Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Develop upstream sync 230130 #1981

Closed
wants to merge 432 commits into from
Closed

Develop upstream sync 230130 #1981

wants to merge 432 commits into from

Conversation

i-chaochen
Copy link

@i-chaochen i-chaochen commented Jan 31, 2023 •
edited
Loading

aartbik and others added 30 commits January 24, 2023 14:46
@aartbik @tensorflower-gardener
[mlir][xla][sparse] test XLA:CPU-next sparse pipeline in isolation
b16afd7 
PiperOrigin-RevId: 504378746
@ddunl @tensorflower-gardener
Add test for compiler/xla/tools:replay_computation_test_{cpu,interpre…
e2d7e0e 
…ter}

PiperOrigin-RevId: 504382614
@jszaday @tensorflower-gardener
Support TPUPartitionedInputV2 with Packed Variables.
149abd5 
PiperOrigin-RevId: 504386021
@tensorflower-gardener
Fix a typo in the doc.
6520d70 
PiperOrigin-RevId: 5043939
@w-xinyi @tensorflower-gardener
Update exit code.
4f2935c 
PiperOrigin-RevId: 504396775
@dansuh17 @tensorflower-gardener
Preserve function aliases through the TF quantizer.
50cdf7a 
Function aliases associate user-defined names to TF functions. This information is in the `MetaGraphDef` and
is used to identify specific functions of interest (the actual function names change every time the model is saved as SavedModel).

This change preserves the function aliases through the TF functions, so that the same functions (quantized, after the quantization passes) are still associated with the same aliases.
This is done by preventing the aliased functions from being inlined by `MarkFunctionsNoinlinePass`. The pass receives a list of names that should not be inlined, so the names should be passed along from the quantizer's python API.

PiperOrigin-RevId: 504396824
@grantjensen @tensorflower-gardener
[tflite-gpu] Remove redundant addr calculations
34b5e57 
PiperOrigin-RevId: 504405735
@angerson @tensorflower-gardener
Update description to Jan 2023
9288c15 
PiperOrigin-RevId: 504406216
@mpcallanan @tensorflower-gardener
Add functionality to detect lost workers in tf.data service dispatcher.
e780748 
PiperOrigin-RevId: 504407156
@qqfish @tensorflower-gardener
Update LOG message to avoid confusion.
b63d9a4 
Github issue:tensorflow#59117

PiperOrigin-RevId: 504408822
@tensorflower-gardener
Merge pull request tensorflow#59101 from trevor-m:tmorris-dtensor-red…
36b9864 
…ucescatter

PiperOrigin-RevId: 504415558
@rohitju @tensorflower-gardener
Disable //tensorflow/python/kernel_tests/math_ops:approx_topk_test_cp…
37d9d15 
…u on OSS

PiperOrigin-RevId: 504416200
@grantjensen @tensorflower-gardener
Update NDK version to r21e
5b77f23 
PiperOrigin-RevId: 504416827
@tensorflower-gardener
Update ops-related pbtxt files.
229b86e 
PiperOrigin-RevId: 504427375
@tensorflower-gardener
Go: Update generated wrapper functions for TensorFlow ops.
aef22c0 
PiperOrigin-RevId: 504431855
@rohitju @tensorflower-gardener
Rollback of PR tensorflow#58640
929c425 
END_PUBLIC

Breaks TF Kokoro builds

*** Original change description ***

BEGIN_OLD_PUBLIC
PR tensorflow#58640: [INTEL oneDNN] OneDNN Refactoring Part 1: Remove obsolete operations

Imported from GitHub PR tensorflow#58640

This PR removes obsolete oneDNN operations (oneDNN blocked format related) as well as MKL layout pass tests and conversion tests (test cases are all related to blocked format).

Results:
   1. 12 source files are deleted
   2. 8 source files are update (bazel files, tests)
   3. Total 8843 lines of code are cleaned away!
Copyba...

***

BEGIN_OLD_PUBLIC
Rollback to unbreak

PiperOrigin-RevId: 504447880
@Maratyszcza @tensorflower-gardener
Update XNNPACK and cpuinfo dependencies
d90b6bc 
PiperOrigin-RevId: 504450668
@tensorflower-gardener
Merge pull request tensorflow#59430 from elfringham:h5py_limit
185b29d 
PiperOrigin-RevId: 504454553
@weinan1997 @tensorflower-gardener
Update preemption watcher metric name to keep consistency.
96fef60 
PiperOrigin-RevId: 504459690
@tensorflower-gardener
Integrate LLVM at llvm/llvm-project@6363320ba6be
905ff6e 
Updates LLVM usage to match
[6363320ba6be](llvm/llvm-project@6363320ba6be)

PiperOrigin-RevId: 504462204
@tensorflower-gardener
Update TFRT dependency to use revision
02f9f0d 
http://github.com/tensorflow/runtime/commit/78367d1146727ed89f4eab3d2197a7ffda260524.

PiperOrigin-RevId: 504463972
@Maratyszcza @tensorflower-gardener
Remove reference to clog license
5d709c7 
clog library is no longer used in TensorFlow

PiperOrigin-RevId: 504469358
@tensorflower-gardener
Extend shim headers visibility for MediaPipe
2a5e328 
PiperOrigin-RevId: 504472780
@akuegel @tensorflower-gardener
Revert Move ROCM build to 5.3
d21b557 
Also add python 3.10 and python 3.11

PiperOrigin-RevId: 504480644
@tensorflower-gardener
compat: Update forward compatibility horizon to 2023-01-25
a334449 
PiperOrigin-RevId: 504489854
@tensorflower-gardener
Update GraphDef version to 1387.
70df6cb 
PiperOrigin-RevId: 504489888
@elfringham
[Linaro:ARM_CI] Exclude tests that fail due to mismatch with tf-io
1852bc3 
These tests currently cannot run as the API to TensorFlow has
been updated and does not match the current release of tensorflow-io
@tensorflower-gardener
Merge pull request tensorflow#59256 from Intel-tensorflow:agramesh/cu…
1027b35 
…da_warnings

PiperOrigin-RevId: 504500145
@pifon2a @tensorflower-gardener
[GmlSt] Migrate transform_matmul_for_cpu away from gml_st.for.
21047f1 
PiperOrigin-RevId: 504506568
@tensorflower-gardener
Rollback of PR tensorflow#52014
1056482 
PR tensorflow#52014: [TFLite] Use integer-only operations for the quantized reference REDUCE_MEAN and REDUCE_SUM kernels

Imported from GitHub PR tensorflow#52014

Hello,

This PR refactors the different quantized reference reduce mean kernels into a single `QuantizedMeanOrSum` solving the following [TODO](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/lite/kernels/reduce.cc#L359) and also removes the now unused code.

The `QuantizedMeanOrSu...

PiperOrigin-RevId: 504510364
vam-google and others added 23 commits January 27, 2023 19:46
@vam-google @tensorflower-gardener
feat: Upgrade TensorFlow protobuf dependency to 3.21.9 (4.21.9 for Py…
84f4092 
…thon)

PiperOrigin-RevId: 505264303
@ezhulenev @tensorflower-gardener
[xla:gpu] PjRtClient: Use thread pool to run Send operations
4c27a8d 
Copying memory from host is a slow and blocking operation, launch it into a thread pool and signal completion via the AsyncValue and se::Event.

PiperOrigin-RevId: 505268768
@tensorflower-gardener
Update GraphDef version to 1390.
fa92339 
PiperOrigin-RevId: 505304324
@tensorflower-gardener
compat: Update forward compatibility horizon to 2023-01-28
cdee5a6 
PiperOrigin-RevId: 505304330
@jreiffers @tensorflower-gardener
mlir_bisect: replace operands with constants, replace return values.
2590713 
Also be slightly smarter about applying strategies.

PiperOrigin-RevId: 505336951
@tensorflower-gardener
Merge pull request tensorflow#59100 from lhutton1:tosa-atan2-tfl-lega…
0e89f1a 
…lization

PiperOrigin-RevId: 505342980
@tensorflower-gardener
Include conditional ops in MLIR preferred ops for legalization
79de5ba 
PiperOrigin-RevId: 505363577
@tensorflower-gardener
Integrate LLVM at llvm/llvm-project@32ac9db7a5c8
530605f 
Updates LLVM usage to match
[32ac9db7a5c8](llvm/llvm-project@32ac9db7a5c8)

PiperOrigin-RevId: 505366611
@tensorflower-gardener
Update TFRT dependency to use revision
7a73997 
http://github.com/tensorflow/runtime/commit/d1c45099b1bc7db1cf4c27772d0b4874803ddec7.

PiperOrigin-RevId: 505367334
@pkanwar23 @tensorflower-gardener
[XLA] Allow MSA prefetch interval picker to use a shape override.
bd473bd 
Using a shape override causes all prefetches to consider the same duration for
prefetches, hence making it more likely to schedule prefetches in roughly FIFO
order.

PiperOrigin-RevId: 505429547
@tensorflower-gardener
Fix for LLVM/MLIR integrate
70fa776 
PiperOrigin-RevId: 505435402
@tensorflower-gardener
Merge pull request tensorflow#59455 from elfringham:mismatch_io
9b30263 
PiperOrigin-RevId: 505438930
@tensorflower-gardener
compat: Update forward compatibility horizon to 2023-01-29
33cdf31 
PiperOrigin-RevId: 505450806
@tensorflower-gardener
Update GraphDef version to 1391.
3c3ed18 
PiperOrigin-RevId: 505450980
@dansuh17 @tensorflower-gardener
Remove an obsolete TODO from remove_vars_in_session_initializer.cc
e22df2d 
PiperOrigin-RevId: 505548162
@tensorflower-gardener
Update GraphDef version to 1392.
a6d0db7 
PiperOrigin-RevId: 505615493
@tensorflower-gardener
compat: Update forward compatibility horizon to 2023-01-30
4bd6b6b 
PiperOrigin-RevId: 505615494
@tensorflower-gardener
Allow JitCompiler to be created from a ModuleOp directly.
889069c 
PiperOrigin-RevId: 505619534
@i-chaochen
init merge
6b8de00
@i-chaochen
resolve all conflicts
6ceea10
@i-chaochen
use ReorganizeMemory based on develop-upstream to rocm_blas.cc
8fb4e30
@i-chaochen
enable rocm/hip for multi_client_test
d29b6d6
@i-chaochen
disable nccl 2gpus tests
4432cf9
@i-chaochen
Copy link
Author

//tensorflow/core/profiler/backends/gpu:device_tracer_test_gpu is failed again.

We didn't disable it at the last week sync.

@i-chaochen
Copy link
Author

close this PR due to #1987

@i-chaochen i-chaochen closed this Feb 3, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@rahulbatra85 rahulbatra85 Awaiting requested review from rahulbatra85

@jayfurmanek jayfurmanek Awaiting requested review from jayfurmanek

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.