy * y * (1 - y)`, where `y = sigmoid(x)`.

PiperOrigin-RevId: 481015565

JitExecutable is not required for AOT compilation, so XlaRuntimeCpuExecutable can own a runtime::Executable directly.

PiperOrigin-RevId: 481012743

PiperOrigin-RevId: 481010554

PiperOrigin-RevId: 481008739

PiperOrigin-RevId: 481007346

PiperOrigin-RevId: 481006537

PiperOrigin-RevId: 481006100

PiperOrigin-RevId: 481004699